| 1 | n/a | /* |
|---|
| 2 | n/a | * This file includes functions to transform a concrete syntax tree (CST) to |
|---|
| 3 | n/a | * an abstract syntax tree (AST). The main function is PyAST_FromNode(). |
|---|
| 4 | n/a | * |
|---|
| 5 | n/a | */ |
|---|
| 6 | n/a | #include "Python.h" |
|---|
| 7 | n/a | #include "Python-ast.h" |
|---|
| 8 | n/a | #include "node.h" |
|---|
| 9 | n/a | #include "ast.h" |
|---|
| 10 | n/a | #include "token.h" |
|---|
| 11 | n/a | |
|---|
| 12 | n/a | #include <assert.h> |
|---|
| 13 | n/a | |
|---|
| 14 | n/a | static int validate_stmts(asdl_seq *); |
|---|
| 15 | n/a | static int validate_exprs(asdl_seq *, expr_context_ty, int); |
|---|
| 16 | n/a | static int validate_nonempty_seq(asdl_seq *, const char *, const char *); |
|---|
| 17 | n/a | static int validate_stmt(stmt_ty); |
|---|
| 18 | n/a | static int validate_expr(expr_ty, expr_context_ty); |
|---|
| 19 | n/a | |
|---|
| 20 | n/a | static int |
|---|
| 21 | n/a | validate_comprehension(asdl_seq *gens) |
|---|
| 22 | n/a | { |
|---|
| 23 | n/a | int i; |
|---|
| 24 | n/a | if (!asdl_seq_LEN(gens)) { |
|---|
| 25 | n/a | PyErr_SetString(PyExc_ValueError, "comprehension with no generators"); |
|---|
| 26 | n/a | return 0; |
|---|
| 27 | n/a | } |
|---|
| 28 | n/a | for (i = 0; i < asdl_seq_LEN(gens); i++) { |
|---|
| 29 | n/a | comprehension_ty comp = asdl_seq_GET(gens, i); |
|---|
| 30 | n/a | if (!validate_expr(comp->target, Store) || |
|---|
| 31 | n/a | !validate_expr(comp->iter, Load) || |
|---|
| 32 | n/a | !validate_exprs(comp->ifs, Load, 0)) |
|---|
| 33 | n/a | return 0; |
|---|
| 34 | n/a | } |
|---|
| 35 | n/a | return 1; |
|---|
| 36 | n/a | } |
|---|
| 37 | n/a | |
|---|
| 38 | n/a | static int |
|---|
| 39 | n/a | validate_slice(slice_ty slice) |
|---|
| 40 | n/a | { |
|---|
| 41 | n/a | switch (slice->kind) { |
|---|
| 42 | n/a | case Slice_kind: |
|---|
| 43 | n/a | return (!slice->v.Slice.lower || validate_expr(slice->v.Slice.lower, Load)) && |
|---|
| 44 | n/a | (!slice->v.Slice.upper || validate_expr(slice->v.Slice.upper, Load)) && |
|---|
| 45 | n/a | (!slice->v.Slice.step || validate_expr(slice->v.Slice.step, Load)); |
|---|
| 46 | n/a | case ExtSlice_kind: { |
|---|
| 47 | n/a | int i; |
|---|
| 48 | n/a | if (!validate_nonempty_seq(slice->v.ExtSlice.dims, "dims", "ExtSlice")) |
|---|
| 49 | n/a | return 0; |
|---|
| 50 | n/a | for (i = 0; i < asdl_seq_LEN(slice->v.ExtSlice.dims); i++) |
|---|
| 51 | n/a | if (!validate_slice(asdl_seq_GET(slice->v.ExtSlice.dims, i))) |
|---|
| 52 | n/a | return 0; |
|---|
| 53 | n/a | return 1; |
|---|
| 54 | n/a | } |
|---|
| 55 | n/a | case Index_kind: |
|---|
| 56 | n/a | return validate_expr(slice->v.Index.value, Load); |
|---|
| 57 | n/a | default: |
|---|
| 58 | n/a | PyErr_SetString(PyExc_SystemError, "unknown slice node"); |
|---|
| 59 | n/a | return 0; |
|---|
| 60 | n/a | } |
|---|
| 61 | n/a | } |
|---|
| 62 | n/a | |
|---|
| 63 | n/a | static int |
|---|
| 64 | n/a | validate_keywords(asdl_seq *keywords) |
|---|
| 65 | n/a | { |
|---|
| 66 | n/a | int i; |
|---|
| 67 | n/a | for (i = 0; i < asdl_seq_LEN(keywords); i++) |
|---|
| 68 | n/a | if (!validate_expr(((keyword_ty)asdl_seq_GET(keywords, i))->value, Load)) |
|---|
| 69 | n/a | return 0; |
|---|
| 70 | n/a | return 1; |
|---|
| 71 | n/a | } |
|---|
| 72 | n/a | |
|---|
| 73 | n/a | static int |
|---|
| 74 | n/a | validate_args(asdl_seq *args) |
|---|
| 75 | n/a | { |
|---|
| 76 | n/a | int i; |
|---|
| 77 | n/a | for (i = 0; i < asdl_seq_LEN(args); i++) { |
|---|
| 78 | n/a | arg_ty arg = asdl_seq_GET(args, i); |
|---|
| 79 | n/a | if (arg->annotation && !validate_expr(arg->annotation, Load)) |
|---|
| 80 | n/a | return 0; |
|---|
| 81 | n/a | } |
|---|
| 82 | n/a | return 1; |
|---|
| 83 | n/a | } |
|---|
| 84 | n/a | |
|---|
| 85 | n/a | static const char * |
|---|
| 86 | n/a | expr_context_name(expr_context_ty ctx) |
|---|
| 87 | n/a | { |
|---|
| 88 | n/a | switch (ctx) { |
|---|
| 89 | n/a | case Load: |
|---|
| 90 | n/a | return "Load"; |
|---|
| 91 | n/a | case Store: |
|---|
| 92 | n/a | return "Store"; |
|---|
| 93 | n/a | case Del: |
|---|
| 94 | n/a | return "Del"; |
|---|
| 95 | n/a | case AugLoad: |
|---|
| 96 | n/a | return "AugLoad"; |
|---|
| 97 | n/a | case AugStore: |
|---|
| 98 | n/a | return "AugStore"; |
|---|
| 99 | n/a | case Param: |
|---|
| 100 | n/a | return "Param"; |
|---|
| 101 | n/a | default: |
|---|
| 102 | n/a | assert(0); |
|---|
| 103 | n/a | return "(unknown)"; |
|---|
| 104 | n/a | } |
|---|
| 105 | n/a | } |
|---|
| 106 | n/a | |
|---|
| 107 | n/a | static int |
|---|
| 108 | n/a | validate_arguments(arguments_ty args) |
|---|
| 109 | n/a | { |
|---|
| 110 | n/a | if (!validate_args(args->args)) |
|---|
| 111 | n/a | return 0; |
|---|
| 112 | n/a | if (args->vararg && args->vararg->annotation |
|---|
| 113 | n/a | && !validate_expr(args->vararg->annotation, Load)) { |
|---|
| 114 | n/a | return 0; |
|---|
| 115 | n/a | } |
|---|
| 116 | n/a | if (!validate_args(args->kwonlyargs)) |
|---|
| 117 | n/a | return 0; |
|---|
| 118 | n/a | if (args->kwarg && args->kwarg->annotation |
|---|
| 119 | n/a | && !validate_expr(args->kwarg->annotation, Load)) { |
|---|
| 120 | n/a | return 0; |
|---|
| 121 | n/a | } |
|---|
| 122 | n/a | if (asdl_seq_LEN(args->defaults) > asdl_seq_LEN(args->args)) { |
|---|
| 123 | n/a | PyErr_SetString(PyExc_ValueError, "more positional defaults than args on arguments"); |
|---|
| 124 | n/a | return 0; |
|---|
| 125 | n/a | } |
|---|
| 126 | n/a | if (asdl_seq_LEN(args->kw_defaults) != asdl_seq_LEN(args->kwonlyargs)) { |
|---|
| 127 | n/a | PyErr_SetString(PyExc_ValueError, "length of kwonlyargs is not the same as " |
|---|
| 128 | n/a | "kw_defaults on arguments"); |
|---|
| 129 | n/a | return 0; |
|---|
| 130 | n/a | } |
|---|
| 131 | n/a | return validate_exprs(args->defaults, Load, 0) && validate_exprs(args->kw_defaults, Load, 1); |
|---|
| 132 | n/a | } |
|---|
| 133 | n/a | |
|---|
| 134 | n/a | static int |
|---|
| 135 | n/a | validate_constant(PyObject *value) |
|---|
| 136 | n/a | { |
|---|
| 137 | n/a | if (value == Py_None || value == Py_Ellipsis) |
|---|
| 138 | n/a | return 1; |
|---|
| 139 | n/a | |
|---|
| 140 | n/a | if (PyLong_CheckExact(value) |
|---|
| 141 | n/a | || PyFloat_CheckExact(value) |
|---|
| 142 | n/a | || PyComplex_CheckExact(value) |
|---|
| 143 | n/a | || PyBool_Check(value) |
|---|
| 144 | n/a | || PyUnicode_CheckExact(value) |
|---|
| 145 | n/a | || PyBytes_CheckExact(value)) |
|---|
| 146 | n/a | return 1; |
|---|
| 147 | n/a | |
|---|
| 148 | n/a | if (PyTuple_CheckExact(value) || PyFrozenSet_CheckExact(value)) { |
|---|
| 149 | n/a | PyObject *it; |
|---|
| 150 | n/a | |
|---|
| 151 | n/a | it = PyObject_GetIter(value); |
|---|
| 152 | n/a | if (it == NULL) |
|---|
| 153 | n/a | return 0; |
|---|
| 154 | n/a | |
|---|
| 155 | n/a | while (1) { |
|---|
| 156 | n/a | PyObject *item = PyIter_Next(it); |
|---|
| 157 | n/a | if (item == NULL) { |
|---|
| 158 | n/a | if (PyErr_Occurred()) { |
|---|
| 159 | n/a | Py_DECREF(it); |
|---|
| 160 | n/a | return 0; |
|---|
| 161 | n/a | } |
|---|
| 162 | n/a | break; |
|---|
| 163 | n/a | } |
|---|
| 164 | n/a | |
|---|
| 165 | n/a | if (!validate_constant(item)) { |
|---|
| 166 | n/a | Py_DECREF(it); |
|---|
| 167 | n/a | Py_DECREF(item); |
|---|
| 168 | n/a | return 0; |
|---|
| 169 | n/a | } |
|---|
| 170 | n/a | Py_DECREF(item); |
|---|
| 171 | n/a | } |
|---|
| 172 | n/a | |
|---|
| 173 | n/a | Py_DECREF(it); |
|---|
| 174 | n/a | return 1; |
|---|
| 175 | n/a | } |
|---|
| 176 | n/a | |
|---|
| 177 | n/a | return 0; |
|---|
| 178 | n/a | } |
|---|
| 179 | n/a | |
|---|
| 180 | n/a | static int |
|---|
| 181 | n/a | validate_expr(expr_ty exp, expr_context_ty ctx) |
|---|
| 182 | n/a | { |
|---|
| 183 | n/a | int check_ctx = 1; |
|---|
| 184 | n/a | expr_context_ty actual_ctx; |
|---|
| 185 | n/a | |
|---|
| 186 | n/a | /* First check expression context. */ |
|---|
| 187 | n/a | switch (exp->kind) { |
|---|
| 188 | n/a | case Attribute_kind: |
|---|
| 189 | n/a | actual_ctx = exp->v.Attribute.ctx; |
|---|
| 190 | n/a | break; |
|---|
| 191 | n/a | case Subscript_kind: |
|---|
| 192 | n/a | actual_ctx = exp->v.Subscript.ctx; |
|---|
| 193 | n/a | break; |
|---|
| 194 | n/a | case Starred_kind: |
|---|
| 195 | n/a | actual_ctx = exp->v.Starred.ctx; |
|---|
| 196 | n/a | break; |
|---|
| 197 | n/a | case Name_kind: |
|---|
| 198 | n/a | actual_ctx = exp->v.Name.ctx; |
|---|
| 199 | n/a | break; |
|---|
| 200 | n/a | case List_kind: |
|---|
| 201 | n/a | actual_ctx = exp->v.List.ctx; |
|---|
| 202 | n/a | break; |
|---|
| 203 | n/a | case Tuple_kind: |
|---|
| 204 | n/a | actual_ctx = exp->v.Tuple.ctx; |
|---|
| 205 | n/a | break; |
|---|
| 206 | n/a | default: |
|---|
| 207 | n/a | if (ctx != Load) { |
|---|
| 208 | n/a | PyErr_Format(PyExc_ValueError, "expression which can't be " |
|---|
| 209 | n/a | "assigned to in %s context", expr_context_name(ctx)); |
|---|
| 210 | n/a | return 0; |
|---|
| 211 | n/a | } |
|---|
| 212 | n/a | check_ctx = 0; |
|---|
| 213 | n/a | /* set actual_ctx to prevent gcc warning */ |
|---|
| 214 | n/a | actual_ctx = 0; |
|---|
| 215 | n/a | } |
|---|
| 216 | n/a | if (check_ctx && actual_ctx != ctx) { |
|---|
| 217 | n/a | PyErr_Format(PyExc_ValueError, "expression must have %s context but has %s instead", |
|---|
| 218 | n/a | expr_context_name(ctx), expr_context_name(actual_ctx)); |
|---|
| 219 | n/a | return 0; |
|---|
| 220 | n/a | } |
|---|
| 221 | n/a | |
|---|
| 222 | n/a | /* Now validate expression. */ |
|---|
| 223 | n/a | switch (exp->kind) { |
|---|
| 224 | n/a | case BoolOp_kind: |
|---|
| 225 | n/a | if (asdl_seq_LEN(exp->v.BoolOp.values) < 2) { |
|---|
| 226 | n/a | PyErr_SetString(PyExc_ValueError, "BoolOp with less than 2 values"); |
|---|
| 227 | n/a | return 0; |
|---|
| 228 | n/a | } |
|---|
| 229 | n/a | return validate_exprs(exp->v.BoolOp.values, Load, 0); |
|---|
| 230 | n/a | case BinOp_kind: |
|---|
| 231 | n/a | return validate_expr(exp->v.BinOp.left, Load) && |
|---|
| 232 | n/a | validate_expr(exp->v.BinOp.right, Load); |
|---|
| 233 | n/a | case UnaryOp_kind: |
|---|
| 234 | n/a | return validate_expr(exp->v.UnaryOp.operand, Load); |
|---|
| 235 | n/a | case Lambda_kind: |
|---|
| 236 | n/a | return validate_arguments(exp->v.Lambda.args) && |
|---|
| 237 | n/a | validate_expr(exp->v.Lambda.body, Load); |
|---|
| 238 | n/a | case IfExp_kind: |
|---|
| 239 | n/a | return validate_expr(exp->v.IfExp.test, Load) && |
|---|
| 240 | n/a | validate_expr(exp->v.IfExp.body, Load) && |
|---|
| 241 | n/a | validate_expr(exp->v.IfExp.orelse, Load); |
|---|
| 242 | n/a | case Dict_kind: |
|---|
| 243 | n/a | if (asdl_seq_LEN(exp->v.Dict.keys) != asdl_seq_LEN(exp->v.Dict.values)) { |
|---|
| 244 | n/a | PyErr_SetString(PyExc_ValueError, |
|---|
| 245 | n/a | "Dict doesn't have the same number of keys as values"); |
|---|
| 246 | n/a | return 0; |
|---|
| 247 | n/a | } |
|---|
| 248 | n/a | /* null_ok=1 for keys expressions to allow dict unpacking to work in |
|---|
| 249 | n/a | dict literals, i.e. ``{**{a:b}}`` */ |
|---|
| 250 | n/a | return validate_exprs(exp->v.Dict.keys, Load, /*null_ok=*/ 1) && |
|---|
| 251 | n/a | validate_exprs(exp->v.Dict.values, Load, /*null_ok=*/ 0); |
|---|
| 252 | n/a | case Set_kind: |
|---|
| 253 | n/a | return validate_exprs(exp->v.Set.elts, Load, 0); |
|---|
| 254 | n/a | #define COMP(NAME) \ |
|---|
| 255 | n/a | case NAME ## _kind: \ |
|---|
| 256 | n/a | return validate_comprehension(exp->v.NAME.generators) && \ |
|---|
| 257 | n/a | validate_expr(exp->v.NAME.elt, Load); |
|---|
| 258 | n/a | COMP(ListComp) |
|---|
| 259 | n/a | COMP(SetComp) |
|---|
| 260 | n/a | COMP(GeneratorExp) |
|---|
| 261 | n/a | #undef COMP |
|---|
| 262 | n/a | case DictComp_kind: |
|---|
| 263 | n/a | return validate_comprehension(exp->v.DictComp.generators) && |
|---|
| 264 | n/a | validate_expr(exp->v.DictComp.key, Load) && |
|---|
| 265 | n/a | validate_expr(exp->v.DictComp.value, Load); |
|---|
| 266 | n/a | case Yield_kind: |
|---|
| 267 | n/a | return !exp->v.Yield.value || validate_expr(exp->v.Yield.value, Load); |
|---|
| 268 | n/a | case YieldFrom_kind: |
|---|
| 269 | n/a | return validate_expr(exp->v.YieldFrom.value, Load); |
|---|
| 270 | n/a | case Await_kind: |
|---|
| 271 | n/a | return validate_expr(exp->v.Await.value, Load); |
|---|
| 272 | n/a | case Compare_kind: |
|---|
| 273 | n/a | if (!asdl_seq_LEN(exp->v.Compare.comparators)) { |
|---|
| 274 | n/a | PyErr_SetString(PyExc_ValueError, "Compare with no comparators"); |
|---|
| 275 | n/a | return 0; |
|---|
| 276 | n/a | } |
|---|
| 277 | n/a | if (asdl_seq_LEN(exp->v.Compare.comparators) != |
|---|
| 278 | n/a | asdl_seq_LEN(exp->v.Compare.ops)) { |
|---|
| 279 | n/a | PyErr_SetString(PyExc_ValueError, "Compare has a different number " |
|---|
| 280 | n/a | "of comparators and operands"); |
|---|
| 281 | n/a | return 0; |
|---|
| 282 | n/a | } |
|---|
| 283 | n/a | return validate_exprs(exp->v.Compare.comparators, Load, 0) && |
|---|
| 284 | n/a | validate_expr(exp->v.Compare.left, Load); |
|---|
| 285 | n/a | case Call_kind: |
|---|
| 286 | n/a | return validate_expr(exp->v.Call.func, Load) && |
|---|
| 287 | n/a | validate_exprs(exp->v.Call.args, Load, 0) && |
|---|
| 288 | n/a | validate_keywords(exp->v.Call.keywords); |
|---|
| 289 | n/a | case Constant_kind: |
|---|
| 290 | n/a | if (!validate_constant(exp->v.Constant.value)) { |
|---|
| 291 | n/a | PyErr_Format(PyExc_TypeError, |
|---|
| 292 | n/a | "got an invalid type in Constant: %s", |
|---|
| 293 | n/a | Py_TYPE(exp->v.Constant.value)->tp_name); |
|---|
| 294 | n/a | return 0; |
|---|
| 295 | n/a | } |
|---|
| 296 | n/a | return 1; |
|---|
| 297 | n/a | case Num_kind: { |
|---|
| 298 | n/a | PyObject *n = exp->v.Num.n; |
|---|
| 299 | n/a | if (!PyLong_CheckExact(n) && !PyFloat_CheckExact(n) && |
|---|
| 300 | n/a | !PyComplex_CheckExact(n)) { |
|---|
| 301 | n/a | PyErr_SetString(PyExc_TypeError, "non-numeric type in Num"); |
|---|
| 302 | n/a | return 0; |
|---|
| 303 | n/a | } |
|---|
| 304 | n/a | return 1; |
|---|
| 305 | n/a | } |
|---|
| 306 | n/a | case Str_kind: { |
|---|
| 307 | n/a | PyObject *s = exp->v.Str.s; |
|---|
| 308 | n/a | if (!PyUnicode_CheckExact(s)) { |
|---|
| 309 | n/a | PyErr_SetString(PyExc_TypeError, "non-string type in Str"); |
|---|
| 310 | n/a | return 0; |
|---|
| 311 | n/a | } |
|---|
| 312 | n/a | return 1; |
|---|
| 313 | n/a | } |
|---|
| 314 | n/a | case JoinedStr_kind: |
|---|
| 315 | n/a | return validate_exprs(exp->v.JoinedStr.values, Load, 0); |
|---|
| 316 | n/a | case FormattedValue_kind: |
|---|
| 317 | n/a | if (validate_expr(exp->v.FormattedValue.value, Load) == 0) |
|---|
| 318 | n/a | return 0; |
|---|
| 319 | n/a | if (exp->v.FormattedValue.format_spec) |
|---|
| 320 | n/a | return validate_expr(exp->v.FormattedValue.format_spec, Load); |
|---|
| 321 | n/a | return 1; |
|---|
| 322 | n/a | case Bytes_kind: { |
|---|
| 323 | n/a | PyObject *b = exp->v.Bytes.s; |
|---|
| 324 | n/a | if (!PyBytes_CheckExact(b)) { |
|---|
| 325 | n/a | PyErr_SetString(PyExc_TypeError, "non-bytes type in Bytes"); |
|---|
| 326 | n/a | return 0; |
|---|
| 327 | n/a | } |
|---|
| 328 | n/a | return 1; |
|---|
| 329 | n/a | } |
|---|
| 330 | n/a | case Attribute_kind: |
|---|
| 331 | n/a | return validate_expr(exp->v.Attribute.value, Load); |
|---|
| 332 | n/a | case Subscript_kind: |
|---|
| 333 | n/a | return validate_slice(exp->v.Subscript.slice) && |
|---|
| 334 | n/a | validate_expr(exp->v.Subscript.value, Load); |
|---|
| 335 | n/a | case Starred_kind: |
|---|
| 336 | n/a | return validate_expr(exp->v.Starred.value, ctx); |
|---|
| 337 | n/a | case List_kind: |
|---|
| 338 | n/a | return validate_exprs(exp->v.List.elts, ctx, 0); |
|---|
| 339 | n/a | case Tuple_kind: |
|---|
| 340 | n/a | return validate_exprs(exp->v.Tuple.elts, ctx, 0); |
|---|
| 341 | n/a | /* These last cases don't have any checking. */ |
|---|
| 342 | n/a | case Name_kind: |
|---|
| 343 | n/a | case NameConstant_kind: |
|---|
| 344 | n/a | case Ellipsis_kind: |
|---|
| 345 | n/a | return 1; |
|---|
| 346 | n/a | default: |
|---|
| 347 | n/a | PyErr_SetString(PyExc_SystemError, "unexpected expression"); |
|---|
| 348 | n/a | return 0; |
|---|
| 349 | n/a | } |
|---|
| 350 | n/a | } |
|---|
| 351 | n/a | |
|---|
| 352 | n/a | static int |
|---|
| 353 | n/a | validate_nonempty_seq(asdl_seq *seq, const char *what, const char *owner) |
|---|
| 354 | n/a | { |
|---|
| 355 | n/a | if (asdl_seq_LEN(seq)) |
|---|
| 356 | n/a | return 1; |
|---|
| 357 | n/a | PyErr_Format(PyExc_ValueError, "empty %s on %s", what, owner); |
|---|
| 358 | n/a | return 0; |
|---|
| 359 | n/a | } |
|---|
| 360 | n/a | |
|---|
| 361 | n/a | static int |
|---|
| 362 | n/a | validate_assignlist(asdl_seq *targets, expr_context_ty ctx) |
|---|
| 363 | n/a | { |
|---|
| 364 | n/a | return validate_nonempty_seq(targets, "targets", ctx == Del ? "Delete" : "Assign") && |
|---|
| 365 | n/a | validate_exprs(targets, ctx, 0); |
|---|
| 366 | n/a | } |
|---|
| 367 | n/a | |
|---|
| 368 | n/a | static int |
|---|
| 369 | n/a | validate_body(asdl_seq *body, const char *owner) |
|---|
| 370 | n/a | { |
|---|
| 371 | n/a | return validate_nonempty_seq(body, "body", owner) && validate_stmts(body); |
|---|
| 372 | n/a | } |
|---|
| 373 | n/a | |
|---|
| 374 | n/a | static int |
|---|
| 375 | n/a | validate_stmt(stmt_ty stmt) |
|---|
| 376 | n/a | { |
|---|
| 377 | n/a | int i; |
|---|
| 378 | n/a | switch (stmt->kind) { |
|---|
| 379 | n/a | case FunctionDef_kind: |
|---|
| 380 | n/a | return validate_body(stmt->v.FunctionDef.body, "FunctionDef") && |
|---|
| 381 | n/a | validate_arguments(stmt->v.FunctionDef.args) && |
|---|
| 382 | n/a | validate_exprs(stmt->v.FunctionDef.decorator_list, Load, 0) && |
|---|
| 383 | n/a | (!stmt->v.FunctionDef.returns || |
|---|
| 384 | n/a | validate_expr(stmt->v.FunctionDef.returns, Load)); |
|---|
| 385 | n/a | case ClassDef_kind: |
|---|
| 386 | n/a | return validate_body(stmt->v.ClassDef.body, "ClassDef") && |
|---|
| 387 | n/a | validate_exprs(stmt->v.ClassDef.bases, Load, 0) && |
|---|
| 388 | n/a | validate_keywords(stmt->v.ClassDef.keywords) && |
|---|
| 389 | n/a | validate_exprs(stmt->v.ClassDef.decorator_list, Load, 0); |
|---|
| 390 | n/a | case Return_kind: |
|---|
| 391 | n/a | return !stmt->v.Return.value || validate_expr(stmt->v.Return.value, Load); |
|---|
| 392 | n/a | case Delete_kind: |
|---|
| 393 | n/a | return validate_assignlist(stmt->v.Delete.targets, Del); |
|---|
| 394 | n/a | case Assign_kind: |
|---|
| 395 | n/a | return validate_assignlist(stmt->v.Assign.targets, Store) && |
|---|
| 396 | n/a | validate_expr(stmt->v.Assign.value, Load); |
|---|
| 397 | n/a | case AugAssign_kind: |
|---|
| 398 | n/a | return validate_expr(stmt->v.AugAssign.target, Store) && |
|---|
| 399 | n/a | validate_expr(stmt->v.AugAssign.value, Load); |
|---|
| 400 | n/a | case AnnAssign_kind: |
|---|
| 401 | n/a | if (stmt->v.AnnAssign.target->kind != Name_kind && |
|---|
| 402 | n/a | stmt->v.AnnAssign.simple) { |
|---|
| 403 | n/a | PyErr_SetString(PyExc_TypeError, |
|---|
| 404 | n/a | "AnnAssign with simple non-Name target"); |
|---|
| 405 | n/a | return 0; |
|---|
| 406 | n/a | } |
|---|
| 407 | n/a | return validate_expr(stmt->v.AnnAssign.target, Store) && |
|---|
| 408 | n/a | (!stmt->v.AnnAssign.value || |
|---|
| 409 | n/a | validate_expr(stmt->v.AnnAssign.value, Load)) && |
|---|
| 410 | n/a | validate_expr(stmt->v.AnnAssign.annotation, Load); |
|---|
| 411 | n/a | case For_kind: |
|---|
| 412 | n/a | return validate_expr(stmt->v.For.target, Store) && |
|---|
| 413 | n/a | validate_expr(stmt->v.For.iter, Load) && |
|---|
| 414 | n/a | validate_body(stmt->v.For.body, "For") && |
|---|
| 415 | n/a | validate_stmts(stmt->v.For.orelse); |
|---|
| 416 | n/a | case AsyncFor_kind: |
|---|
| 417 | n/a | return validate_expr(stmt->v.AsyncFor.target, Store) && |
|---|
| 418 | n/a | validate_expr(stmt->v.AsyncFor.iter, Load) && |
|---|
| 419 | n/a | validate_body(stmt->v.AsyncFor.body, "AsyncFor") && |
|---|
| 420 | n/a | validate_stmts(stmt->v.AsyncFor.orelse); |
|---|
| 421 | n/a | case While_kind: |
|---|
| 422 | n/a | return validate_expr(stmt->v.While.test, Load) && |
|---|
| 423 | n/a | validate_body(stmt->v.While.body, "While") && |
|---|
| 424 | n/a | validate_stmts(stmt->v.While.orelse); |
|---|
| 425 | n/a | case If_kind: |
|---|
| 426 | n/a | return validate_expr(stmt->v.If.test, Load) && |
|---|
| 427 | n/a | validate_body(stmt->v.If.body, "If") && |
|---|
| 428 | n/a | validate_stmts(stmt->v.If.orelse); |
|---|
| 429 | n/a | case With_kind: |
|---|
| 430 | n/a | if (!validate_nonempty_seq(stmt->v.With.items, "items", "With")) |
|---|
| 431 | n/a | return 0; |
|---|
| 432 | n/a | for (i = 0; i < asdl_seq_LEN(stmt->v.With.items); i++) { |
|---|
| 433 | n/a | withitem_ty item = asdl_seq_GET(stmt->v.With.items, i); |
|---|
| 434 | n/a | if (!validate_expr(item->context_expr, Load) || |
|---|
| 435 | n/a | (item->optional_vars && !validate_expr(item->optional_vars, Store))) |
|---|
| 436 | n/a | return 0; |
|---|
| 437 | n/a | } |
|---|
| 438 | n/a | return validate_body(stmt->v.With.body, "With"); |
|---|
| 439 | n/a | case AsyncWith_kind: |
|---|
| 440 | n/a | if (!validate_nonempty_seq(stmt->v.AsyncWith.items, "items", "AsyncWith")) |
|---|
| 441 | n/a | return 0; |
|---|
| 442 | n/a | for (i = 0; i < asdl_seq_LEN(stmt->v.AsyncWith.items); i++) { |
|---|
| 443 | n/a | withitem_ty item = asdl_seq_GET(stmt->v.AsyncWith.items, i); |
|---|
| 444 | n/a | if (!validate_expr(item->context_expr, Load) || |
|---|
| 445 | n/a | (item->optional_vars && !validate_expr(item->optional_vars, Store))) |
|---|
| 446 | n/a | return 0; |
|---|
| 447 | n/a | } |
|---|
| 448 | n/a | return validate_body(stmt->v.AsyncWith.body, "AsyncWith"); |
|---|
| 449 | n/a | case Raise_kind: |
|---|
| 450 | n/a | if (stmt->v.Raise.exc) { |
|---|
| 451 | n/a | return validate_expr(stmt->v.Raise.exc, Load) && |
|---|
| 452 | n/a | (!stmt->v.Raise.cause || validate_expr(stmt->v.Raise.cause, Load)); |
|---|
| 453 | n/a | } |
|---|
| 454 | n/a | if (stmt->v.Raise.cause) { |
|---|
| 455 | n/a | PyErr_SetString(PyExc_ValueError, "Raise with cause but no exception"); |
|---|
| 456 | n/a | return 0; |
|---|
| 457 | n/a | } |
|---|
| 458 | n/a | return 1; |
|---|
| 459 | n/a | case Try_kind: |
|---|
| 460 | n/a | if (!validate_body(stmt->v.Try.body, "Try")) |
|---|
| 461 | n/a | return 0; |
|---|
| 462 | n/a | if (!asdl_seq_LEN(stmt->v.Try.handlers) && |
|---|
| 463 | n/a | !asdl_seq_LEN(stmt->v.Try.finalbody)) { |
|---|
| 464 | n/a | PyErr_SetString(PyExc_ValueError, "Try has neither except handlers nor finalbody"); |
|---|
| 465 | n/a | return 0; |
|---|
| 466 | n/a | } |
|---|
| 467 | n/a | if (!asdl_seq_LEN(stmt->v.Try.handlers) && |
|---|
| 468 | n/a | asdl_seq_LEN(stmt->v.Try.orelse)) { |
|---|
| 469 | n/a | PyErr_SetString(PyExc_ValueError, "Try has orelse but no except handlers"); |
|---|
| 470 | n/a | return 0; |
|---|
| 471 | n/a | } |
|---|
| 472 | n/a | for (i = 0; i < asdl_seq_LEN(stmt->v.Try.handlers); i++) { |
|---|
| 473 | n/a | excepthandler_ty handler = asdl_seq_GET(stmt->v.Try.handlers, i); |
|---|
| 474 | n/a | if ((handler->v.ExceptHandler.type && |
|---|
| 475 | n/a | !validate_expr(handler->v.ExceptHandler.type, Load)) || |
|---|
| 476 | n/a | !validate_body(handler->v.ExceptHandler.body, "ExceptHandler")) |
|---|
| 477 | n/a | return 0; |
|---|
| 478 | n/a | } |
|---|
| 479 | n/a | return (!asdl_seq_LEN(stmt->v.Try.finalbody) || |
|---|
| 480 | n/a | validate_stmts(stmt->v.Try.finalbody)) && |
|---|
| 481 | n/a | (!asdl_seq_LEN(stmt->v.Try.orelse) || |
|---|
| 482 | n/a | validate_stmts(stmt->v.Try.orelse)); |
|---|
| 483 | n/a | case Assert_kind: |
|---|
| 484 | n/a | return validate_expr(stmt->v.Assert.test, Load) && |
|---|
| 485 | n/a | (!stmt->v.Assert.msg || validate_expr(stmt->v.Assert.msg, Load)); |
|---|
| 486 | n/a | case Import_kind: |
|---|
| 487 | n/a | return validate_nonempty_seq(stmt->v.Import.names, "names", "Import"); |
|---|
| 488 | n/a | case ImportFrom_kind: |
|---|
| 489 | n/a | if (stmt->v.ImportFrom.level < 0) { |
|---|
| 490 | n/a | PyErr_SetString(PyExc_ValueError, "Negative ImportFrom level"); |
|---|
| 491 | n/a | return 0; |
|---|
| 492 | n/a | } |
|---|
| 493 | n/a | return validate_nonempty_seq(stmt->v.ImportFrom.names, "names", "ImportFrom"); |
|---|
| 494 | n/a | case Global_kind: |
|---|
| 495 | n/a | return validate_nonempty_seq(stmt->v.Global.names, "names", "Global"); |
|---|
| 496 | n/a | case Nonlocal_kind: |
|---|
| 497 | n/a | return validate_nonempty_seq(stmt->v.Nonlocal.names, "names", "Nonlocal"); |
|---|
| 498 | n/a | case Expr_kind: |
|---|
| 499 | n/a | return validate_expr(stmt->v.Expr.value, Load); |
|---|
| 500 | n/a | case AsyncFunctionDef_kind: |
|---|
| 501 | n/a | return validate_body(stmt->v.AsyncFunctionDef.body, "AsyncFunctionDef") && |
|---|
| 502 | n/a | validate_arguments(stmt->v.AsyncFunctionDef.args) && |
|---|
| 503 | n/a | validate_exprs(stmt->v.AsyncFunctionDef.decorator_list, Load, 0) && |
|---|
| 504 | n/a | (!stmt->v.AsyncFunctionDef.returns || |
|---|
| 505 | n/a | validate_expr(stmt->v.AsyncFunctionDef.returns, Load)); |
|---|
| 506 | n/a | case Pass_kind: |
|---|
| 507 | n/a | case Break_kind: |
|---|
| 508 | n/a | case Continue_kind: |
|---|
| 509 | n/a | return 1; |
|---|
| 510 | n/a | default: |
|---|
| 511 | n/a | PyErr_SetString(PyExc_SystemError, "unexpected statement"); |
|---|
| 512 | n/a | return 0; |
|---|
| 513 | n/a | } |
|---|
| 514 | n/a | } |
|---|
| 515 | n/a | |
|---|
| 516 | n/a | static int |
|---|
| 517 | n/a | validate_stmts(asdl_seq *seq) |
|---|
| 518 | n/a | { |
|---|
| 519 | n/a | int i; |
|---|
| 520 | n/a | for (i = 0; i < asdl_seq_LEN(seq); i++) { |
|---|
| 521 | n/a | stmt_ty stmt = asdl_seq_GET(seq, i); |
|---|
| 522 | n/a | if (stmt) { |
|---|
| 523 | n/a | if (!validate_stmt(stmt)) |
|---|
| 524 | n/a | return 0; |
|---|
| 525 | n/a | } |
|---|
| 526 | n/a | else { |
|---|
| 527 | n/a | PyErr_SetString(PyExc_ValueError, |
|---|
| 528 | n/a | "None disallowed in statement list"); |
|---|
| 529 | n/a | return 0; |
|---|
| 530 | n/a | } |
|---|
| 531 | n/a | } |
|---|
| 532 | n/a | return 1; |
|---|
| 533 | n/a | } |
|---|
| 534 | n/a | |
|---|
| 535 | n/a | static int |
|---|
| 536 | n/a | validate_exprs(asdl_seq *exprs, expr_context_ty ctx, int null_ok) |
|---|
| 537 | n/a | { |
|---|
| 538 | n/a | int i; |
|---|
| 539 | n/a | for (i = 0; i < asdl_seq_LEN(exprs); i++) { |
|---|
| 540 | n/a | expr_ty expr = asdl_seq_GET(exprs, i); |
|---|
| 541 | n/a | if (expr) { |
|---|
| 542 | n/a | if (!validate_expr(expr, ctx)) |
|---|
| 543 | n/a | return 0; |
|---|
| 544 | n/a | } |
|---|
| 545 | n/a | else if (!null_ok) { |
|---|
| 546 | n/a | PyErr_SetString(PyExc_ValueError, |
|---|
| 547 | n/a | "None disallowed in expression list"); |
|---|
| 548 | n/a | return 0; |
|---|
| 549 | n/a | } |
|---|
| 550 | n/a | |
|---|
| 551 | n/a | } |
|---|
| 552 | n/a | return 1; |
|---|
| 553 | n/a | } |
|---|
| 554 | n/a | |
|---|
| 555 | n/a | int |
|---|
| 556 | n/a | PyAST_Validate(mod_ty mod) |
|---|
| 557 | n/a | { |
|---|
| 558 | n/a | int res = 0; |
|---|
| 559 | n/a | |
|---|
| 560 | n/a | switch (mod->kind) { |
|---|
| 561 | n/a | case Module_kind: |
|---|
| 562 | n/a | res = validate_stmts(mod->v.Module.body); |
|---|
| 563 | n/a | break; |
|---|
| 564 | n/a | case Interactive_kind: |
|---|
| 565 | n/a | res = validate_stmts(mod->v.Interactive.body); |
|---|
| 566 | n/a | break; |
|---|
| 567 | n/a | case Expression_kind: |
|---|
| 568 | n/a | res = validate_expr(mod->v.Expression.body, Load); |
|---|
| 569 | n/a | break; |
|---|
| 570 | n/a | case Suite_kind: |
|---|
| 571 | n/a | PyErr_SetString(PyExc_ValueError, "Suite is not valid in the CPython compiler"); |
|---|
| 572 | n/a | break; |
|---|
| 573 | n/a | default: |
|---|
| 574 | n/a | PyErr_SetString(PyExc_SystemError, "impossible module node"); |
|---|
| 575 | n/a | res = 0; |
|---|
| 576 | n/a | break; |
|---|
| 577 | n/a | } |
|---|
| 578 | n/a | return res; |
|---|
| 579 | n/a | } |
|---|
| 580 | n/a | |
|---|
| 581 | n/a | /* This is done here, so defines like "test" don't interfere with AST use above. */ |
|---|
| 582 | n/a | #include "grammar.h" |
|---|
| 583 | n/a | #include "parsetok.h" |
|---|
| 584 | n/a | #include "graminit.h" |
|---|
| 585 | n/a | |
|---|
| 586 | n/a | /* Data structure used internally */ |
|---|
| 587 | n/a | struct compiling { |
|---|
| 588 | n/a | PyArena *c_arena; /* Arena for allocating memory. */ |
|---|
| 589 | n/a | PyObject *c_filename; /* filename */ |
|---|
| 590 | n/a | PyObject *c_normalize; /* Normalization function from unicodedata. */ |
|---|
| 591 | n/a | PyObject *c_normalize_args; /* Normalization argument tuple. */ |
|---|
| 592 | n/a | }; |
|---|
| 593 | n/a | |
|---|
| 594 | n/a | static asdl_seq *seq_for_testlist(struct compiling *, const node *); |
|---|
| 595 | n/a | static expr_ty ast_for_expr(struct compiling *, const node *); |
|---|
| 596 | n/a | static stmt_ty ast_for_stmt(struct compiling *, const node *); |
|---|
| 597 | n/a | static asdl_seq *ast_for_suite(struct compiling *, const node *); |
|---|
| 598 | n/a | static asdl_seq *ast_for_exprlist(struct compiling *, const node *, |
|---|
| 599 | n/a | expr_context_ty); |
|---|
| 600 | n/a | static expr_ty ast_for_testlist(struct compiling *, const node *); |
|---|
| 601 | n/a | static stmt_ty ast_for_classdef(struct compiling *, const node *, asdl_seq *); |
|---|
| 602 | n/a | |
|---|
| 603 | n/a | static stmt_ty ast_for_with_stmt(struct compiling *, const node *, int); |
|---|
| 604 | n/a | static stmt_ty ast_for_for_stmt(struct compiling *, const node *, int); |
|---|
| 605 | n/a | |
|---|
| 606 | n/a | /* Note different signature for ast_for_call */ |
|---|
| 607 | n/a | static expr_ty ast_for_call(struct compiling *, const node *, expr_ty); |
|---|
| 608 | n/a | |
|---|
| 609 | n/a | static PyObject *parsenumber(struct compiling *, const char *); |
|---|
| 610 | n/a | static expr_ty parsestrplus(struct compiling *, const node *n); |
|---|
| 611 | n/a | |
|---|
| 612 | n/a | #define COMP_GENEXP 0 |
|---|
| 613 | n/a | #define COMP_LISTCOMP 1 |
|---|
| 614 | n/a | #define COMP_SETCOMP 2 |
|---|
| 615 | n/a | |
|---|
| 616 | n/a | static int |
|---|
| 617 | n/a | init_normalization(struct compiling *c) |
|---|
| 618 | n/a | { |
|---|
| 619 | n/a | PyObject *m = PyImport_ImportModuleNoBlock("unicodedata"); |
|---|
| 620 | n/a | if (!m) |
|---|
| 621 | n/a | return 0; |
|---|
| 622 | n/a | c->c_normalize = PyObject_GetAttrString(m, "normalize"); |
|---|
| 623 | n/a | Py_DECREF(m); |
|---|
| 624 | n/a | if (!c->c_normalize) |
|---|
| 625 | n/a | return 0; |
|---|
| 626 | n/a | c->c_normalize_args = Py_BuildValue("(sN)", "NFKC", Py_None); |
|---|
| 627 | n/a | if (!c->c_normalize_args) { |
|---|
| 628 | n/a | Py_CLEAR(c->c_normalize); |
|---|
| 629 | n/a | return 0; |
|---|
| 630 | n/a | } |
|---|
| 631 | n/a | PyTuple_SET_ITEM(c->c_normalize_args, 1, NULL); |
|---|
| 632 | n/a | return 1; |
|---|
| 633 | n/a | } |
|---|
| 634 | n/a | |
|---|
| 635 | n/a | static identifier |
|---|
| 636 | n/a | new_identifier(const char *n, struct compiling *c) |
|---|
| 637 | n/a | { |
|---|
| 638 | n/a | PyObject *id = PyUnicode_DecodeUTF8(n, strlen(n), NULL); |
|---|
| 639 | n/a | if (!id) |
|---|
| 640 | n/a | return NULL; |
|---|
| 641 | n/a | /* PyUnicode_DecodeUTF8 should always return a ready string. */ |
|---|
| 642 | n/a | assert(PyUnicode_IS_READY(id)); |
|---|
| 643 | n/a | /* Check whether there are non-ASCII characters in the |
|---|
| 644 | n/a | identifier; if so, normalize to NFKC. */ |
|---|
| 645 | n/a | if (!PyUnicode_IS_ASCII(id)) { |
|---|
| 646 | n/a | PyObject *id2; |
|---|
| 647 | n/a | if (!c->c_normalize && !init_normalization(c)) { |
|---|
| 648 | n/a | Py_DECREF(id); |
|---|
| 649 | n/a | return NULL; |
|---|
| 650 | n/a | } |
|---|
| 651 | n/a | PyTuple_SET_ITEM(c->c_normalize_args, 1, id); |
|---|
| 652 | n/a | id2 = PyObject_Call(c->c_normalize, c->c_normalize_args, NULL); |
|---|
| 653 | n/a | Py_DECREF(id); |
|---|
| 654 | n/a | if (!id2) |
|---|
| 655 | n/a | return NULL; |
|---|
| 656 | n/a | id = id2; |
|---|
| 657 | n/a | } |
|---|
| 658 | n/a | PyUnicode_InternInPlace(&id); |
|---|
| 659 | n/a | if (PyArena_AddPyObject(c->c_arena, id) < 0) { |
|---|
| 660 | n/a | Py_DECREF(id); |
|---|
| 661 | n/a | return NULL; |
|---|
| 662 | n/a | } |
|---|
| 663 | n/a | return id; |
|---|
| 664 | n/a | } |
|---|
| 665 | n/a | |
|---|
| 666 | n/a | #define NEW_IDENTIFIER(n) new_identifier(STR(n), c) |
|---|
| 667 | n/a | |
|---|
| 668 | n/a | static int |
|---|
| 669 | n/a | ast_error(struct compiling *c, const node *n, const char *errmsg) |
|---|
| 670 | n/a | { |
|---|
| 671 | n/a | PyObject *value, *errstr, *loc, *tmp; |
|---|
| 672 | n/a | |
|---|
| 673 | n/a | loc = PyErr_ProgramTextObject(c->c_filename, LINENO(n)); |
|---|
| 674 | n/a | if (!loc) { |
|---|
| 675 | n/a | Py_INCREF(Py_None); |
|---|
| 676 | n/a | loc = Py_None; |
|---|
| 677 | n/a | } |
|---|
| 678 | n/a | tmp = Py_BuildValue("(OiiN)", c->c_filename, LINENO(n), n->n_col_offset, loc); |
|---|
| 679 | n/a | if (!tmp) |
|---|
| 680 | n/a | return 0; |
|---|
| 681 | n/a | errstr = PyUnicode_FromString(errmsg); |
|---|
| 682 | n/a | if (!errstr) { |
|---|
| 683 | n/a | Py_DECREF(tmp); |
|---|
| 684 | n/a | return 0; |
|---|
| 685 | n/a | } |
|---|
| 686 | n/a | value = PyTuple_Pack(2, errstr, tmp); |
|---|
| 687 | n/a | Py_DECREF(errstr); |
|---|
| 688 | n/a | Py_DECREF(tmp); |
|---|
| 689 | n/a | if (value) { |
|---|
| 690 | n/a | PyErr_SetObject(PyExc_SyntaxError, value); |
|---|
| 691 | n/a | Py_DECREF(value); |
|---|
| 692 | n/a | } |
|---|
| 693 | n/a | return 0; |
|---|
| 694 | n/a | } |
|---|
| 695 | n/a | |
|---|
| 696 | n/a | /* num_stmts() returns number of contained statements. |
|---|
| 697 | n/a | |
|---|
| 698 | n/a | Use this routine to determine how big a sequence is needed for |
|---|
| 699 | n/a | the statements in a parse tree. Its raison d'etre is this bit of |
|---|
| 700 | n/a | grammar: |
|---|
| 701 | n/a | |
|---|
| 702 | n/a | stmt: simple_stmt | compound_stmt |
|---|
| 703 | n/a | simple_stmt: small_stmt (';' small_stmt)* [';'] NEWLINE |
|---|
| 704 | n/a | |
|---|
| 705 | n/a | A simple_stmt can contain multiple small_stmt elements joined |
|---|
| 706 | n/a | by semicolons. If the arg is a simple_stmt, the number of |
|---|
| 707 | n/a | small_stmt elements is returned. |
|---|
| 708 | n/a | */ |
|---|
| 709 | n/a | |
|---|
| 710 | n/a | static int |
|---|
| 711 | n/a | num_stmts(const node *n) |
|---|
| 712 | n/a | { |
|---|
| 713 | n/a | int i, l; |
|---|
| 714 | n/a | node *ch; |
|---|
| 715 | n/a | |
|---|
| 716 | n/a | switch (TYPE(n)) { |
|---|
| 717 | n/a | case single_input: |
|---|
| 718 | n/a | if (TYPE(CHILD(n, 0)) == NEWLINE) |
|---|
| 719 | n/a | return 0; |
|---|
| 720 | n/a | else |
|---|
| 721 | n/a | return num_stmts(CHILD(n, 0)); |
|---|
| 722 | n/a | case file_input: |
|---|
| 723 | n/a | l = 0; |
|---|
| 724 | n/a | for (i = 0; i < NCH(n); i++) { |
|---|
| 725 | n/a | ch = CHILD(n, i); |
|---|
| 726 | n/a | if (TYPE(ch) == stmt) |
|---|
| 727 | n/a | l += num_stmts(ch); |
|---|
| 728 | n/a | } |
|---|
| 729 | n/a | return l; |
|---|
| 730 | n/a | case stmt: |
|---|
| 731 | n/a | return num_stmts(CHILD(n, 0)); |
|---|
| 732 | n/a | case compound_stmt: |
|---|
| 733 | n/a | return 1; |
|---|
| 734 | n/a | case simple_stmt: |
|---|
| 735 | n/a | return NCH(n) / 2; /* Divide by 2 to remove count of semi-colons */ |
|---|
| 736 | n/a | case suite: |
|---|
| 737 | n/a | if (NCH(n) == 1) |
|---|
| 738 | n/a | return num_stmts(CHILD(n, 0)); |
|---|
| 739 | n/a | else { |
|---|
| 740 | n/a | l = 0; |
|---|
| 741 | n/a | for (i = 2; i < (NCH(n) - 1); i++) |
|---|
| 742 | n/a | l += num_stmts(CHILD(n, i)); |
|---|
| 743 | n/a | return l; |
|---|
| 744 | n/a | } |
|---|
| 745 | n/a | default: { |
|---|
| 746 | n/a | char buf[128]; |
|---|
| 747 | n/a | |
|---|
| 748 | n/a | sprintf(buf, "Non-statement found: %d %d", |
|---|
| 749 | n/a | TYPE(n), NCH(n)); |
|---|
| 750 | n/a | Py_FatalError(buf); |
|---|
| 751 | n/a | } |
|---|
| 752 | n/a | } |
|---|
| 753 | n/a | assert(0); |
|---|
| 754 | n/a | return 0; |
|---|
| 755 | n/a | } |
|---|
| 756 | n/a | |
|---|
| 757 | n/a | /* Transform the CST rooted at node * to the appropriate AST |
|---|
| 758 | n/a | */ |
|---|
| 759 | n/a | |
|---|
| 760 | n/a | mod_ty |
|---|
| 761 | n/a | PyAST_FromNodeObject(const node *n, PyCompilerFlags *flags, |
|---|
| 762 | n/a | PyObject *filename, PyArena *arena) |
|---|
| 763 | n/a | { |
|---|
| 764 | n/a | int i, j, k, num; |
|---|
| 765 | n/a | asdl_seq *stmts = NULL; |
|---|
| 766 | n/a | stmt_ty s; |
|---|
| 767 | n/a | node *ch; |
|---|
| 768 | n/a | struct compiling c; |
|---|
| 769 | n/a | mod_ty res = NULL; |
|---|
| 770 | n/a | |
|---|
| 771 | n/a | c.c_arena = arena; |
|---|
| 772 | n/a | /* borrowed reference */ |
|---|
| 773 | n/a | c.c_filename = filename; |
|---|
| 774 | n/a | c.c_normalize = NULL; |
|---|
| 775 | n/a | c.c_normalize_args = NULL; |
|---|
| 776 | n/a | |
|---|
| 777 | n/a | if (TYPE(n) == encoding_decl) |
|---|
| 778 | n/a | n = CHILD(n, 0); |
|---|
| 779 | n/a | |
|---|
| 780 | n/a | k = 0; |
|---|
| 781 | n/a | switch (TYPE(n)) { |
|---|
| 782 | n/a | case file_input: |
|---|
| 783 | n/a | stmts = _Py_asdl_seq_new(num_stmts(n), arena); |
|---|
| 784 | n/a | if (!stmts) |
|---|
| 785 | n/a | goto out; |
|---|
| 786 | n/a | for (i = 0; i < NCH(n) - 1; i++) { |
|---|
| 787 | n/a | ch = CHILD(n, i); |
|---|
| 788 | n/a | if (TYPE(ch) == NEWLINE) |
|---|
| 789 | n/a | continue; |
|---|
| 790 | n/a | REQ(ch, stmt); |
|---|
| 791 | n/a | num = num_stmts(ch); |
|---|
| 792 | n/a | if (num == 1) { |
|---|
| 793 | n/a | s = ast_for_stmt(&c, ch); |
|---|
| 794 | n/a | if (!s) |
|---|
| 795 | n/a | goto out; |
|---|
| 796 | n/a | asdl_seq_SET(stmts, k++, s); |
|---|
| 797 | n/a | } |
|---|
| 798 | n/a | else { |
|---|
| 799 | n/a | ch = CHILD(ch, 0); |
|---|
| 800 | n/a | REQ(ch, simple_stmt); |
|---|
| 801 | n/a | for (j = 0; j < num; j++) { |
|---|
| 802 | n/a | s = ast_for_stmt(&c, CHILD(ch, j * 2)); |
|---|
| 803 | n/a | if (!s) |
|---|
| 804 | n/a | goto out; |
|---|
| 805 | n/a | asdl_seq_SET(stmts, k++, s); |
|---|
| 806 | n/a | } |
|---|
| 807 | n/a | } |
|---|
| 808 | n/a | } |
|---|
| 809 | n/a | res = Module(stmts, arena); |
|---|
| 810 | n/a | break; |
|---|
| 811 | n/a | case eval_input: { |
|---|
| 812 | n/a | expr_ty testlist_ast; |
|---|
| 813 | n/a | |
|---|
| 814 | n/a | /* XXX Why not comp_for here? */ |
|---|
| 815 | n/a | testlist_ast = ast_for_testlist(&c, CHILD(n, 0)); |
|---|
| 816 | n/a | if (!testlist_ast) |
|---|
| 817 | n/a | goto out; |
|---|
| 818 | n/a | res = Expression(testlist_ast, arena); |
|---|
| 819 | n/a | break; |
|---|
| 820 | n/a | } |
|---|
| 821 | n/a | case single_input: |
|---|
| 822 | n/a | if (TYPE(CHILD(n, 0)) == NEWLINE) { |
|---|
| 823 | n/a | stmts = _Py_asdl_seq_new(1, arena); |
|---|
| 824 | n/a | if (!stmts) |
|---|
| 825 | n/a | goto out; |
|---|
| 826 | n/a | asdl_seq_SET(stmts, 0, Pass(n->n_lineno, n->n_col_offset, |
|---|
| 827 | n/a | arena)); |
|---|
| 828 | n/a | if (!asdl_seq_GET(stmts, 0)) |
|---|
| 829 | n/a | goto out; |
|---|
| 830 | n/a | res = Interactive(stmts, arena); |
|---|
| 831 | n/a | } |
|---|
| 832 | n/a | else { |
|---|
| 833 | n/a | n = CHILD(n, 0); |
|---|
| 834 | n/a | num = num_stmts(n); |
|---|
| 835 | n/a | stmts = _Py_asdl_seq_new(num, arena); |
|---|
| 836 | n/a | if (!stmts) |
|---|
| 837 | n/a | goto out; |
|---|
| 838 | n/a | if (num == 1) { |
|---|
| 839 | n/a | s = ast_for_stmt(&c, n); |
|---|
| 840 | n/a | if (!s) |
|---|
| 841 | n/a | goto out; |
|---|
| 842 | n/a | asdl_seq_SET(stmts, 0, s); |
|---|
| 843 | n/a | } |
|---|
| 844 | n/a | else { |
|---|
| 845 | n/a | /* Only a simple_stmt can contain multiple statements. */ |
|---|
| 846 | n/a | REQ(n, simple_stmt); |
|---|
| 847 | n/a | for (i = 0; i < NCH(n); i += 2) { |
|---|
| 848 | n/a | if (TYPE(CHILD(n, i)) == NEWLINE) |
|---|
| 849 | n/a | break; |
|---|
| 850 | n/a | s = ast_for_stmt(&c, CHILD(n, i)); |
|---|
| 851 | n/a | if (!s) |
|---|
| 852 | n/a | goto out; |
|---|
| 853 | n/a | asdl_seq_SET(stmts, i / 2, s); |
|---|
| 854 | n/a | } |
|---|
| 855 | n/a | } |
|---|
| 856 | n/a | |
|---|
| 857 | n/a | res = Interactive(stmts, arena); |
|---|
| 858 | n/a | } |
|---|
| 859 | n/a | break; |
|---|
| 860 | n/a | default: |
|---|
| 861 | n/a | PyErr_Format(PyExc_SystemError, |
|---|
| 862 | n/a | "invalid node %d for PyAST_FromNode", TYPE(n)); |
|---|
| 863 | n/a | goto out; |
|---|
| 864 | n/a | } |
|---|
| 865 | n/a | out: |
|---|
| 866 | n/a | if (c.c_normalize) { |
|---|
| 867 | n/a | Py_DECREF(c.c_normalize); |
|---|
| 868 | n/a | PyTuple_SET_ITEM(c.c_normalize_args, 1, NULL); |
|---|
| 869 | n/a | Py_DECREF(c.c_normalize_args); |
|---|
| 870 | n/a | } |
|---|
| 871 | n/a | return res; |
|---|
| 872 | n/a | } |
|---|
| 873 | n/a | |
|---|
| 874 | n/a | mod_ty |
|---|
| 875 | n/a | PyAST_FromNode(const node *n, PyCompilerFlags *flags, const char *filename_str, |
|---|
| 876 | n/a | PyArena *arena) |
|---|
| 877 | n/a | { |
|---|
| 878 | n/a | mod_ty mod; |
|---|
| 879 | n/a | PyObject *filename; |
|---|
| 880 | n/a | filename = PyUnicode_DecodeFSDefault(filename_str); |
|---|
| 881 | n/a | if (filename == NULL) |
|---|
| 882 | n/a | return NULL; |
|---|
| 883 | n/a | mod = PyAST_FromNodeObject(n, flags, filename, arena); |
|---|
| 884 | n/a | Py_DECREF(filename); |
|---|
| 885 | n/a | return mod; |
|---|
| 886 | n/a | |
|---|
| 887 | n/a | } |
|---|
| 888 | n/a | |
|---|
| 889 | n/a | /* Return the AST repr. of the operator represented as syntax (|, ^, etc.) |
|---|
| 890 | n/a | */ |
|---|
| 891 | n/a | |
|---|
| 892 | n/a | static operator_ty |
|---|
| 893 | n/a | get_operator(const node *n) |
|---|
| 894 | n/a | { |
|---|
| 895 | n/a | switch (TYPE(n)) { |
|---|
| 896 | n/a | case VBAR: |
|---|
| 897 | n/a | return BitOr; |
|---|
| 898 | n/a | case CIRCUMFLEX: |
|---|
| 899 | n/a | return BitXor; |
|---|
| 900 | n/a | case AMPER: |
|---|
| 901 | n/a | return BitAnd; |
|---|
| 902 | n/a | case LEFTSHIFT: |
|---|
| 903 | n/a | return LShift; |
|---|
| 904 | n/a | case RIGHTSHIFT: |
|---|
| 905 | n/a | return RShift; |
|---|
| 906 | n/a | case PLUS: |
|---|
| 907 | n/a | return Add; |
|---|
| 908 | n/a | case MINUS: |
|---|
| 909 | n/a | return Sub; |
|---|
| 910 | n/a | case STAR: |
|---|
| 911 | n/a | return Mult; |
|---|
| 912 | n/a | case AT: |
|---|
| 913 | n/a | return MatMult; |
|---|
| 914 | n/a | case SLASH: |
|---|
| 915 | n/a | return Div; |
|---|
| 916 | n/a | case DOUBLESLASH: |
|---|
| 917 | n/a | return FloorDiv; |
|---|
| 918 | n/a | case PERCENT: |
|---|
| 919 | n/a | return Mod; |
|---|
| 920 | n/a | default: |
|---|
| 921 | n/a | return (operator_ty)0; |
|---|
| 922 | n/a | } |
|---|
| 923 | n/a | } |
|---|
| 924 | n/a | |
|---|
| 925 | n/a | static const char * const FORBIDDEN[] = { |
|---|
| 926 | n/a | "None", |
|---|
| 927 | n/a | "True", |
|---|
| 928 | n/a | "False", |
|---|
| 929 | n/a | NULL, |
|---|
| 930 | n/a | }; |
|---|
| 931 | n/a | |
|---|
| 932 | n/a | static int |
|---|
| 933 | n/a | forbidden_name(struct compiling *c, identifier name, const node *n, |
|---|
| 934 | n/a | int full_checks) |
|---|
| 935 | n/a | { |
|---|
| 936 | n/a | assert(PyUnicode_Check(name)); |
|---|
| 937 | n/a | if (_PyUnicode_EqualToASCIIString(name, "__debug__")) { |
|---|
| 938 | n/a | ast_error(c, n, "assignment to keyword"); |
|---|
| 939 | n/a | return 1; |
|---|
| 940 | n/a | } |
|---|
| 941 | n/a | if (_PyUnicode_EqualToASCIIString(name, "async") || |
|---|
| 942 | n/a | _PyUnicode_EqualToASCIIString(name, "await")) |
|---|
| 943 | n/a | { |
|---|
| 944 | n/a | PyObject *message = PyUnicode_FromString( |
|---|
| 945 | n/a | "'async' and 'await' will become reserved keywords" |
|---|
| 946 | n/a | " in Python 3.7"); |
|---|
| 947 | n/a | int ret; |
|---|
| 948 | n/a | if (message == NULL) { |
|---|
| 949 | n/a | return 1; |
|---|
| 950 | n/a | } |
|---|
| 951 | n/a | ret = PyErr_WarnExplicitObject( |
|---|
| 952 | n/a | PyExc_DeprecationWarning, |
|---|
| 953 | n/a | message, |
|---|
| 954 | n/a | c->c_filename, |
|---|
| 955 | n/a | LINENO(n), |
|---|
| 956 | n/a | NULL, |
|---|
| 957 | n/a | NULL); |
|---|
| 958 | n/a | Py_DECREF(message); |
|---|
| 959 | n/a | if (ret < 0) { |
|---|
| 960 | n/a | return 1; |
|---|
| 961 | n/a | } |
|---|
| 962 | n/a | } |
|---|
| 963 | n/a | if (full_checks) { |
|---|
| 964 | n/a | const char * const *p; |
|---|
| 965 | n/a | for (p = FORBIDDEN; *p; p++) { |
|---|
| 966 | n/a | if (_PyUnicode_EqualToASCIIString(name, *p)) { |
|---|
| 967 | n/a | ast_error(c, n, "assignment to keyword"); |
|---|
| 968 | n/a | return 1; |
|---|
| 969 | n/a | } |
|---|
| 970 | n/a | } |
|---|
| 971 | n/a | } |
|---|
| 972 | n/a | return 0; |
|---|
| 973 | n/a | } |
|---|
| 974 | n/a | |
|---|
| 975 | n/a | /* Set the context ctx for expr_ty e, recursively traversing e. |
|---|
| 976 | n/a | |
|---|
| 977 | n/a | Only sets context for expr kinds that "can appear in assignment context" |
|---|
| 978 | n/a | (according to ../Parser/Python.asdl). For other expr kinds, it sets |
|---|
| 979 | n/a | an appropriate syntax error and returns false. |
|---|
| 980 | n/a | */ |
|---|
| 981 | n/a | |
|---|
| 982 | n/a | static int |
|---|
| 983 | n/a | set_context(struct compiling *c, expr_ty e, expr_context_ty ctx, const node *n) |
|---|
| 984 | n/a | { |
|---|
| 985 | n/a | asdl_seq *s = NULL; |
|---|
| 986 | n/a | /* If a particular expression type can't be used for assign / delete, |
|---|
| 987 | n/a | set expr_name to its name and an error message will be generated. |
|---|
| 988 | n/a | */ |
|---|
| 989 | n/a | const char* expr_name = NULL; |
|---|
| 990 | n/a | |
|---|
| 991 | n/a | /* The ast defines augmented store and load contexts, but the |
|---|
| 992 | n/a | implementation here doesn't actually use them. The code may be |
|---|
| 993 | n/a | a little more complex than necessary as a result. It also means |
|---|
| 994 | n/a | that expressions in an augmented assignment have a Store context. |
|---|
| 995 | n/a | Consider restructuring so that augmented assignment uses |
|---|
| 996 | n/a | set_context(), too. |
|---|
| 997 | n/a | */ |
|---|
| 998 | n/a | assert(ctx != AugStore && ctx != AugLoad); |
|---|
| 999 | n/a | |
|---|
| 1000 | n/a | switch (e->kind) { |
|---|
| 1001 | n/a | case Attribute_kind: |
|---|
| 1002 | n/a | e->v.Attribute.ctx = ctx; |
|---|
| 1003 | n/a | if (ctx == Store && forbidden_name(c, e->v.Attribute.attr, n, 1)) |
|---|
| 1004 | n/a | return 0; |
|---|
| 1005 | n/a | break; |
|---|
| 1006 | n/a | case Subscript_kind: |
|---|
| 1007 | n/a | e->v.Subscript.ctx = ctx; |
|---|
| 1008 | n/a | break; |
|---|
| 1009 | n/a | case Starred_kind: |
|---|
| 1010 | n/a | e->v.Starred.ctx = ctx; |
|---|
| 1011 | n/a | if (!set_context(c, e->v.Starred.value, ctx, n)) |
|---|
| 1012 | n/a | return 0; |
|---|
| 1013 | n/a | break; |
|---|
| 1014 | n/a | case Name_kind: |
|---|
| 1015 | n/a | if (ctx == Store) { |
|---|
| 1016 | n/a | if (forbidden_name(c, e->v.Name.id, n, 0)) |
|---|
| 1017 | n/a | return 0; /* forbidden_name() calls ast_error() */ |
|---|
| 1018 | n/a | } |
|---|
| 1019 | n/a | e->v.Name.ctx = ctx; |
|---|
| 1020 | n/a | break; |
|---|
| 1021 | n/a | case List_kind: |
|---|
| 1022 | n/a | e->v.List.ctx = ctx; |
|---|
| 1023 | n/a | s = e->v.List.elts; |
|---|
| 1024 | n/a | break; |
|---|
| 1025 | n/a | case Tuple_kind: |
|---|
| 1026 | n/a | e->v.Tuple.ctx = ctx; |
|---|
| 1027 | n/a | s = e->v.Tuple.elts; |
|---|
| 1028 | n/a | break; |
|---|
| 1029 | n/a | case Lambda_kind: |
|---|
| 1030 | n/a | expr_name = "lambda"; |
|---|
| 1031 | n/a | break; |
|---|
| 1032 | n/a | case Call_kind: |
|---|
| 1033 | n/a | expr_name = "function call"; |
|---|
| 1034 | n/a | break; |
|---|
| 1035 | n/a | case BoolOp_kind: |
|---|
| 1036 | n/a | case BinOp_kind: |
|---|
| 1037 | n/a | case UnaryOp_kind: |
|---|
| 1038 | n/a | expr_name = "operator"; |
|---|
| 1039 | n/a | break; |
|---|
| 1040 | n/a | case GeneratorExp_kind: |
|---|
| 1041 | n/a | expr_name = "generator expression"; |
|---|
| 1042 | n/a | break; |
|---|
| 1043 | n/a | case Yield_kind: |
|---|
| 1044 | n/a | case YieldFrom_kind: |
|---|
| 1045 | n/a | expr_name = "yield expression"; |
|---|
| 1046 | n/a | break; |
|---|
| 1047 | n/a | case Await_kind: |
|---|
| 1048 | n/a | expr_name = "await expression"; |
|---|
| 1049 | n/a | break; |
|---|
| 1050 | n/a | case ListComp_kind: |
|---|
| 1051 | n/a | expr_name = "list comprehension"; |
|---|
| 1052 | n/a | break; |
|---|
| 1053 | n/a | case SetComp_kind: |
|---|
| 1054 | n/a | expr_name = "set comprehension"; |
|---|
| 1055 | n/a | break; |
|---|
| 1056 | n/a | case DictComp_kind: |
|---|
| 1057 | n/a | expr_name = "dict comprehension"; |
|---|
| 1058 | n/a | break; |
|---|
| 1059 | n/a | case Dict_kind: |
|---|
| 1060 | n/a | case Set_kind: |
|---|
| 1061 | n/a | case Num_kind: |
|---|
| 1062 | n/a | case Str_kind: |
|---|
| 1063 | n/a | case Bytes_kind: |
|---|
| 1064 | n/a | case JoinedStr_kind: |
|---|
| 1065 | n/a | case FormattedValue_kind: |
|---|
| 1066 | n/a | expr_name = "literal"; |
|---|
| 1067 | n/a | break; |
|---|
| 1068 | n/a | case NameConstant_kind: |
|---|
| 1069 | n/a | expr_name = "keyword"; |
|---|
| 1070 | n/a | break; |
|---|
| 1071 | n/a | case Ellipsis_kind: |
|---|
| 1072 | n/a | expr_name = "Ellipsis"; |
|---|
| 1073 | n/a | break; |
|---|
| 1074 | n/a | case Compare_kind: |
|---|
| 1075 | n/a | expr_name = "comparison"; |
|---|
| 1076 | n/a | break; |
|---|
| 1077 | n/a | case IfExp_kind: |
|---|
| 1078 | n/a | expr_name = "conditional expression"; |
|---|
| 1079 | n/a | break; |
|---|
| 1080 | n/a | default: |
|---|
| 1081 | n/a | PyErr_Format(PyExc_SystemError, |
|---|
| 1082 | n/a | "unexpected expression in assignment %d (line %d)", |
|---|
| 1083 | n/a | e->kind, e->lineno); |
|---|
| 1084 | n/a | return 0; |
|---|
| 1085 | n/a | } |
|---|
| 1086 | n/a | /* Check for error string set by switch */ |
|---|
| 1087 | n/a | if (expr_name) { |
|---|
| 1088 | n/a | char buf[300]; |
|---|
| 1089 | n/a | PyOS_snprintf(buf, sizeof(buf), |
|---|
| 1090 | n/a | "can't %s %s", |
|---|
| 1091 | n/a | ctx == Store ? "assign to" : "delete", |
|---|
| 1092 | n/a | expr_name); |
|---|
| 1093 | n/a | return ast_error(c, n, buf); |
|---|
| 1094 | n/a | } |
|---|
| 1095 | n/a | |
|---|
| 1096 | n/a | /* If the LHS is a list or tuple, we need to set the assignment |
|---|
| 1097 | n/a | context for all the contained elements. |
|---|
| 1098 | n/a | */ |
|---|
| 1099 | n/a | if (s) { |
|---|
| 1100 | n/a | int i; |
|---|
| 1101 | n/a | |
|---|
| 1102 | n/a | for (i = 0; i < asdl_seq_LEN(s); i++) { |
|---|
| 1103 | n/a | if (!set_context(c, (expr_ty)asdl_seq_GET(s, i), ctx, n)) |
|---|
| 1104 | n/a | return 0; |
|---|
| 1105 | n/a | } |
|---|
| 1106 | n/a | } |
|---|
| 1107 | n/a | return 1; |
|---|
| 1108 | n/a | } |
|---|
| 1109 | n/a | |
|---|
| 1110 | n/a | static operator_ty |
|---|
| 1111 | n/a | ast_for_augassign(struct compiling *c, const node *n) |
|---|
| 1112 | n/a | { |
|---|
| 1113 | n/a | REQ(n, augassign); |
|---|
| 1114 | n/a | n = CHILD(n, 0); |
|---|
| 1115 | n/a | switch (STR(n)[0]) { |
|---|
| 1116 | n/a | case '+': |
|---|
| 1117 | n/a | return Add; |
|---|
| 1118 | n/a | case '-': |
|---|
| 1119 | n/a | return Sub; |
|---|
| 1120 | n/a | case '/': |
|---|
| 1121 | n/a | if (STR(n)[1] == '/') |
|---|
| 1122 | n/a | return FloorDiv; |
|---|
| 1123 | n/a | else |
|---|
| 1124 | n/a | return Div; |
|---|
| 1125 | n/a | case '%': |
|---|
| 1126 | n/a | return Mod; |
|---|
| 1127 | n/a | case '<': |
|---|
| 1128 | n/a | return LShift; |
|---|
| 1129 | n/a | case '>': |
|---|
| 1130 | n/a | return RShift; |
|---|
| 1131 | n/a | case '&': |
|---|
| 1132 | n/a | return BitAnd; |
|---|
| 1133 | n/a | case '^': |
|---|
| 1134 | n/a | return BitXor; |
|---|
| 1135 | n/a | case '|': |
|---|
| 1136 | n/a | return BitOr; |
|---|
| 1137 | n/a | case '*': |
|---|
| 1138 | n/a | if (STR(n)[1] == '*') |
|---|
| 1139 | n/a | return Pow; |
|---|
| 1140 | n/a | else |
|---|
| 1141 | n/a | return Mult; |
|---|
| 1142 | n/a | case '@': |
|---|
| 1143 | n/a | return MatMult; |
|---|
| 1144 | n/a | default: |
|---|
| 1145 | n/a | PyErr_Format(PyExc_SystemError, "invalid augassign: %s", STR(n)); |
|---|
| 1146 | n/a | return (operator_ty)0; |
|---|
| 1147 | n/a | } |
|---|
| 1148 | n/a | } |
|---|
| 1149 | n/a | |
|---|
| 1150 | n/a | static cmpop_ty |
|---|
| 1151 | n/a | ast_for_comp_op(struct compiling *c, const node *n) |
|---|
| 1152 | n/a | { |
|---|
| 1153 | n/a | /* comp_op: '<'|'>'|'=='|'>='|'<='|'!='|'in'|'not' 'in'|'is' |
|---|
| 1154 | n/a | |'is' 'not' |
|---|
| 1155 | n/a | */ |
|---|
| 1156 | n/a | REQ(n, comp_op); |
|---|
| 1157 | n/a | if (NCH(n) == 1) { |
|---|
| 1158 | n/a | n = CHILD(n, 0); |
|---|
| 1159 | n/a | switch (TYPE(n)) { |
|---|
| 1160 | n/a | case LESS: |
|---|
| 1161 | n/a | return Lt; |
|---|
| 1162 | n/a | case GREATER: |
|---|
| 1163 | n/a | return Gt; |
|---|
| 1164 | n/a | case EQEQUAL: /* == */ |
|---|
| 1165 | n/a | return Eq; |
|---|
| 1166 | n/a | case LESSEQUAL: |
|---|
| 1167 | n/a | return LtE; |
|---|
| 1168 | n/a | case GREATEREQUAL: |
|---|
| 1169 | n/a | return GtE; |
|---|
| 1170 | n/a | case NOTEQUAL: |
|---|
| 1171 | n/a | return NotEq; |
|---|
| 1172 | n/a | case NAME: |
|---|
| 1173 | n/a | if (strcmp(STR(n), "in") == 0) |
|---|
| 1174 | n/a | return In; |
|---|
| 1175 | n/a | if (strcmp(STR(n), "is") == 0) |
|---|
| 1176 | n/a | return Is; |
|---|
| 1177 | n/a | default: |
|---|
| 1178 | n/a | PyErr_Format(PyExc_SystemError, "invalid comp_op: %s", |
|---|
| 1179 | n/a | STR(n)); |
|---|
| 1180 | n/a | return (cmpop_ty)0; |
|---|
| 1181 | n/a | } |
|---|
| 1182 | n/a | } |
|---|
| 1183 | n/a | else if (NCH(n) == 2) { |
|---|
| 1184 | n/a | /* handle "not in" and "is not" */ |
|---|
| 1185 | n/a | switch (TYPE(CHILD(n, 0))) { |
|---|
| 1186 | n/a | case NAME: |
|---|
| 1187 | n/a | if (strcmp(STR(CHILD(n, 1)), "in") == 0) |
|---|
| 1188 | n/a | return NotIn; |
|---|
| 1189 | n/a | if (strcmp(STR(CHILD(n, 0)), "is") == 0) |
|---|
| 1190 | n/a | return IsNot; |
|---|
| 1191 | n/a | default: |
|---|
| 1192 | n/a | PyErr_Format(PyExc_SystemError, "invalid comp_op: %s %s", |
|---|
| 1193 | n/a | STR(CHILD(n, 0)), STR(CHILD(n, 1))); |
|---|
| 1194 | n/a | return (cmpop_ty)0; |
|---|
| 1195 | n/a | } |
|---|
| 1196 | n/a | } |
|---|
| 1197 | n/a | PyErr_Format(PyExc_SystemError, "invalid comp_op: has %d children", |
|---|
| 1198 | n/a | NCH(n)); |
|---|
| 1199 | n/a | return (cmpop_ty)0; |
|---|
| 1200 | n/a | } |
|---|
| 1201 | n/a | |
|---|
| 1202 | n/a | static asdl_seq * |
|---|
| 1203 | n/a | seq_for_testlist(struct compiling *c, const node *n) |
|---|
| 1204 | n/a | { |
|---|
| 1205 | n/a | /* testlist: test (',' test)* [','] |
|---|
| 1206 | n/a | testlist_star_expr: test|star_expr (',' test|star_expr)* [','] |
|---|
| 1207 | n/a | */ |
|---|
| 1208 | n/a | asdl_seq *seq; |
|---|
| 1209 | n/a | expr_ty expression; |
|---|
| 1210 | n/a | int i; |
|---|
| 1211 | n/a | assert(TYPE(n) == testlist || TYPE(n) == testlist_star_expr || TYPE(n) == testlist_comp); |
|---|
| 1212 | n/a | |
|---|
| 1213 | n/a | seq = _Py_asdl_seq_new((NCH(n) + 1) / 2, c->c_arena); |
|---|
| 1214 | n/a | if (!seq) |
|---|
| 1215 | n/a | return NULL; |
|---|
| 1216 | n/a | |
|---|
| 1217 | n/a | for (i = 0; i < NCH(n); i += 2) { |
|---|
| 1218 | n/a | const node *ch = CHILD(n, i); |
|---|
| 1219 | n/a | assert(TYPE(ch) == test || TYPE(ch) == test_nocond || TYPE(ch) == star_expr); |
|---|
| 1220 | n/a | |
|---|
| 1221 | n/a | expression = ast_for_expr(c, ch); |
|---|
| 1222 | n/a | if (!expression) |
|---|
| 1223 | n/a | return NULL; |
|---|
| 1224 | n/a | |
|---|
| 1225 | n/a | assert(i / 2 < seq->size); |
|---|
| 1226 | n/a | asdl_seq_SET(seq, i / 2, expression); |
|---|
| 1227 | n/a | } |
|---|
| 1228 | n/a | return seq; |
|---|
| 1229 | n/a | } |
|---|
| 1230 | n/a | |
|---|
| 1231 | n/a | static arg_ty |
|---|
| 1232 | n/a | ast_for_arg(struct compiling *c, const node *n) |
|---|
| 1233 | n/a | { |
|---|
| 1234 | n/a | identifier name; |
|---|
| 1235 | n/a | expr_ty annotation = NULL; |
|---|
| 1236 | n/a | node *ch; |
|---|
| 1237 | n/a | arg_ty ret; |
|---|
| 1238 | n/a | |
|---|
| 1239 | n/a | assert(TYPE(n) == tfpdef || TYPE(n) == vfpdef); |
|---|
| 1240 | n/a | ch = CHILD(n, 0); |
|---|
| 1241 | n/a | name = NEW_IDENTIFIER(ch); |
|---|
| 1242 | n/a | if (!name) |
|---|
| 1243 | n/a | return NULL; |
|---|
| 1244 | n/a | if (forbidden_name(c, name, ch, 0)) |
|---|
| 1245 | n/a | return NULL; |
|---|
| 1246 | n/a | |
|---|
| 1247 | n/a | if (NCH(n) == 3 && TYPE(CHILD(n, 1)) == COLON) { |
|---|
| 1248 | n/a | annotation = ast_for_expr(c, CHILD(n, 2)); |
|---|
| 1249 | n/a | if (!annotation) |
|---|
| 1250 | n/a | return NULL; |
|---|
| 1251 | n/a | } |
|---|
| 1252 | n/a | |
|---|
| 1253 | n/a | ret = arg(name, annotation, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 1254 | n/a | if (!ret) |
|---|
| 1255 | n/a | return NULL; |
|---|
| 1256 | n/a | return ret; |
|---|
| 1257 | n/a | } |
|---|
| 1258 | n/a | |
|---|
| 1259 | n/a | /* returns -1 if failed to handle keyword only arguments |
|---|
| 1260 | n/a | returns new position to keep processing if successful |
|---|
| 1261 | n/a | (',' tfpdef ['=' test])* |
|---|
| 1262 | n/a | ^^^ |
|---|
| 1263 | n/a | start pointing here |
|---|
| 1264 | n/a | */ |
|---|
| 1265 | n/a | static int |
|---|
| 1266 | n/a | handle_keywordonly_args(struct compiling *c, const node *n, int start, |
|---|
| 1267 | n/a | asdl_seq *kwonlyargs, asdl_seq *kwdefaults) |
|---|
| 1268 | n/a | { |
|---|
| 1269 | n/a | PyObject *argname; |
|---|
| 1270 | n/a | node *ch; |
|---|
| 1271 | n/a | expr_ty expression, annotation; |
|---|
| 1272 | n/a | arg_ty arg; |
|---|
| 1273 | n/a | int i = start; |
|---|
| 1274 | n/a | int j = 0; /* index for kwdefaults and kwonlyargs */ |
|---|
| 1275 | n/a | |
|---|
| 1276 | n/a | if (kwonlyargs == NULL) { |
|---|
| 1277 | n/a | ast_error(c, CHILD(n, start), "named arguments must follow bare *"); |
|---|
| 1278 | n/a | return -1; |
|---|
| 1279 | n/a | } |
|---|
| 1280 | n/a | assert(kwdefaults != NULL); |
|---|
| 1281 | n/a | while (i < NCH(n)) { |
|---|
| 1282 | n/a | ch = CHILD(n, i); |
|---|
| 1283 | n/a | switch (TYPE(ch)) { |
|---|
| 1284 | n/a | case vfpdef: |
|---|
| 1285 | n/a | case tfpdef: |
|---|
| 1286 | n/a | if (i + 1 < NCH(n) && TYPE(CHILD(n, i + 1)) == EQUAL) { |
|---|
| 1287 | n/a | expression = ast_for_expr(c, CHILD(n, i + 2)); |
|---|
| 1288 | n/a | if (!expression) |
|---|
| 1289 | n/a | goto error; |
|---|
| 1290 | n/a | asdl_seq_SET(kwdefaults, j, expression); |
|---|
| 1291 | n/a | i += 2; /* '=' and test */ |
|---|
| 1292 | n/a | } |
|---|
| 1293 | n/a | else { /* setting NULL if no default value exists */ |
|---|
| 1294 | n/a | asdl_seq_SET(kwdefaults, j, NULL); |
|---|
| 1295 | n/a | } |
|---|
| 1296 | n/a | if (NCH(ch) == 3) { |
|---|
| 1297 | n/a | /* ch is NAME ':' test */ |
|---|
| 1298 | n/a | annotation = ast_for_expr(c, CHILD(ch, 2)); |
|---|
| 1299 | n/a | if (!annotation) |
|---|
| 1300 | n/a | goto error; |
|---|
| 1301 | n/a | } |
|---|
| 1302 | n/a | else { |
|---|
| 1303 | n/a | annotation = NULL; |
|---|
| 1304 | n/a | } |
|---|
| 1305 | n/a | ch = CHILD(ch, 0); |
|---|
| 1306 | n/a | argname = NEW_IDENTIFIER(ch); |
|---|
| 1307 | n/a | if (!argname) |
|---|
| 1308 | n/a | goto error; |
|---|
| 1309 | n/a | if (forbidden_name(c, argname, ch, 0)) |
|---|
| 1310 | n/a | goto error; |
|---|
| 1311 | n/a | arg = arg(argname, annotation, LINENO(ch), ch->n_col_offset, |
|---|
| 1312 | n/a | c->c_arena); |
|---|
| 1313 | n/a | if (!arg) |
|---|
| 1314 | n/a | goto error; |
|---|
| 1315 | n/a | asdl_seq_SET(kwonlyargs, j++, arg); |
|---|
| 1316 | n/a | i += 2; /* the name and the comma */ |
|---|
| 1317 | n/a | break; |
|---|
| 1318 | n/a | case DOUBLESTAR: |
|---|
| 1319 | n/a | return i; |
|---|
| 1320 | n/a | default: |
|---|
| 1321 | n/a | ast_error(c, ch, "unexpected node"); |
|---|
| 1322 | n/a | goto error; |
|---|
| 1323 | n/a | } |
|---|
| 1324 | n/a | } |
|---|
| 1325 | n/a | return i; |
|---|
| 1326 | n/a | error: |
|---|
| 1327 | n/a | return -1; |
|---|
| 1328 | n/a | } |
|---|
| 1329 | n/a | |
|---|
| 1330 | n/a | /* Create AST for argument list. */ |
|---|
| 1331 | n/a | |
|---|
| 1332 | n/a | static arguments_ty |
|---|
| 1333 | n/a | ast_for_arguments(struct compiling *c, const node *n) |
|---|
| 1334 | n/a | { |
|---|
| 1335 | n/a | /* This function handles both typedargslist (function definition) |
|---|
| 1336 | n/a | and varargslist (lambda definition). |
|---|
| 1337 | n/a | |
|---|
| 1338 | n/a | parameters: '(' [typedargslist] ')' |
|---|
| 1339 | n/a | typedargslist: (tfpdef ['=' test] (',' tfpdef ['=' test])* [',' [ |
|---|
| 1340 | n/a | '*' [tfpdef] (',' tfpdef ['=' test])* [',' ['**' tfpdef [',']]] |
|---|
| 1341 | n/a | | '**' tfpdef [',']]] |
|---|
| 1342 | n/a | | '*' [tfpdef] (',' tfpdef ['=' test])* [',' ['**' tfpdef [',']]] |
|---|
| 1343 | n/a | | '**' tfpdef [',']) |
|---|
| 1344 | n/a | tfpdef: NAME [':' test] |
|---|
| 1345 | n/a | varargslist: (vfpdef ['=' test] (',' vfpdef ['=' test])* [',' [ |
|---|
| 1346 | n/a | '*' [vfpdef] (',' vfpdef ['=' test])* [',' ['**' vfpdef [',']]] |
|---|
| 1347 | n/a | | '**' vfpdef [',']]] |
|---|
| 1348 | n/a | | '*' [vfpdef] (',' vfpdef ['=' test])* [',' ['**' vfpdef [',']]] |
|---|
| 1349 | n/a | | '**' vfpdef [','] |
|---|
| 1350 | n/a | ) |
|---|
| 1351 | n/a | vfpdef: NAME |
|---|
| 1352 | n/a | |
|---|
| 1353 | n/a | */ |
|---|
| 1354 | n/a | int i, j, k, nposargs = 0, nkwonlyargs = 0; |
|---|
| 1355 | n/a | int nposdefaults = 0, found_default = 0; |
|---|
| 1356 | n/a | asdl_seq *posargs, *posdefaults, *kwonlyargs, *kwdefaults; |
|---|
| 1357 | n/a | arg_ty vararg = NULL, kwarg = NULL; |
|---|
| 1358 | n/a | arg_ty arg; |
|---|
| 1359 | n/a | node *ch; |
|---|
| 1360 | n/a | |
|---|
| 1361 | n/a | if (TYPE(n) == parameters) { |
|---|
| 1362 | n/a | if (NCH(n) == 2) /* () as argument list */ |
|---|
| 1363 | n/a | return arguments(NULL, NULL, NULL, NULL, NULL, NULL, c->c_arena); |
|---|
| 1364 | n/a | n = CHILD(n, 1); |
|---|
| 1365 | n/a | } |
|---|
| 1366 | n/a | assert(TYPE(n) == typedargslist || TYPE(n) == varargslist); |
|---|
| 1367 | n/a | |
|---|
| 1368 | n/a | /* First count the number of positional args & defaults. The |
|---|
| 1369 | n/a | variable i is the loop index for this for loop and the next. |
|---|
| 1370 | n/a | The next loop picks up where the first leaves off. |
|---|
| 1371 | n/a | */ |
|---|
| 1372 | n/a | for (i = 0; i < NCH(n); i++) { |
|---|
| 1373 | n/a | ch = CHILD(n, i); |
|---|
| 1374 | n/a | if (TYPE(ch) == STAR) { |
|---|
| 1375 | n/a | /* skip star */ |
|---|
| 1376 | n/a | i++; |
|---|
| 1377 | n/a | if (i < NCH(n) && /* skip argument following star */ |
|---|
| 1378 | n/a | (TYPE(CHILD(n, i)) == tfpdef || |
|---|
| 1379 | n/a | TYPE(CHILD(n, i)) == vfpdef)) { |
|---|
| 1380 | n/a | i++; |
|---|
| 1381 | n/a | } |
|---|
| 1382 | n/a | break; |
|---|
| 1383 | n/a | } |
|---|
| 1384 | n/a | if (TYPE(ch) == DOUBLESTAR) break; |
|---|
| 1385 | n/a | if (TYPE(ch) == vfpdef || TYPE(ch) == tfpdef) nposargs++; |
|---|
| 1386 | n/a | if (TYPE(ch) == EQUAL) nposdefaults++; |
|---|
| 1387 | n/a | } |
|---|
| 1388 | n/a | /* count the number of keyword only args & |
|---|
| 1389 | n/a | defaults for keyword only args */ |
|---|
| 1390 | n/a | for ( ; i < NCH(n); ++i) { |
|---|
| 1391 | n/a | ch = CHILD(n, i); |
|---|
| 1392 | n/a | if (TYPE(ch) == DOUBLESTAR) break; |
|---|
| 1393 | n/a | if (TYPE(ch) == tfpdef || TYPE(ch) == vfpdef) nkwonlyargs++; |
|---|
| 1394 | n/a | } |
|---|
| 1395 | n/a | posargs = (nposargs ? _Py_asdl_seq_new(nposargs, c->c_arena) : NULL); |
|---|
| 1396 | n/a | if (!posargs && nposargs) |
|---|
| 1397 | n/a | return NULL; |
|---|
| 1398 | n/a | kwonlyargs = (nkwonlyargs ? |
|---|
| 1399 | n/a | _Py_asdl_seq_new(nkwonlyargs, c->c_arena) : NULL); |
|---|
| 1400 | n/a | if (!kwonlyargs && nkwonlyargs) |
|---|
| 1401 | n/a | return NULL; |
|---|
| 1402 | n/a | posdefaults = (nposdefaults ? |
|---|
| 1403 | n/a | _Py_asdl_seq_new(nposdefaults, c->c_arena) : NULL); |
|---|
| 1404 | n/a | if (!posdefaults && nposdefaults) |
|---|
| 1405 | n/a | return NULL; |
|---|
| 1406 | n/a | /* The length of kwonlyargs and kwdefaults are same |
|---|
| 1407 | n/a | since we set NULL as default for keyword only argument w/o default |
|---|
| 1408 | n/a | - we have sequence data structure, but no dictionary */ |
|---|
| 1409 | n/a | kwdefaults = (nkwonlyargs ? |
|---|
| 1410 | n/a | _Py_asdl_seq_new(nkwonlyargs, c->c_arena) : NULL); |
|---|
| 1411 | n/a | if (!kwdefaults && nkwonlyargs) |
|---|
| 1412 | n/a | return NULL; |
|---|
| 1413 | n/a | |
|---|
| 1414 | n/a | /* tfpdef: NAME [':' test] |
|---|
| 1415 | n/a | vfpdef: NAME |
|---|
| 1416 | n/a | */ |
|---|
| 1417 | n/a | i = 0; |
|---|
| 1418 | n/a | j = 0; /* index for defaults */ |
|---|
| 1419 | n/a | k = 0; /* index for args */ |
|---|
| 1420 | n/a | while (i < NCH(n)) { |
|---|
| 1421 | n/a | ch = CHILD(n, i); |
|---|
| 1422 | n/a | switch (TYPE(ch)) { |
|---|
| 1423 | n/a | case tfpdef: |
|---|
| 1424 | n/a | case vfpdef: |
|---|
| 1425 | n/a | /* XXX Need to worry about checking if TYPE(CHILD(n, i+1)) is |
|---|
| 1426 | n/a | anything other than EQUAL or a comma? */ |
|---|
| 1427 | n/a | /* XXX Should NCH(n) check be made a separate check? */ |
|---|
| 1428 | n/a | if (i + 1 < NCH(n) && TYPE(CHILD(n, i + 1)) == EQUAL) { |
|---|
| 1429 | n/a | expr_ty expression = ast_for_expr(c, CHILD(n, i + 2)); |
|---|
| 1430 | n/a | if (!expression) |
|---|
| 1431 | n/a | return NULL; |
|---|
| 1432 | n/a | assert(posdefaults != NULL); |
|---|
| 1433 | n/a | asdl_seq_SET(posdefaults, j++, expression); |
|---|
| 1434 | n/a | i += 2; |
|---|
| 1435 | n/a | found_default = 1; |
|---|
| 1436 | n/a | } |
|---|
| 1437 | n/a | else if (found_default) { |
|---|
| 1438 | n/a | ast_error(c, n, |
|---|
| 1439 | n/a | "non-default argument follows default argument"); |
|---|
| 1440 | n/a | return NULL; |
|---|
| 1441 | n/a | } |
|---|
| 1442 | n/a | arg = ast_for_arg(c, ch); |
|---|
| 1443 | n/a | if (!arg) |
|---|
| 1444 | n/a | return NULL; |
|---|
| 1445 | n/a | asdl_seq_SET(posargs, k++, arg); |
|---|
| 1446 | n/a | i += 2; /* the name and the comma */ |
|---|
| 1447 | n/a | break; |
|---|
| 1448 | n/a | case STAR: |
|---|
| 1449 | n/a | if (i+1 >= NCH(n) || |
|---|
| 1450 | n/a | (i+2 == NCH(n) && TYPE(CHILD(n, i+1)) == COMMA)) { |
|---|
| 1451 | n/a | ast_error(c, CHILD(n, i), |
|---|
| 1452 | n/a | "named arguments must follow bare *"); |
|---|
| 1453 | n/a | return NULL; |
|---|
| 1454 | n/a | } |
|---|
| 1455 | n/a | ch = CHILD(n, i+1); /* tfpdef or COMMA */ |
|---|
| 1456 | n/a | if (TYPE(ch) == COMMA) { |
|---|
| 1457 | n/a | int res = 0; |
|---|
| 1458 | n/a | i += 2; /* now follows keyword only arguments */ |
|---|
| 1459 | n/a | res = handle_keywordonly_args(c, n, i, |
|---|
| 1460 | n/a | kwonlyargs, kwdefaults); |
|---|
| 1461 | n/a | if (res == -1) return NULL; |
|---|
| 1462 | n/a | i = res; /* res has new position to process */ |
|---|
| 1463 | n/a | } |
|---|
| 1464 | n/a | else { |
|---|
| 1465 | n/a | vararg = ast_for_arg(c, ch); |
|---|
| 1466 | n/a | if (!vararg) |
|---|
| 1467 | n/a | return NULL; |
|---|
| 1468 | n/a | |
|---|
| 1469 | n/a | i += 3; |
|---|
| 1470 | n/a | if (i < NCH(n) && (TYPE(CHILD(n, i)) == tfpdef |
|---|
| 1471 | n/a | || TYPE(CHILD(n, i)) == vfpdef)) { |
|---|
| 1472 | n/a | int res = 0; |
|---|
| 1473 | n/a | res = handle_keywordonly_args(c, n, i, |
|---|
| 1474 | n/a | kwonlyargs, kwdefaults); |
|---|
| 1475 | n/a | if (res == -1) return NULL; |
|---|
| 1476 | n/a | i = res; /* res has new position to process */ |
|---|
| 1477 | n/a | } |
|---|
| 1478 | n/a | } |
|---|
| 1479 | n/a | break; |
|---|
| 1480 | n/a | case DOUBLESTAR: |
|---|
| 1481 | n/a | ch = CHILD(n, i+1); /* tfpdef */ |
|---|
| 1482 | n/a | assert(TYPE(ch) == tfpdef || TYPE(ch) == vfpdef); |
|---|
| 1483 | n/a | kwarg = ast_for_arg(c, ch); |
|---|
| 1484 | n/a | if (!kwarg) |
|---|
| 1485 | n/a | return NULL; |
|---|
| 1486 | n/a | i += 3; |
|---|
| 1487 | n/a | break; |
|---|
| 1488 | n/a | default: |
|---|
| 1489 | n/a | PyErr_Format(PyExc_SystemError, |
|---|
| 1490 | n/a | "unexpected node in varargslist: %d @ %d", |
|---|
| 1491 | n/a | TYPE(ch), i); |
|---|
| 1492 | n/a | return NULL; |
|---|
| 1493 | n/a | } |
|---|
| 1494 | n/a | } |
|---|
| 1495 | n/a | return arguments(posargs, vararg, kwonlyargs, kwdefaults, kwarg, posdefaults, c->c_arena); |
|---|
| 1496 | n/a | } |
|---|
| 1497 | n/a | |
|---|
| 1498 | n/a | static expr_ty |
|---|
| 1499 | n/a | ast_for_dotted_name(struct compiling *c, const node *n) |
|---|
| 1500 | n/a | { |
|---|
| 1501 | n/a | expr_ty e; |
|---|
| 1502 | n/a | identifier id; |
|---|
| 1503 | n/a | int lineno, col_offset; |
|---|
| 1504 | n/a | int i; |
|---|
| 1505 | n/a | |
|---|
| 1506 | n/a | REQ(n, dotted_name); |
|---|
| 1507 | n/a | |
|---|
| 1508 | n/a | lineno = LINENO(n); |
|---|
| 1509 | n/a | col_offset = n->n_col_offset; |
|---|
| 1510 | n/a | |
|---|
| 1511 | n/a | id = NEW_IDENTIFIER(CHILD(n, 0)); |
|---|
| 1512 | n/a | if (!id) |
|---|
| 1513 | n/a | return NULL; |
|---|
| 1514 | n/a | e = Name(id, Load, lineno, col_offset, c->c_arena); |
|---|
| 1515 | n/a | if (!e) |
|---|
| 1516 | n/a | return NULL; |
|---|
| 1517 | n/a | |
|---|
| 1518 | n/a | for (i = 2; i < NCH(n); i+=2) { |
|---|
| 1519 | n/a | id = NEW_IDENTIFIER(CHILD(n, i)); |
|---|
| 1520 | n/a | if (!id) |
|---|
| 1521 | n/a | return NULL; |
|---|
| 1522 | n/a | e = Attribute(e, id, Load, lineno, col_offset, c->c_arena); |
|---|
| 1523 | n/a | if (!e) |
|---|
| 1524 | n/a | return NULL; |
|---|
| 1525 | n/a | } |
|---|
| 1526 | n/a | |
|---|
| 1527 | n/a | return e; |
|---|
| 1528 | n/a | } |
|---|
| 1529 | n/a | |
|---|
| 1530 | n/a | static expr_ty |
|---|
| 1531 | n/a | ast_for_decorator(struct compiling *c, const node *n) |
|---|
| 1532 | n/a | { |
|---|
| 1533 | n/a | /* decorator: '@' dotted_name [ '(' [arglist] ')' ] NEWLINE */ |
|---|
| 1534 | n/a | expr_ty d = NULL; |
|---|
| 1535 | n/a | expr_ty name_expr; |
|---|
| 1536 | n/a | |
|---|
| 1537 | n/a | REQ(n, decorator); |
|---|
| 1538 | n/a | REQ(CHILD(n, 0), AT); |
|---|
| 1539 | n/a | REQ(RCHILD(n, -1), NEWLINE); |
|---|
| 1540 | n/a | |
|---|
| 1541 | n/a | name_expr = ast_for_dotted_name(c, CHILD(n, 1)); |
|---|
| 1542 | n/a | if (!name_expr) |
|---|
| 1543 | n/a | return NULL; |
|---|
| 1544 | n/a | |
|---|
| 1545 | n/a | if (NCH(n) == 3) { /* No arguments */ |
|---|
| 1546 | n/a | d = name_expr; |
|---|
| 1547 | n/a | name_expr = NULL; |
|---|
| 1548 | n/a | } |
|---|
| 1549 | n/a | else if (NCH(n) == 5) { /* Call with no arguments */ |
|---|
| 1550 | n/a | d = Call(name_expr, NULL, NULL, LINENO(n), |
|---|
| 1551 | n/a | n->n_col_offset, c->c_arena); |
|---|
| 1552 | n/a | if (!d) |
|---|
| 1553 | n/a | return NULL; |
|---|
| 1554 | n/a | name_expr = NULL; |
|---|
| 1555 | n/a | } |
|---|
| 1556 | n/a | else { |
|---|
| 1557 | n/a | d = ast_for_call(c, CHILD(n, 3), name_expr); |
|---|
| 1558 | n/a | if (!d) |
|---|
| 1559 | n/a | return NULL; |
|---|
| 1560 | n/a | name_expr = NULL; |
|---|
| 1561 | n/a | } |
|---|
| 1562 | n/a | |
|---|
| 1563 | n/a | return d; |
|---|
| 1564 | n/a | } |
|---|
| 1565 | n/a | |
|---|
| 1566 | n/a | static asdl_seq* |
|---|
| 1567 | n/a | ast_for_decorators(struct compiling *c, const node *n) |
|---|
| 1568 | n/a | { |
|---|
| 1569 | n/a | asdl_seq* decorator_seq; |
|---|
| 1570 | n/a | expr_ty d; |
|---|
| 1571 | n/a | int i; |
|---|
| 1572 | n/a | |
|---|
| 1573 | n/a | REQ(n, decorators); |
|---|
| 1574 | n/a | decorator_seq = _Py_asdl_seq_new(NCH(n), c->c_arena); |
|---|
| 1575 | n/a | if (!decorator_seq) |
|---|
| 1576 | n/a | return NULL; |
|---|
| 1577 | n/a | |
|---|
| 1578 | n/a | for (i = 0; i < NCH(n); i++) { |
|---|
| 1579 | n/a | d = ast_for_decorator(c, CHILD(n, i)); |
|---|
| 1580 | n/a | if (!d) |
|---|
| 1581 | n/a | return NULL; |
|---|
| 1582 | n/a | asdl_seq_SET(decorator_seq, i, d); |
|---|
| 1583 | n/a | } |
|---|
| 1584 | n/a | return decorator_seq; |
|---|
| 1585 | n/a | } |
|---|
| 1586 | n/a | |
|---|
| 1587 | n/a | static stmt_ty |
|---|
| 1588 | n/a | ast_for_funcdef_impl(struct compiling *c, const node *n, |
|---|
| 1589 | n/a | asdl_seq *decorator_seq, int is_async) |
|---|
| 1590 | n/a | { |
|---|
| 1591 | n/a | /* funcdef: 'def' NAME parameters ['->' test] ':' suite */ |
|---|
| 1592 | n/a | identifier name; |
|---|
| 1593 | n/a | arguments_ty args; |
|---|
| 1594 | n/a | asdl_seq *body; |
|---|
| 1595 | n/a | expr_ty returns = NULL; |
|---|
| 1596 | n/a | int name_i = 1; |
|---|
| 1597 | n/a | |
|---|
| 1598 | n/a | REQ(n, funcdef); |
|---|
| 1599 | n/a | |
|---|
| 1600 | n/a | name = NEW_IDENTIFIER(CHILD(n, name_i)); |
|---|
| 1601 | n/a | if (!name) |
|---|
| 1602 | n/a | return NULL; |
|---|
| 1603 | n/a | if (forbidden_name(c, name, CHILD(n, name_i), 0)) |
|---|
| 1604 | n/a | return NULL; |
|---|
| 1605 | n/a | args = ast_for_arguments(c, CHILD(n, name_i + 1)); |
|---|
| 1606 | n/a | if (!args) |
|---|
| 1607 | n/a | return NULL; |
|---|
| 1608 | n/a | if (TYPE(CHILD(n, name_i+2)) == RARROW) { |
|---|
| 1609 | n/a | returns = ast_for_expr(c, CHILD(n, name_i + 3)); |
|---|
| 1610 | n/a | if (!returns) |
|---|
| 1611 | n/a | return NULL; |
|---|
| 1612 | n/a | name_i += 2; |
|---|
| 1613 | n/a | } |
|---|
| 1614 | n/a | body = ast_for_suite(c, CHILD(n, name_i + 3)); |
|---|
| 1615 | n/a | if (!body) |
|---|
| 1616 | n/a | return NULL; |
|---|
| 1617 | n/a | |
|---|
| 1618 | n/a | if (is_async) |
|---|
| 1619 | n/a | return AsyncFunctionDef(name, args, body, decorator_seq, returns, |
|---|
| 1620 | n/a | LINENO(n), |
|---|
| 1621 | n/a | n->n_col_offset, c->c_arena); |
|---|
| 1622 | n/a | else |
|---|
| 1623 | n/a | return FunctionDef(name, args, body, decorator_seq, returns, |
|---|
| 1624 | n/a | LINENO(n), |
|---|
| 1625 | n/a | n->n_col_offset, c->c_arena); |
|---|
| 1626 | n/a | } |
|---|
| 1627 | n/a | |
|---|
| 1628 | n/a | static stmt_ty |
|---|
| 1629 | n/a | ast_for_async_funcdef(struct compiling *c, const node *n, asdl_seq *decorator_seq) |
|---|
| 1630 | n/a | { |
|---|
| 1631 | n/a | /* async_funcdef: ASYNC funcdef */ |
|---|
| 1632 | n/a | REQ(n, async_funcdef); |
|---|
| 1633 | n/a | REQ(CHILD(n, 0), ASYNC); |
|---|
| 1634 | n/a | REQ(CHILD(n, 1), funcdef); |
|---|
| 1635 | n/a | |
|---|
| 1636 | n/a | return ast_for_funcdef_impl(c, CHILD(n, 1), decorator_seq, |
|---|
| 1637 | n/a | 1 /* is_async */); |
|---|
| 1638 | n/a | } |
|---|
| 1639 | n/a | |
|---|
| 1640 | n/a | static stmt_ty |
|---|
| 1641 | n/a | ast_for_funcdef(struct compiling *c, const node *n, asdl_seq *decorator_seq) |
|---|
| 1642 | n/a | { |
|---|
| 1643 | n/a | /* funcdef: 'def' NAME parameters ['->' test] ':' suite */ |
|---|
| 1644 | n/a | return ast_for_funcdef_impl(c, n, decorator_seq, |
|---|
| 1645 | n/a | 0 /* is_async */); |
|---|
| 1646 | n/a | } |
|---|
| 1647 | n/a | |
|---|
| 1648 | n/a | |
|---|
| 1649 | n/a | static stmt_ty |
|---|
| 1650 | n/a | ast_for_async_stmt(struct compiling *c, const node *n) |
|---|
| 1651 | n/a | { |
|---|
| 1652 | n/a | /* async_stmt: ASYNC (funcdef | with_stmt | for_stmt) */ |
|---|
| 1653 | n/a | REQ(n, async_stmt); |
|---|
| 1654 | n/a | REQ(CHILD(n, 0), ASYNC); |
|---|
| 1655 | n/a | |
|---|
| 1656 | n/a | switch (TYPE(CHILD(n, 1))) { |
|---|
| 1657 | n/a | case funcdef: |
|---|
| 1658 | n/a | return ast_for_funcdef_impl(c, CHILD(n, 1), NULL, |
|---|
| 1659 | n/a | 1 /* is_async */); |
|---|
| 1660 | n/a | case with_stmt: |
|---|
| 1661 | n/a | return ast_for_with_stmt(c, CHILD(n, 1), |
|---|
| 1662 | n/a | 1 /* is_async */); |
|---|
| 1663 | n/a | |
|---|
| 1664 | n/a | case for_stmt: |
|---|
| 1665 | n/a | return ast_for_for_stmt(c, CHILD(n, 1), |
|---|
| 1666 | n/a | 1 /* is_async */); |
|---|
| 1667 | n/a | |
|---|
| 1668 | n/a | default: |
|---|
| 1669 | n/a | PyErr_Format(PyExc_SystemError, |
|---|
| 1670 | n/a | "invalid async stament: %s", |
|---|
| 1671 | n/a | STR(CHILD(n, 1))); |
|---|
| 1672 | n/a | return NULL; |
|---|
| 1673 | n/a | } |
|---|
| 1674 | n/a | } |
|---|
| 1675 | n/a | |
|---|
| 1676 | n/a | static stmt_ty |
|---|
| 1677 | n/a | ast_for_decorated(struct compiling *c, const node *n) |
|---|
| 1678 | n/a | { |
|---|
| 1679 | n/a | /* decorated: decorators (classdef | funcdef | async_funcdef) */ |
|---|
| 1680 | n/a | stmt_ty thing = NULL; |
|---|
| 1681 | n/a | asdl_seq *decorator_seq = NULL; |
|---|
| 1682 | n/a | |
|---|
| 1683 | n/a | REQ(n, decorated); |
|---|
| 1684 | n/a | |
|---|
| 1685 | n/a | decorator_seq = ast_for_decorators(c, CHILD(n, 0)); |
|---|
| 1686 | n/a | if (!decorator_seq) |
|---|
| 1687 | n/a | return NULL; |
|---|
| 1688 | n/a | |
|---|
| 1689 | n/a | assert(TYPE(CHILD(n, 1)) == funcdef || |
|---|
| 1690 | n/a | TYPE(CHILD(n, 1)) == async_funcdef || |
|---|
| 1691 | n/a | TYPE(CHILD(n, 1)) == classdef); |
|---|
| 1692 | n/a | |
|---|
| 1693 | n/a | if (TYPE(CHILD(n, 1)) == funcdef) { |
|---|
| 1694 | n/a | thing = ast_for_funcdef(c, CHILD(n, 1), decorator_seq); |
|---|
| 1695 | n/a | } else if (TYPE(CHILD(n, 1)) == classdef) { |
|---|
| 1696 | n/a | thing = ast_for_classdef(c, CHILD(n, 1), decorator_seq); |
|---|
| 1697 | n/a | } else if (TYPE(CHILD(n, 1)) == async_funcdef) { |
|---|
| 1698 | n/a | thing = ast_for_async_funcdef(c, CHILD(n, 1), decorator_seq); |
|---|
| 1699 | n/a | } |
|---|
| 1700 | n/a | /* we count the decorators in when talking about the class' or |
|---|
| 1701 | n/a | * function's line number */ |
|---|
| 1702 | n/a | if (thing) { |
|---|
| 1703 | n/a | thing->lineno = LINENO(n); |
|---|
| 1704 | n/a | thing->col_offset = n->n_col_offset; |
|---|
| 1705 | n/a | } |
|---|
| 1706 | n/a | return thing; |
|---|
| 1707 | n/a | } |
|---|
| 1708 | n/a | |
|---|
| 1709 | n/a | static expr_ty |
|---|
| 1710 | n/a | ast_for_lambdef(struct compiling *c, const node *n) |
|---|
| 1711 | n/a | { |
|---|
| 1712 | n/a | /* lambdef: 'lambda' [varargslist] ':' test |
|---|
| 1713 | n/a | lambdef_nocond: 'lambda' [varargslist] ':' test_nocond */ |
|---|
| 1714 | n/a | arguments_ty args; |
|---|
| 1715 | n/a | expr_ty expression; |
|---|
| 1716 | n/a | |
|---|
| 1717 | n/a | if (NCH(n) == 3) { |
|---|
| 1718 | n/a | args = arguments(NULL, NULL, NULL, NULL, NULL, NULL, c->c_arena); |
|---|
| 1719 | n/a | if (!args) |
|---|
| 1720 | n/a | return NULL; |
|---|
| 1721 | n/a | expression = ast_for_expr(c, CHILD(n, 2)); |
|---|
| 1722 | n/a | if (!expression) |
|---|
| 1723 | n/a | return NULL; |
|---|
| 1724 | n/a | } |
|---|
| 1725 | n/a | else { |
|---|
| 1726 | n/a | args = ast_for_arguments(c, CHILD(n, 1)); |
|---|
| 1727 | n/a | if (!args) |
|---|
| 1728 | n/a | return NULL; |
|---|
| 1729 | n/a | expression = ast_for_expr(c, CHILD(n, 3)); |
|---|
| 1730 | n/a | if (!expression) |
|---|
| 1731 | n/a | return NULL; |
|---|
| 1732 | n/a | } |
|---|
| 1733 | n/a | |
|---|
| 1734 | n/a | return Lambda(args, expression, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 1735 | n/a | } |
|---|
| 1736 | n/a | |
|---|
| 1737 | n/a | static expr_ty |
|---|
| 1738 | n/a | ast_for_ifexpr(struct compiling *c, const node *n) |
|---|
| 1739 | n/a | { |
|---|
| 1740 | n/a | /* test: or_test 'if' or_test 'else' test */ |
|---|
| 1741 | n/a | expr_ty expression, body, orelse; |
|---|
| 1742 | n/a | |
|---|
| 1743 | n/a | assert(NCH(n) == 5); |
|---|
| 1744 | n/a | body = ast_for_expr(c, CHILD(n, 0)); |
|---|
| 1745 | n/a | if (!body) |
|---|
| 1746 | n/a | return NULL; |
|---|
| 1747 | n/a | expression = ast_for_expr(c, CHILD(n, 2)); |
|---|
| 1748 | n/a | if (!expression) |
|---|
| 1749 | n/a | return NULL; |
|---|
| 1750 | n/a | orelse = ast_for_expr(c, CHILD(n, 4)); |
|---|
| 1751 | n/a | if (!orelse) |
|---|
| 1752 | n/a | return NULL; |
|---|
| 1753 | n/a | return IfExp(expression, body, orelse, LINENO(n), n->n_col_offset, |
|---|
| 1754 | n/a | c->c_arena); |
|---|
| 1755 | n/a | } |
|---|
| 1756 | n/a | |
|---|
| 1757 | n/a | /* |
|---|
| 1758 | n/a | Count the number of 'for' loops in a comprehension. |
|---|
| 1759 | n/a | |
|---|
| 1760 | n/a | Helper for ast_for_comprehension(). |
|---|
| 1761 | n/a | */ |
|---|
| 1762 | n/a | |
|---|
| 1763 | n/a | static int |
|---|
| 1764 | n/a | count_comp_fors(struct compiling *c, const node *n) |
|---|
| 1765 | n/a | { |
|---|
| 1766 | n/a | int n_fors = 0; |
|---|
| 1767 | n/a | int is_async; |
|---|
| 1768 | n/a | |
|---|
| 1769 | n/a | count_comp_for: |
|---|
| 1770 | n/a | is_async = 0; |
|---|
| 1771 | n/a | n_fors++; |
|---|
| 1772 | n/a | REQ(n, comp_for); |
|---|
| 1773 | n/a | if (TYPE(CHILD(n, 0)) == ASYNC) { |
|---|
| 1774 | n/a | is_async = 1; |
|---|
| 1775 | n/a | } |
|---|
| 1776 | n/a | if (NCH(n) == (5 + is_async)) { |
|---|
| 1777 | n/a | n = CHILD(n, 4 + is_async); |
|---|
| 1778 | n/a | } |
|---|
| 1779 | n/a | else { |
|---|
| 1780 | n/a | return n_fors; |
|---|
| 1781 | n/a | } |
|---|
| 1782 | n/a | count_comp_iter: |
|---|
| 1783 | n/a | REQ(n, comp_iter); |
|---|
| 1784 | n/a | n = CHILD(n, 0); |
|---|
| 1785 | n/a | if (TYPE(n) == comp_for) |
|---|
| 1786 | n/a | goto count_comp_for; |
|---|
| 1787 | n/a | else if (TYPE(n) == comp_if) { |
|---|
| 1788 | n/a | if (NCH(n) == 3) { |
|---|
| 1789 | n/a | n = CHILD(n, 2); |
|---|
| 1790 | n/a | goto count_comp_iter; |
|---|
| 1791 | n/a | } |
|---|
| 1792 | n/a | else |
|---|
| 1793 | n/a | return n_fors; |
|---|
| 1794 | n/a | } |
|---|
| 1795 | n/a | |
|---|
| 1796 | n/a | /* Should never be reached */ |
|---|
| 1797 | n/a | PyErr_SetString(PyExc_SystemError, |
|---|
| 1798 | n/a | "logic error in count_comp_fors"); |
|---|
| 1799 | n/a | return -1; |
|---|
| 1800 | n/a | } |
|---|
| 1801 | n/a | |
|---|
| 1802 | n/a | /* Count the number of 'if' statements in a comprehension. |
|---|
| 1803 | n/a | |
|---|
| 1804 | n/a | Helper for ast_for_comprehension(). |
|---|
| 1805 | n/a | */ |
|---|
| 1806 | n/a | |
|---|
| 1807 | n/a | static int |
|---|
| 1808 | n/a | count_comp_ifs(struct compiling *c, const node *n) |
|---|
| 1809 | n/a | { |
|---|
| 1810 | n/a | int n_ifs = 0; |
|---|
| 1811 | n/a | |
|---|
| 1812 | n/a | while (1) { |
|---|
| 1813 | n/a | REQ(n, comp_iter); |
|---|
| 1814 | n/a | if (TYPE(CHILD(n, 0)) == comp_for) |
|---|
| 1815 | n/a | return n_ifs; |
|---|
| 1816 | n/a | n = CHILD(n, 0); |
|---|
| 1817 | n/a | REQ(n, comp_if); |
|---|
| 1818 | n/a | n_ifs++; |
|---|
| 1819 | n/a | if (NCH(n) == 2) |
|---|
| 1820 | n/a | return n_ifs; |
|---|
| 1821 | n/a | n = CHILD(n, 2); |
|---|
| 1822 | n/a | } |
|---|
| 1823 | n/a | } |
|---|
| 1824 | n/a | |
|---|
| 1825 | n/a | static asdl_seq * |
|---|
| 1826 | n/a | ast_for_comprehension(struct compiling *c, const node *n) |
|---|
| 1827 | n/a | { |
|---|
| 1828 | n/a | int i, n_fors; |
|---|
| 1829 | n/a | asdl_seq *comps; |
|---|
| 1830 | n/a | |
|---|
| 1831 | n/a | n_fors = count_comp_fors(c, n); |
|---|
| 1832 | n/a | if (n_fors == -1) |
|---|
| 1833 | n/a | return NULL; |
|---|
| 1834 | n/a | |
|---|
| 1835 | n/a | comps = _Py_asdl_seq_new(n_fors, c->c_arena); |
|---|
| 1836 | n/a | if (!comps) |
|---|
| 1837 | n/a | return NULL; |
|---|
| 1838 | n/a | |
|---|
| 1839 | n/a | for (i = 0; i < n_fors; i++) { |
|---|
| 1840 | n/a | comprehension_ty comp; |
|---|
| 1841 | n/a | asdl_seq *t; |
|---|
| 1842 | n/a | expr_ty expression, first; |
|---|
| 1843 | n/a | node *for_ch; |
|---|
| 1844 | n/a | int is_async = 0; |
|---|
| 1845 | n/a | |
|---|
| 1846 | n/a | REQ(n, comp_for); |
|---|
| 1847 | n/a | |
|---|
| 1848 | n/a | if (TYPE(CHILD(n, 0)) == ASYNC) { |
|---|
| 1849 | n/a | is_async = 1; |
|---|
| 1850 | n/a | } |
|---|
| 1851 | n/a | |
|---|
| 1852 | n/a | for_ch = CHILD(n, 1 + is_async); |
|---|
| 1853 | n/a | t = ast_for_exprlist(c, for_ch, Store); |
|---|
| 1854 | n/a | if (!t) |
|---|
| 1855 | n/a | return NULL; |
|---|
| 1856 | n/a | expression = ast_for_expr(c, CHILD(n, 3 + is_async)); |
|---|
| 1857 | n/a | if (!expression) |
|---|
| 1858 | n/a | return NULL; |
|---|
| 1859 | n/a | |
|---|
| 1860 | n/a | /* Check the # of children rather than the length of t, since |
|---|
| 1861 | n/a | (x for x, in ...) has 1 element in t, but still requires a Tuple. */ |
|---|
| 1862 | n/a | first = (expr_ty)asdl_seq_GET(t, 0); |
|---|
| 1863 | n/a | if (NCH(for_ch) == 1) |
|---|
| 1864 | n/a | comp = comprehension(first, expression, NULL, |
|---|
| 1865 | n/a | is_async, c->c_arena); |
|---|
| 1866 | n/a | else |
|---|
| 1867 | n/a | comp = comprehension(Tuple(t, Store, first->lineno, |
|---|
| 1868 | n/a | first->col_offset, c->c_arena), |
|---|
| 1869 | n/a | expression, NULL, is_async, c->c_arena); |
|---|
| 1870 | n/a | if (!comp) |
|---|
| 1871 | n/a | return NULL; |
|---|
| 1872 | n/a | |
|---|
| 1873 | n/a | if (NCH(n) == (5 + is_async)) { |
|---|
| 1874 | n/a | int j, n_ifs; |
|---|
| 1875 | n/a | asdl_seq *ifs; |
|---|
| 1876 | n/a | |
|---|
| 1877 | n/a | n = CHILD(n, 4 + is_async); |
|---|
| 1878 | n/a | n_ifs = count_comp_ifs(c, n); |
|---|
| 1879 | n/a | if (n_ifs == -1) |
|---|
| 1880 | n/a | return NULL; |
|---|
| 1881 | n/a | |
|---|
| 1882 | n/a | ifs = _Py_asdl_seq_new(n_ifs, c->c_arena); |
|---|
| 1883 | n/a | if (!ifs) |
|---|
| 1884 | n/a | return NULL; |
|---|
| 1885 | n/a | |
|---|
| 1886 | n/a | for (j = 0; j < n_ifs; j++) { |
|---|
| 1887 | n/a | REQ(n, comp_iter); |
|---|
| 1888 | n/a | n = CHILD(n, 0); |
|---|
| 1889 | n/a | REQ(n, comp_if); |
|---|
| 1890 | n/a | |
|---|
| 1891 | n/a | expression = ast_for_expr(c, CHILD(n, 1)); |
|---|
| 1892 | n/a | if (!expression) |
|---|
| 1893 | n/a | return NULL; |
|---|
| 1894 | n/a | asdl_seq_SET(ifs, j, expression); |
|---|
| 1895 | n/a | if (NCH(n) == 3) |
|---|
| 1896 | n/a | n = CHILD(n, 2); |
|---|
| 1897 | n/a | } |
|---|
| 1898 | n/a | /* on exit, must guarantee that n is a comp_for */ |
|---|
| 1899 | n/a | if (TYPE(n) == comp_iter) |
|---|
| 1900 | n/a | n = CHILD(n, 0); |
|---|
| 1901 | n/a | comp->ifs = ifs; |
|---|
| 1902 | n/a | } |
|---|
| 1903 | n/a | asdl_seq_SET(comps, i, comp); |
|---|
| 1904 | n/a | } |
|---|
| 1905 | n/a | return comps; |
|---|
| 1906 | n/a | } |
|---|
| 1907 | n/a | |
|---|
| 1908 | n/a | static expr_ty |
|---|
| 1909 | n/a | ast_for_itercomp(struct compiling *c, const node *n, int type) |
|---|
| 1910 | n/a | { |
|---|
| 1911 | n/a | /* testlist_comp: (test|star_expr) |
|---|
| 1912 | n/a | * ( comp_for | (',' (test|star_expr))* [','] ) */ |
|---|
| 1913 | n/a | expr_ty elt; |
|---|
| 1914 | n/a | asdl_seq *comps; |
|---|
| 1915 | n/a | node *ch; |
|---|
| 1916 | n/a | |
|---|
| 1917 | n/a | assert(NCH(n) > 1); |
|---|
| 1918 | n/a | |
|---|
| 1919 | n/a | ch = CHILD(n, 0); |
|---|
| 1920 | n/a | elt = ast_for_expr(c, ch); |
|---|
| 1921 | n/a | if (!elt) |
|---|
| 1922 | n/a | return NULL; |
|---|
| 1923 | n/a | if (elt->kind == Starred_kind) { |
|---|
| 1924 | n/a | ast_error(c, ch, "iterable unpacking cannot be used in comprehension"); |
|---|
| 1925 | n/a | return NULL; |
|---|
| 1926 | n/a | } |
|---|
| 1927 | n/a | |
|---|
| 1928 | n/a | comps = ast_for_comprehension(c, CHILD(n, 1)); |
|---|
| 1929 | n/a | if (!comps) |
|---|
| 1930 | n/a | return NULL; |
|---|
| 1931 | n/a | |
|---|
| 1932 | n/a | if (type == COMP_GENEXP) |
|---|
| 1933 | n/a | return GeneratorExp(elt, comps, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 1934 | n/a | else if (type == COMP_LISTCOMP) |
|---|
| 1935 | n/a | return ListComp(elt, comps, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 1936 | n/a | else if (type == COMP_SETCOMP) |
|---|
| 1937 | n/a | return SetComp(elt, comps, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 1938 | n/a | else |
|---|
| 1939 | n/a | /* Should never happen */ |
|---|
| 1940 | n/a | return NULL; |
|---|
| 1941 | n/a | } |
|---|
| 1942 | n/a | |
|---|
| 1943 | n/a | /* Fills in the key, value pair corresponding to the dict element. In case |
|---|
| 1944 | n/a | * of an unpacking, key is NULL. *i is advanced by the number of ast |
|---|
| 1945 | n/a | * elements. Iff successful, nonzero is returned. |
|---|
| 1946 | n/a | */ |
|---|
| 1947 | n/a | static int |
|---|
| 1948 | n/a | ast_for_dictelement(struct compiling *c, const node *n, int *i, |
|---|
| 1949 | n/a | expr_ty *key, expr_ty *value) |
|---|
| 1950 | n/a | { |
|---|
| 1951 | n/a | expr_ty expression; |
|---|
| 1952 | n/a | if (TYPE(CHILD(n, *i)) == DOUBLESTAR) { |
|---|
| 1953 | n/a | assert(NCH(n) - *i >= 2); |
|---|
| 1954 | n/a | |
|---|
| 1955 | n/a | expression = ast_for_expr(c, CHILD(n, *i + 1)); |
|---|
| 1956 | n/a | if (!expression) |
|---|
| 1957 | n/a | return 0; |
|---|
| 1958 | n/a | *key = NULL; |
|---|
| 1959 | n/a | *value = expression; |
|---|
| 1960 | n/a | |
|---|
| 1961 | n/a | *i += 2; |
|---|
| 1962 | n/a | } |
|---|
| 1963 | n/a | else { |
|---|
| 1964 | n/a | assert(NCH(n) - *i >= 3); |
|---|
| 1965 | n/a | |
|---|
| 1966 | n/a | expression = ast_for_expr(c, CHILD(n, *i)); |
|---|
| 1967 | n/a | if (!expression) |
|---|
| 1968 | n/a | return 0; |
|---|
| 1969 | n/a | *key = expression; |
|---|
| 1970 | n/a | |
|---|
| 1971 | n/a | REQ(CHILD(n, *i + 1), COLON); |
|---|
| 1972 | n/a | |
|---|
| 1973 | n/a | expression = ast_for_expr(c, CHILD(n, *i + 2)); |
|---|
| 1974 | n/a | if (!expression) |
|---|
| 1975 | n/a | return 0; |
|---|
| 1976 | n/a | *value = expression; |
|---|
| 1977 | n/a | |
|---|
| 1978 | n/a | *i += 3; |
|---|
| 1979 | n/a | } |
|---|
| 1980 | n/a | return 1; |
|---|
| 1981 | n/a | } |
|---|
| 1982 | n/a | |
|---|
| 1983 | n/a | static expr_ty |
|---|
| 1984 | n/a | ast_for_dictcomp(struct compiling *c, const node *n) |
|---|
| 1985 | n/a | { |
|---|
| 1986 | n/a | expr_ty key, value; |
|---|
| 1987 | n/a | asdl_seq *comps; |
|---|
| 1988 | n/a | int i = 0; |
|---|
| 1989 | n/a | |
|---|
| 1990 | n/a | if (!ast_for_dictelement(c, n, &i, &key, &value)) |
|---|
| 1991 | n/a | return NULL; |
|---|
| 1992 | n/a | assert(key); |
|---|
| 1993 | n/a | assert(NCH(n) - i >= 1); |
|---|
| 1994 | n/a | |
|---|
| 1995 | n/a | comps = ast_for_comprehension(c, CHILD(n, i)); |
|---|
| 1996 | n/a | if (!comps) |
|---|
| 1997 | n/a | return NULL; |
|---|
| 1998 | n/a | |
|---|
| 1999 | n/a | return DictComp(key, value, comps, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2000 | n/a | } |
|---|
| 2001 | n/a | |
|---|
| 2002 | n/a | static expr_ty |
|---|
| 2003 | n/a | ast_for_dictdisplay(struct compiling *c, const node *n) |
|---|
| 2004 | n/a | { |
|---|
| 2005 | n/a | int i; |
|---|
| 2006 | n/a | int j; |
|---|
| 2007 | n/a | int size; |
|---|
| 2008 | n/a | asdl_seq *keys, *values; |
|---|
| 2009 | n/a | |
|---|
| 2010 | n/a | size = (NCH(n) + 1) / 3; /* +1 in case no trailing comma */ |
|---|
| 2011 | n/a | keys = _Py_asdl_seq_new(size, c->c_arena); |
|---|
| 2012 | n/a | if (!keys) |
|---|
| 2013 | n/a | return NULL; |
|---|
| 2014 | n/a | |
|---|
| 2015 | n/a | values = _Py_asdl_seq_new(size, c->c_arena); |
|---|
| 2016 | n/a | if (!values) |
|---|
| 2017 | n/a | return NULL; |
|---|
| 2018 | n/a | |
|---|
| 2019 | n/a | j = 0; |
|---|
| 2020 | n/a | for (i = 0; i < NCH(n); i++) { |
|---|
| 2021 | n/a | expr_ty key, value; |
|---|
| 2022 | n/a | |
|---|
| 2023 | n/a | if (!ast_for_dictelement(c, n, &i, &key, &value)) |
|---|
| 2024 | n/a | return NULL; |
|---|
| 2025 | n/a | asdl_seq_SET(keys, j, key); |
|---|
| 2026 | n/a | asdl_seq_SET(values, j, value); |
|---|
| 2027 | n/a | |
|---|
| 2028 | n/a | j++; |
|---|
| 2029 | n/a | } |
|---|
| 2030 | n/a | keys->size = j; |
|---|
| 2031 | n/a | values->size = j; |
|---|
| 2032 | n/a | return Dict(keys, values, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2033 | n/a | } |
|---|
| 2034 | n/a | |
|---|
| 2035 | n/a | static expr_ty |
|---|
| 2036 | n/a | ast_for_genexp(struct compiling *c, const node *n) |
|---|
| 2037 | n/a | { |
|---|
| 2038 | n/a | assert(TYPE(n) == (testlist_comp) || TYPE(n) == (argument)); |
|---|
| 2039 | n/a | return ast_for_itercomp(c, n, COMP_GENEXP); |
|---|
| 2040 | n/a | } |
|---|
| 2041 | n/a | |
|---|
| 2042 | n/a | static expr_ty |
|---|
| 2043 | n/a | ast_for_listcomp(struct compiling *c, const node *n) |
|---|
| 2044 | n/a | { |
|---|
| 2045 | n/a | assert(TYPE(n) == (testlist_comp)); |
|---|
| 2046 | n/a | return ast_for_itercomp(c, n, COMP_LISTCOMP); |
|---|
| 2047 | n/a | } |
|---|
| 2048 | n/a | |
|---|
| 2049 | n/a | static expr_ty |
|---|
| 2050 | n/a | ast_for_setcomp(struct compiling *c, const node *n) |
|---|
| 2051 | n/a | { |
|---|
| 2052 | n/a | assert(TYPE(n) == (dictorsetmaker)); |
|---|
| 2053 | n/a | return ast_for_itercomp(c, n, COMP_SETCOMP); |
|---|
| 2054 | n/a | } |
|---|
| 2055 | n/a | |
|---|
| 2056 | n/a | static expr_ty |
|---|
| 2057 | n/a | ast_for_setdisplay(struct compiling *c, const node *n) |
|---|
| 2058 | n/a | { |
|---|
| 2059 | n/a | int i; |
|---|
| 2060 | n/a | int size; |
|---|
| 2061 | n/a | asdl_seq *elts; |
|---|
| 2062 | n/a | |
|---|
| 2063 | n/a | assert(TYPE(n) == (dictorsetmaker)); |
|---|
| 2064 | n/a | size = (NCH(n) + 1) / 2; /* +1 in case no trailing comma */ |
|---|
| 2065 | n/a | elts = _Py_asdl_seq_new(size, c->c_arena); |
|---|
| 2066 | n/a | if (!elts) |
|---|
| 2067 | n/a | return NULL; |
|---|
| 2068 | n/a | for (i = 0; i < NCH(n); i += 2) { |
|---|
| 2069 | n/a | expr_ty expression; |
|---|
| 2070 | n/a | expression = ast_for_expr(c, CHILD(n, i)); |
|---|
| 2071 | n/a | if (!expression) |
|---|
| 2072 | n/a | return NULL; |
|---|
| 2073 | n/a | asdl_seq_SET(elts, i / 2, expression); |
|---|
| 2074 | n/a | } |
|---|
| 2075 | n/a | return Set(elts, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2076 | n/a | } |
|---|
| 2077 | n/a | |
|---|
| 2078 | n/a | static expr_ty |
|---|
| 2079 | n/a | ast_for_atom(struct compiling *c, const node *n) |
|---|
| 2080 | n/a | { |
|---|
| 2081 | n/a | /* atom: '(' [yield_expr|testlist_comp] ')' | '[' [testlist_comp] ']' |
|---|
| 2082 | n/a | | '{' [dictmaker|testlist_comp] '}' | NAME | NUMBER | STRING+ |
|---|
| 2083 | n/a | | '...' | 'None' | 'True' | 'False' |
|---|
| 2084 | n/a | */ |
|---|
| 2085 | n/a | node *ch = CHILD(n, 0); |
|---|
| 2086 | n/a | |
|---|
| 2087 | n/a | switch (TYPE(ch)) { |
|---|
| 2088 | n/a | case NAME: { |
|---|
| 2089 | n/a | PyObject *name; |
|---|
| 2090 | n/a | const char *s = STR(ch); |
|---|
| 2091 | n/a | size_t len = strlen(s); |
|---|
| 2092 | n/a | if (len >= 4 && len <= 5) { |
|---|
| 2093 | n/a | if (!strcmp(s, "None")) |
|---|
| 2094 | n/a | return NameConstant(Py_None, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2095 | n/a | if (!strcmp(s, "True")) |
|---|
| 2096 | n/a | return NameConstant(Py_True, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2097 | n/a | if (!strcmp(s, "False")) |
|---|
| 2098 | n/a | return NameConstant(Py_False, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2099 | n/a | } |
|---|
| 2100 | n/a | name = new_identifier(s, c); |
|---|
| 2101 | n/a | if (!name) |
|---|
| 2102 | n/a | return NULL; |
|---|
| 2103 | n/a | /* All names start in Load context, but may later be changed. */ |
|---|
| 2104 | n/a | return Name(name, Load, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2105 | n/a | } |
|---|
| 2106 | n/a | case STRING: { |
|---|
| 2107 | n/a | expr_ty str = parsestrplus(c, n); |
|---|
| 2108 | n/a | if (!str) { |
|---|
| 2109 | n/a | const char *errtype = NULL; |
|---|
| 2110 | n/a | if (PyErr_ExceptionMatches(PyExc_UnicodeError)) |
|---|
| 2111 | n/a | errtype = "unicode error"; |
|---|
| 2112 | n/a | else if (PyErr_ExceptionMatches(PyExc_ValueError)) |
|---|
| 2113 | n/a | errtype = "value error"; |
|---|
| 2114 | n/a | if (errtype) { |
|---|
| 2115 | n/a | char buf[128]; |
|---|
| 2116 | n/a | const char *s = NULL; |
|---|
| 2117 | n/a | PyObject *type, *value, *tback, *errstr; |
|---|
| 2118 | n/a | PyErr_Fetch(&type, &value, &tback); |
|---|
| 2119 | n/a | errstr = PyObject_Str(value); |
|---|
| 2120 | n/a | if (errstr) |
|---|
| 2121 | n/a | s = PyUnicode_AsUTF8(errstr); |
|---|
| 2122 | n/a | if (s) { |
|---|
| 2123 | n/a | PyOS_snprintf(buf, sizeof(buf), "(%s) %s", errtype, s); |
|---|
| 2124 | n/a | } else { |
|---|
| 2125 | n/a | PyErr_Clear(); |
|---|
| 2126 | n/a | PyOS_snprintf(buf, sizeof(buf), "(%s) unknown error", errtype); |
|---|
| 2127 | n/a | } |
|---|
| 2128 | n/a | Py_XDECREF(errstr); |
|---|
| 2129 | n/a | ast_error(c, n, buf); |
|---|
| 2130 | n/a | Py_DECREF(type); |
|---|
| 2131 | n/a | Py_XDECREF(value); |
|---|
| 2132 | n/a | Py_XDECREF(tback); |
|---|
| 2133 | n/a | } |
|---|
| 2134 | n/a | return NULL; |
|---|
| 2135 | n/a | } |
|---|
| 2136 | n/a | return str; |
|---|
| 2137 | n/a | } |
|---|
| 2138 | n/a | case NUMBER: { |
|---|
| 2139 | n/a | PyObject *pynum = parsenumber(c, STR(ch)); |
|---|
| 2140 | n/a | if (!pynum) |
|---|
| 2141 | n/a | return NULL; |
|---|
| 2142 | n/a | |
|---|
| 2143 | n/a | if (PyArena_AddPyObject(c->c_arena, pynum) < 0) { |
|---|
| 2144 | n/a | Py_DECREF(pynum); |
|---|
| 2145 | n/a | return NULL; |
|---|
| 2146 | n/a | } |
|---|
| 2147 | n/a | return Num(pynum, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2148 | n/a | } |
|---|
| 2149 | n/a | case ELLIPSIS: /* Ellipsis */ |
|---|
| 2150 | n/a | return Ellipsis(LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2151 | n/a | case LPAR: /* some parenthesized expressions */ |
|---|
| 2152 | n/a | ch = CHILD(n, 1); |
|---|
| 2153 | n/a | |
|---|
| 2154 | n/a | if (TYPE(ch) == RPAR) |
|---|
| 2155 | n/a | return Tuple(NULL, Load, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2156 | n/a | |
|---|
| 2157 | n/a | if (TYPE(ch) == yield_expr) |
|---|
| 2158 | n/a | return ast_for_expr(c, ch); |
|---|
| 2159 | n/a | |
|---|
| 2160 | n/a | /* testlist_comp: test ( comp_for | (',' test)* [','] ) */ |
|---|
| 2161 | n/a | if ((NCH(ch) > 1) && (TYPE(CHILD(ch, 1)) == comp_for)) |
|---|
| 2162 | n/a | return ast_for_genexp(c, ch); |
|---|
| 2163 | n/a | |
|---|
| 2164 | n/a | return ast_for_testlist(c, ch); |
|---|
| 2165 | n/a | case LSQB: /* list (or list comprehension) */ |
|---|
| 2166 | n/a | ch = CHILD(n, 1); |
|---|
| 2167 | n/a | |
|---|
| 2168 | n/a | if (TYPE(ch) == RSQB) |
|---|
| 2169 | n/a | return List(NULL, Load, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2170 | n/a | |
|---|
| 2171 | n/a | REQ(ch, testlist_comp); |
|---|
| 2172 | n/a | if (NCH(ch) == 1 || TYPE(CHILD(ch, 1)) == COMMA) { |
|---|
| 2173 | n/a | asdl_seq *elts = seq_for_testlist(c, ch); |
|---|
| 2174 | n/a | if (!elts) |
|---|
| 2175 | n/a | return NULL; |
|---|
| 2176 | n/a | |
|---|
| 2177 | n/a | return List(elts, Load, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2178 | n/a | } |
|---|
| 2179 | n/a | else |
|---|
| 2180 | n/a | return ast_for_listcomp(c, ch); |
|---|
| 2181 | n/a | case LBRACE: { |
|---|
| 2182 | n/a | /* dictorsetmaker: ( ((test ':' test | '**' test) |
|---|
| 2183 | n/a | * (comp_for | (',' (test ':' test | '**' test))* [','])) | |
|---|
| 2184 | n/a | * ((test | '*' test) |
|---|
| 2185 | n/a | * (comp_for | (',' (test | '*' test))* [','])) ) */ |
|---|
| 2186 | n/a | expr_ty res; |
|---|
| 2187 | n/a | ch = CHILD(n, 1); |
|---|
| 2188 | n/a | if (TYPE(ch) == RBRACE) { |
|---|
| 2189 | n/a | /* It's an empty dict. */ |
|---|
| 2190 | n/a | return Dict(NULL, NULL, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2191 | n/a | } |
|---|
| 2192 | n/a | else { |
|---|
| 2193 | n/a | int is_dict = (TYPE(CHILD(ch, 0)) == DOUBLESTAR); |
|---|
| 2194 | n/a | if (NCH(ch) == 1 || |
|---|
| 2195 | n/a | (NCH(ch) > 1 && |
|---|
| 2196 | n/a | TYPE(CHILD(ch, 1)) == COMMA)) { |
|---|
| 2197 | n/a | /* It's a set display. */ |
|---|
| 2198 | n/a | res = ast_for_setdisplay(c, ch); |
|---|
| 2199 | n/a | } |
|---|
| 2200 | n/a | else if (NCH(ch) > 1 && |
|---|
| 2201 | n/a | TYPE(CHILD(ch, 1)) == comp_for) { |
|---|
| 2202 | n/a | /* It's a set comprehension. */ |
|---|
| 2203 | n/a | res = ast_for_setcomp(c, ch); |
|---|
| 2204 | n/a | } |
|---|
| 2205 | n/a | else if (NCH(ch) > 3 - is_dict && |
|---|
| 2206 | n/a | TYPE(CHILD(ch, 3 - is_dict)) == comp_for) { |
|---|
| 2207 | n/a | /* It's a dictionary comprehension. */ |
|---|
| 2208 | n/a | if (is_dict) { |
|---|
| 2209 | n/a | ast_error(c, n, "dict unpacking cannot be used in " |
|---|
| 2210 | n/a | "dict comprehension"); |
|---|
| 2211 | n/a | return NULL; |
|---|
| 2212 | n/a | } |
|---|
| 2213 | n/a | res = ast_for_dictcomp(c, ch); |
|---|
| 2214 | n/a | } |
|---|
| 2215 | n/a | else { |
|---|
| 2216 | n/a | /* It's a dictionary display. */ |
|---|
| 2217 | n/a | res = ast_for_dictdisplay(c, ch); |
|---|
| 2218 | n/a | } |
|---|
| 2219 | n/a | if (res) { |
|---|
| 2220 | n/a | res->lineno = LINENO(n); |
|---|
| 2221 | n/a | res->col_offset = n->n_col_offset; |
|---|
| 2222 | n/a | } |
|---|
| 2223 | n/a | return res; |
|---|
| 2224 | n/a | } |
|---|
| 2225 | n/a | } |
|---|
| 2226 | n/a | default: |
|---|
| 2227 | n/a | PyErr_Format(PyExc_SystemError, "unhandled atom %d", TYPE(ch)); |
|---|
| 2228 | n/a | return NULL; |
|---|
| 2229 | n/a | } |
|---|
| 2230 | n/a | } |
|---|
| 2231 | n/a | |
|---|
| 2232 | n/a | static slice_ty |
|---|
| 2233 | n/a | ast_for_slice(struct compiling *c, const node *n) |
|---|
| 2234 | n/a | { |
|---|
| 2235 | n/a | node *ch; |
|---|
| 2236 | n/a | expr_ty lower = NULL, upper = NULL, step = NULL; |
|---|
| 2237 | n/a | |
|---|
| 2238 | n/a | REQ(n, subscript); |
|---|
| 2239 | n/a | |
|---|
| 2240 | n/a | /* |
|---|
| 2241 | n/a | subscript: test | [test] ':' [test] [sliceop] |
|---|
| 2242 | n/a | sliceop: ':' [test] |
|---|
| 2243 | n/a | */ |
|---|
| 2244 | n/a | ch = CHILD(n, 0); |
|---|
| 2245 | n/a | if (NCH(n) == 1 && TYPE(ch) == test) { |
|---|
| 2246 | n/a | /* 'step' variable hold no significance in terms of being used over |
|---|
| 2247 | n/a | other vars */ |
|---|
| 2248 | n/a | step = ast_for_expr(c, ch); |
|---|
| 2249 | n/a | if (!step) |
|---|
| 2250 | n/a | return NULL; |
|---|
| 2251 | n/a | |
|---|
| 2252 | n/a | return Index(step, c->c_arena); |
|---|
| 2253 | n/a | } |
|---|
| 2254 | n/a | |
|---|
| 2255 | n/a | if (TYPE(ch) == test) { |
|---|
| 2256 | n/a | lower = ast_for_expr(c, ch); |
|---|
| 2257 | n/a | if (!lower) |
|---|
| 2258 | n/a | return NULL; |
|---|
| 2259 | n/a | } |
|---|
| 2260 | n/a | |
|---|
| 2261 | n/a | /* If there's an upper bound it's in the second or third position. */ |
|---|
| 2262 | n/a | if (TYPE(ch) == COLON) { |
|---|
| 2263 | n/a | if (NCH(n) > 1) { |
|---|
| 2264 | n/a | node *n2 = CHILD(n, 1); |
|---|
| 2265 | n/a | |
|---|
| 2266 | n/a | if (TYPE(n2) == test) { |
|---|
| 2267 | n/a | upper = ast_for_expr(c, n2); |
|---|
| 2268 | n/a | if (!upper) |
|---|
| 2269 | n/a | return NULL; |
|---|
| 2270 | n/a | } |
|---|
| 2271 | n/a | } |
|---|
| 2272 | n/a | } else if (NCH(n) > 2) { |
|---|
| 2273 | n/a | node *n2 = CHILD(n, 2); |
|---|
| 2274 | n/a | |
|---|
| 2275 | n/a | if (TYPE(n2) == test) { |
|---|
| 2276 | n/a | upper = ast_for_expr(c, n2); |
|---|
| 2277 | n/a | if (!upper) |
|---|
| 2278 | n/a | return NULL; |
|---|
| 2279 | n/a | } |
|---|
| 2280 | n/a | } |
|---|
| 2281 | n/a | |
|---|
| 2282 | n/a | ch = CHILD(n, NCH(n) - 1); |
|---|
| 2283 | n/a | if (TYPE(ch) == sliceop) { |
|---|
| 2284 | n/a | if (NCH(ch) != 1) { |
|---|
| 2285 | n/a | ch = CHILD(ch, 1); |
|---|
| 2286 | n/a | if (TYPE(ch) == test) { |
|---|
| 2287 | n/a | step = ast_for_expr(c, ch); |
|---|
| 2288 | n/a | if (!step) |
|---|
| 2289 | n/a | return NULL; |
|---|
| 2290 | n/a | } |
|---|
| 2291 | n/a | } |
|---|
| 2292 | n/a | } |
|---|
| 2293 | n/a | |
|---|
| 2294 | n/a | return Slice(lower, upper, step, c->c_arena); |
|---|
| 2295 | n/a | } |
|---|
| 2296 | n/a | |
|---|
| 2297 | n/a | static expr_ty |
|---|
| 2298 | n/a | ast_for_binop(struct compiling *c, const node *n) |
|---|
| 2299 | n/a | { |
|---|
| 2300 | n/a | /* Must account for a sequence of expressions. |
|---|
| 2301 | n/a | How should A op B op C by represented? |
|---|
| 2302 | n/a | BinOp(BinOp(A, op, B), op, C). |
|---|
| 2303 | n/a | */ |
|---|
| 2304 | n/a | |
|---|
| 2305 | n/a | int i, nops; |
|---|
| 2306 | n/a | expr_ty expr1, expr2, result; |
|---|
| 2307 | n/a | operator_ty newoperator; |
|---|
| 2308 | n/a | |
|---|
| 2309 | n/a | expr1 = ast_for_expr(c, CHILD(n, 0)); |
|---|
| 2310 | n/a | if (!expr1) |
|---|
| 2311 | n/a | return NULL; |
|---|
| 2312 | n/a | |
|---|
| 2313 | n/a | expr2 = ast_for_expr(c, CHILD(n, 2)); |
|---|
| 2314 | n/a | if (!expr2) |
|---|
| 2315 | n/a | return NULL; |
|---|
| 2316 | n/a | |
|---|
| 2317 | n/a | newoperator = get_operator(CHILD(n, 1)); |
|---|
| 2318 | n/a | if (!newoperator) |
|---|
| 2319 | n/a | return NULL; |
|---|
| 2320 | n/a | |
|---|
| 2321 | n/a | result = BinOp(expr1, newoperator, expr2, LINENO(n), n->n_col_offset, |
|---|
| 2322 | n/a | c->c_arena); |
|---|
| 2323 | n/a | if (!result) |
|---|
| 2324 | n/a | return NULL; |
|---|
| 2325 | n/a | |
|---|
| 2326 | n/a | nops = (NCH(n) - 1) / 2; |
|---|
| 2327 | n/a | for (i = 1; i < nops; i++) { |
|---|
| 2328 | n/a | expr_ty tmp_result, tmp; |
|---|
| 2329 | n/a | const node* next_oper = CHILD(n, i * 2 + 1); |
|---|
| 2330 | n/a | |
|---|
| 2331 | n/a | newoperator = get_operator(next_oper); |
|---|
| 2332 | n/a | if (!newoperator) |
|---|
| 2333 | n/a | return NULL; |
|---|
| 2334 | n/a | |
|---|
| 2335 | n/a | tmp = ast_for_expr(c, CHILD(n, i * 2 + 2)); |
|---|
| 2336 | n/a | if (!tmp) |
|---|
| 2337 | n/a | return NULL; |
|---|
| 2338 | n/a | |
|---|
| 2339 | n/a | tmp_result = BinOp(result, newoperator, tmp, |
|---|
| 2340 | n/a | LINENO(next_oper), next_oper->n_col_offset, |
|---|
| 2341 | n/a | c->c_arena); |
|---|
| 2342 | n/a | if (!tmp_result) |
|---|
| 2343 | n/a | return NULL; |
|---|
| 2344 | n/a | result = tmp_result; |
|---|
| 2345 | n/a | } |
|---|
| 2346 | n/a | return result; |
|---|
| 2347 | n/a | } |
|---|
| 2348 | n/a | |
|---|
| 2349 | n/a | static expr_ty |
|---|
| 2350 | n/a | ast_for_trailer(struct compiling *c, const node *n, expr_ty left_expr) |
|---|
| 2351 | n/a | { |
|---|
| 2352 | n/a | /* trailer: '(' [arglist] ')' | '[' subscriptlist ']' | '.' NAME |
|---|
| 2353 | n/a | subscriptlist: subscript (',' subscript)* [','] |
|---|
| 2354 | n/a | subscript: '.' '.' '.' | test | [test] ':' [test] [sliceop] |
|---|
| 2355 | n/a | */ |
|---|
| 2356 | n/a | REQ(n, trailer); |
|---|
| 2357 | n/a | if (TYPE(CHILD(n, 0)) == LPAR) { |
|---|
| 2358 | n/a | if (NCH(n) == 2) |
|---|
| 2359 | n/a | return Call(left_expr, NULL, NULL, LINENO(n), |
|---|
| 2360 | n/a | n->n_col_offset, c->c_arena); |
|---|
| 2361 | n/a | else |
|---|
| 2362 | n/a | return ast_for_call(c, CHILD(n, 1), left_expr); |
|---|
| 2363 | n/a | } |
|---|
| 2364 | n/a | else if (TYPE(CHILD(n, 0)) == DOT) { |
|---|
| 2365 | n/a | PyObject *attr_id = NEW_IDENTIFIER(CHILD(n, 1)); |
|---|
| 2366 | n/a | if (!attr_id) |
|---|
| 2367 | n/a | return NULL; |
|---|
| 2368 | n/a | return Attribute(left_expr, attr_id, Load, |
|---|
| 2369 | n/a | LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2370 | n/a | } |
|---|
| 2371 | n/a | else { |
|---|
| 2372 | n/a | REQ(CHILD(n, 0), LSQB); |
|---|
| 2373 | n/a | REQ(CHILD(n, 2), RSQB); |
|---|
| 2374 | n/a | n = CHILD(n, 1); |
|---|
| 2375 | n/a | if (NCH(n) == 1) { |
|---|
| 2376 | n/a | slice_ty slc = ast_for_slice(c, CHILD(n, 0)); |
|---|
| 2377 | n/a | if (!slc) |
|---|
| 2378 | n/a | return NULL; |
|---|
| 2379 | n/a | return Subscript(left_expr, slc, Load, LINENO(n), n->n_col_offset, |
|---|
| 2380 | n/a | c->c_arena); |
|---|
| 2381 | n/a | } |
|---|
| 2382 | n/a | else { |
|---|
| 2383 | n/a | /* The grammar is ambiguous here. The ambiguity is resolved |
|---|
| 2384 | n/a | by treating the sequence as a tuple literal if there are |
|---|
| 2385 | n/a | no slice features. |
|---|
| 2386 | n/a | */ |
|---|
| 2387 | n/a | int j; |
|---|
| 2388 | n/a | slice_ty slc; |
|---|
| 2389 | n/a | expr_ty e; |
|---|
| 2390 | n/a | int simple = 1; |
|---|
| 2391 | n/a | asdl_seq *slices, *elts; |
|---|
| 2392 | n/a | slices = _Py_asdl_seq_new((NCH(n) + 1) / 2, c->c_arena); |
|---|
| 2393 | n/a | if (!slices) |
|---|
| 2394 | n/a | return NULL; |
|---|
| 2395 | n/a | for (j = 0; j < NCH(n); j += 2) { |
|---|
| 2396 | n/a | slc = ast_for_slice(c, CHILD(n, j)); |
|---|
| 2397 | n/a | if (!slc) |
|---|
| 2398 | n/a | return NULL; |
|---|
| 2399 | n/a | if (slc->kind != Index_kind) |
|---|
| 2400 | n/a | simple = 0; |
|---|
| 2401 | n/a | asdl_seq_SET(slices, j / 2, slc); |
|---|
| 2402 | n/a | } |
|---|
| 2403 | n/a | if (!simple) { |
|---|
| 2404 | n/a | return Subscript(left_expr, ExtSlice(slices, c->c_arena), |
|---|
| 2405 | n/a | Load, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2406 | n/a | } |
|---|
| 2407 | n/a | /* extract Index values and put them in a Tuple */ |
|---|
| 2408 | n/a | elts = _Py_asdl_seq_new(asdl_seq_LEN(slices), c->c_arena); |
|---|
| 2409 | n/a | if (!elts) |
|---|
| 2410 | n/a | return NULL; |
|---|
| 2411 | n/a | for (j = 0; j < asdl_seq_LEN(slices); ++j) { |
|---|
| 2412 | n/a | slc = (slice_ty)asdl_seq_GET(slices, j); |
|---|
| 2413 | n/a | assert(slc->kind == Index_kind && slc->v.Index.value); |
|---|
| 2414 | n/a | asdl_seq_SET(elts, j, slc->v.Index.value); |
|---|
| 2415 | n/a | } |
|---|
| 2416 | n/a | e = Tuple(elts, Load, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2417 | n/a | if (!e) |
|---|
| 2418 | n/a | return NULL; |
|---|
| 2419 | n/a | return Subscript(left_expr, Index(e, c->c_arena), |
|---|
| 2420 | n/a | Load, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2421 | n/a | } |
|---|
| 2422 | n/a | } |
|---|
| 2423 | n/a | } |
|---|
| 2424 | n/a | |
|---|
| 2425 | n/a | static expr_ty |
|---|
| 2426 | n/a | ast_for_factor(struct compiling *c, const node *n) |
|---|
| 2427 | n/a | { |
|---|
| 2428 | n/a | expr_ty expression; |
|---|
| 2429 | n/a | |
|---|
| 2430 | n/a | expression = ast_for_expr(c, CHILD(n, 1)); |
|---|
| 2431 | n/a | if (!expression) |
|---|
| 2432 | n/a | return NULL; |
|---|
| 2433 | n/a | |
|---|
| 2434 | n/a | switch (TYPE(CHILD(n, 0))) { |
|---|
| 2435 | n/a | case PLUS: |
|---|
| 2436 | n/a | return UnaryOp(UAdd, expression, LINENO(n), n->n_col_offset, |
|---|
| 2437 | n/a | c->c_arena); |
|---|
| 2438 | n/a | case MINUS: |
|---|
| 2439 | n/a | return UnaryOp(USub, expression, LINENO(n), n->n_col_offset, |
|---|
| 2440 | n/a | c->c_arena); |
|---|
| 2441 | n/a | case TILDE: |
|---|
| 2442 | n/a | return UnaryOp(Invert, expression, LINENO(n), |
|---|
| 2443 | n/a | n->n_col_offset, c->c_arena); |
|---|
| 2444 | n/a | } |
|---|
| 2445 | n/a | PyErr_Format(PyExc_SystemError, "unhandled factor: %d", |
|---|
| 2446 | n/a | TYPE(CHILD(n, 0))); |
|---|
| 2447 | n/a | return NULL; |
|---|
| 2448 | n/a | } |
|---|
| 2449 | n/a | |
|---|
| 2450 | n/a | static expr_ty |
|---|
| 2451 | n/a | ast_for_atom_expr(struct compiling *c, const node *n) |
|---|
| 2452 | n/a | { |
|---|
| 2453 | n/a | int i, nch, start = 0; |
|---|
| 2454 | n/a | expr_ty e, tmp; |
|---|
| 2455 | n/a | |
|---|
| 2456 | n/a | REQ(n, atom_expr); |
|---|
| 2457 | n/a | nch = NCH(n); |
|---|
| 2458 | n/a | |
|---|
| 2459 | n/a | if (TYPE(CHILD(n, 0)) == AWAIT) { |
|---|
| 2460 | n/a | start = 1; |
|---|
| 2461 | n/a | assert(nch > 1); |
|---|
| 2462 | n/a | } |
|---|
| 2463 | n/a | |
|---|
| 2464 | n/a | e = ast_for_atom(c, CHILD(n, start)); |
|---|
| 2465 | n/a | if (!e) |
|---|
| 2466 | n/a | return NULL; |
|---|
| 2467 | n/a | if (nch == 1) |
|---|
| 2468 | n/a | return e; |
|---|
| 2469 | n/a | if (start && nch == 2) { |
|---|
| 2470 | n/a | return Await(e, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2471 | n/a | } |
|---|
| 2472 | n/a | |
|---|
| 2473 | n/a | for (i = start + 1; i < nch; i++) { |
|---|
| 2474 | n/a | node *ch = CHILD(n, i); |
|---|
| 2475 | n/a | if (TYPE(ch) != trailer) |
|---|
| 2476 | n/a | break; |
|---|
| 2477 | n/a | tmp = ast_for_trailer(c, ch, e); |
|---|
| 2478 | n/a | if (!tmp) |
|---|
| 2479 | n/a | return NULL; |
|---|
| 2480 | n/a | tmp->lineno = e->lineno; |
|---|
| 2481 | n/a | tmp->col_offset = e->col_offset; |
|---|
| 2482 | n/a | e = tmp; |
|---|
| 2483 | n/a | } |
|---|
| 2484 | n/a | |
|---|
| 2485 | n/a | if (start) { |
|---|
| 2486 | n/a | /* there was an AWAIT */ |
|---|
| 2487 | n/a | return Await(e, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2488 | n/a | } |
|---|
| 2489 | n/a | else { |
|---|
| 2490 | n/a | return e; |
|---|
| 2491 | n/a | } |
|---|
| 2492 | n/a | } |
|---|
| 2493 | n/a | |
|---|
| 2494 | n/a | static expr_ty |
|---|
| 2495 | n/a | ast_for_power(struct compiling *c, const node *n) |
|---|
| 2496 | n/a | { |
|---|
| 2497 | n/a | /* power: atom trailer* ('**' factor)* |
|---|
| 2498 | n/a | */ |
|---|
| 2499 | n/a | expr_ty e; |
|---|
| 2500 | n/a | REQ(n, power); |
|---|
| 2501 | n/a | e = ast_for_atom_expr(c, CHILD(n, 0)); |
|---|
| 2502 | n/a | if (!e) |
|---|
| 2503 | n/a | return NULL; |
|---|
| 2504 | n/a | if (NCH(n) == 1) |
|---|
| 2505 | n/a | return e; |
|---|
| 2506 | n/a | if (TYPE(CHILD(n, NCH(n) - 1)) == factor) { |
|---|
| 2507 | n/a | expr_ty f = ast_for_expr(c, CHILD(n, NCH(n) - 1)); |
|---|
| 2508 | n/a | if (!f) |
|---|
| 2509 | n/a | return NULL; |
|---|
| 2510 | n/a | e = BinOp(e, Pow, f, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2511 | n/a | } |
|---|
| 2512 | n/a | return e; |
|---|
| 2513 | n/a | } |
|---|
| 2514 | n/a | |
|---|
| 2515 | n/a | static expr_ty |
|---|
| 2516 | n/a | ast_for_starred(struct compiling *c, const node *n) |
|---|
| 2517 | n/a | { |
|---|
| 2518 | n/a | expr_ty tmp; |
|---|
| 2519 | n/a | REQ(n, star_expr); |
|---|
| 2520 | n/a | |
|---|
| 2521 | n/a | tmp = ast_for_expr(c, CHILD(n, 1)); |
|---|
| 2522 | n/a | if (!tmp) |
|---|
| 2523 | n/a | return NULL; |
|---|
| 2524 | n/a | |
|---|
| 2525 | n/a | /* The Load context is changed later. */ |
|---|
| 2526 | n/a | return Starred(tmp, Load, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2527 | n/a | } |
|---|
| 2528 | n/a | |
|---|
| 2529 | n/a | |
|---|
| 2530 | n/a | /* Do not name a variable 'expr'! Will cause a compile error. |
|---|
| 2531 | n/a | */ |
|---|
| 2532 | n/a | |
|---|
| 2533 | n/a | static expr_ty |
|---|
| 2534 | n/a | ast_for_expr(struct compiling *c, const node *n) |
|---|
| 2535 | n/a | { |
|---|
| 2536 | n/a | /* handle the full range of simple expressions |
|---|
| 2537 | n/a | test: or_test ['if' or_test 'else' test] | lambdef |
|---|
| 2538 | n/a | test_nocond: or_test | lambdef_nocond |
|---|
| 2539 | n/a | or_test: and_test ('or' and_test)* |
|---|
| 2540 | n/a | and_test: not_test ('and' not_test)* |
|---|
| 2541 | n/a | not_test: 'not' not_test | comparison |
|---|
| 2542 | n/a | comparison: expr (comp_op expr)* |
|---|
| 2543 | n/a | expr: xor_expr ('|' xor_expr)* |
|---|
| 2544 | n/a | xor_expr: and_expr ('^' and_expr)* |
|---|
| 2545 | n/a | and_expr: shift_expr ('&' shift_expr)* |
|---|
| 2546 | n/a | shift_expr: arith_expr (('<<'|'>>') arith_expr)* |
|---|
| 2547 | n/a | arith_expr: term (('+'|'-') term)* |
|---|
| 2548 | n/a | term: factor (('*'|'@'|'/'|'%'|'//') factor)* |
|---|
| 2549 | n/a | factor: ('+'|'-'|'~') factor | power |
|---|
| 2550 | n/a | power: atom_expr ['**' factor] |
|---|
| 2551 | n/a | atom_expr: [AWAIT] atom trailer* |
|---|
| 2552 | n/a | yield_expr: 'yield' [yield_arg] |
|---|
| 2553 | n/a | */ |
|---|
| 2554 | n/a | |
|---|
| 2555 | n/a | asdl_seq *seq; |
|---|
| 2556 | n/a | int i; |
|---|
| 2557 | n/a | |
|---|
| 2558 | n/a | loop: |
|---|
| 2559 | n/a | switch (TYPE(n)) { |
|---|
| 2560 | n/a | case test: |
|---|
| 2561 | n/a | case test_nocond: |
|---|
| 2562 | n/a | if (TYPE(CHILD(n, 0)) == lambdef || |
|---|
| 2563 | n/a | TYPE(CHILD(n, 0)) == lambdef_nocond) |
|---|
| 2564 | n/a | return ast_for_lambdef(c, CHILD(n, 0)); |
|---|
| 2565 | n/a | else if (NCH(n) > 1) |
|---|
| 2566 | n/a | return ast_for_ifexpr(c, n); |
|---|
| 2567 | n/a | /* Fallthrough */ |
|---|
| 2568 | n/a | case or_test: |
|---|
| 2569 | n/a | case and_test: |
|---|
| 2570 | n/a | if (NCH(n) == 1) { |
|---|
| 2571 | n/a | n = CHILD(n, 0); |
|---|
| 2572 | n/a | goto loop; |
|---|
| 2573 | n/a | } |
|---|
| 2574 | n/a | seq = _Py_asdl_seq_new((NCH(n) + 1) / 2, c->c_arena); |
|---|
| 2575 | n/a | if (!seq) |
|---|
| 2576 | n/a | return NULL; |
|---|
| 2577 | n/a | for (i = 0; i < NCH(n); i += 2) { |
|---|
| 2578 | n/a | expr_ty e = ast_for_expr(c, CHILD(n, i)); |
|---|
| 2579 | n/a | if (!e) |
|---|
| 2580 | n/a | return NULL; |
|---|
| 2581 | n/a | asdl_seq_SET(seq, i / 2, e); |
|---|
| 2582 | n/a | } |
|---|
| 2583 | n/a | if (!strcmp(STR(CHILD(n, 1)), "and")) |
|---|
| 2584 | n/a | return BoolOp(And, seq, LINENO(n), n->n_col_offset, |
|---|
| 2585 | n/a | c->c_arena); |
|---|
| 2586 | n/a | assert(!strcmp(STR(CHILD(n, 1)), "or")); |
|---|
| 2587 | n/a | return BoolOp(Or, seq, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2588 | n/a | case not_test: |
|---|
| 2589 | n/a | if (NCH(n) == 1) { |
|---|
| 2590 | n/a | n = CHILD(n, 0); |
|---|
| 2591 | n/a | goto loop; |
|---|
| 2592 | n/a | } |
|---|
| 2593 | n/a | else { |
|---|
| 2594 | n/a | expr_ty expression = ast_for_expr(c, CHILD(n, 1)); |
|---|
| 2595 | n/a | if (!expression) |
|---|
| 2596 | n/a | return NULL; |
|---|
| 2597 | n/a | |
|---|
| 2598 | n/a | return UnaryOp(Not, expression, LINENO(n), n->n_col_offset, |
|---|
| 2599 | n/a | c->c_arena); |
|---|
| 2600 | n/a | } |
|---|
| 2601 | n/a | case comparison: |
|---|
| 2602 | n/a | if (NCH(n) == 1) { |
|---|
| 2603 | n/a | n = CHILD(n, 0); |
|---|
| 2604 | n/a | goto loop; |
|---|
| 2605 | n/a | } |
|---|
| 2606 | n/a | else { |
|---|
| 2607 | n/a | expr_ty expression; |
|---|
| 2608 | n/a | asdl_int_seq *ops; |
|---|
| 2609 | n/a | asdl_seq *cmps; |
|---|
| 2610 | n/a | ops = _Py_asdl_int_seq_new(NCH(n) / 2, c->c_arena); |
|---|
| 2611 | n/a | if (!ops) |
|---|
| 2612 | n/a | return NULL; |
|---|
| 2613 | n/a | cmps = _Py_asdl_seq_new(NCH(n) / 2, c->c_arena); |
|---|
| 2614 | n/a | if (!cmps) { |
|---|
| 2615 | n/a | return NULL; |
|---|
| 2616 | n/a | } |
|---|
| 2617 | n/a | for (i = 1; i < NCH(n); i += 2) { |
|---|
| 2618 | n/a | cmpop_ty newoperator; |
|---|
| 2619 | n/a | |
|---|
| 2620 | n/a | newoperator = ast_for_comp_op(c, CHILD(n, i)); |
|---|
| 2621 | n/a | if (!newoperator) { |
|---|
| 2622 | n/a | return NULL; |
|---|
| 2623 | n/a | } |
|---|
| 2624 | n/a | |
|---|
| 2625 | n/a | expression = ast_for_expr(c, CHILD(n, i + 1)); |
|---|
| 2626 | n/a | if (!expression) { |
|---|
| 2627 | n/a | return NULL; |
|---|
| 2628 | n/a | } |
|---|
| 2629 | n/a | |
|---|
| 2630 | n/a | asdl_seq_SET(ops, i / 2, newoperator); |
|---|
| 2631 | n/a | asdl_seq_SET(cmps, i / 2, expression); |
|---|
| 2632 | n/a | } |
|---|
| 2633 | n/a | expression = ast_for_expr(c, CHILD(n, 0)); |
|---|
| 2634 | n/a | if (!expression) { |
|---|
| 2635 | n/a | return NULL; |
|---|
| 2636 | n/a | } |
|---|
| 2637 | n/a | |
|---|
| 2638 | n/a | return Compare(expression, ops, cmps, LINENO(n), |
|---|
| 2639 | n/a | n->n_col_offset, c->c_arena); |
|---|
| 2640 | n/a | } |
|---|
| 2641 | n/a | break; |
|---|
| 2642 | n/a | |
|---|
| 2643 | n/a | case star_expr: |
|---|
| 2644 | n/a | return ast_for_starred(c, n); |
|---|
| 2645 | n/a | /* The next five cases all handle BinOps. The main body of code |
|---|
| 2646 | n/a | is the same in each case, but the switch turned inside out to |
|---|
| 2647 | n/a | reuse the code for each type of operator. |
|---|
| 2648 | n/a | */ |
|---|
| 2649 | n/a | case expr: |
|---|
| 2650 | n/a | case xor_expr: |
|---|
| 2651 | n/a | case and_expr: |
|---|
| 2652 | n/a | case shift_expr: |
|---|
| 2653 | n/a | case arith_expr: |
|---|
| 2654 | n/a | case term: |
|---|
| 2655 | n/a | if (NCH(n) == 1) { |
|---|
| 2656 | n/a | n = CHILD(n, 0); |
|---|
| 2657 | n/a | goto loop; |
|---|
| 2658 | n/a | } |
|---|
| 2659 | n/a | return ast_for_binop(c, n); |
|---|
| 2660 | n/a | case yield_expr: { |
|---|
| 2661 | n/a | node *an = NULL; |
|---|
| 2662 | n/a | node *en = NULL; |
|---|
| 2663 | n/a | int is_from = 0; |
|---|
| 2664 | n/a | expr_ty exp = NULL; |
|---|
| 2665 | n/a | if (NCH(n) > 1) |
|---|
| 2666 | n/a | an = CHILD(n, 1); /* yield_arg */ |
|---|
| 2667 | n/a | if (an) { |
|---|
| 2668 | n/a | en = CHILD(an, NCH(an) - 1); |
|---|
| 2669 | n/a | if (NCH(an) == 2) { |
|---|
| 2670 | n/a | is_from = 1; |
|---|
| 2671 | n/a | exp = ast_for_expr(c, en); |
|---|
| 2672 | n/a | } |
|---|
| 2673 | n/a | else |
|---|
| 2674 | n/a | exp = ast_for_testlist(c, en); |
|---|
| 2675 | n/a | if (!exp) |
|---|
| 2676 | n/a | return NULL; |
|---|
| 2677 | n/a | } |
|---|
| 2678 | n/a | if (is_from) |
|---|
| 2679 | n/a | return YieldFrom(exp, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2680 | n/a | return Yield(exp, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2681 | n/a | } |
|---|
| 2682 | n/a | case factor: |
|---|
| 2683 | n/a | if (NCH(n) == 1) { |
|---|
| 2684 | n/a | n = CHILD(n, 0); |
|---|
| 2685 | n/a | goto loop; |
|---|
| 2686 | n/a | } |
|---|
| 2687 | n/a | return ast_for_factor(c, n); |
|---|
| 2688 | n/a | case power: |
|---|
| 2689 | n/a | return ast_for_power(c, n); |
|---|
| 2690 | n/a | default: |
|---|
| 2691 | n/a | PyErr_Format(PyExc_SystemError, "unhandled expr: %d", TYPE(n)); |
|---|
| 2692 | n/a | return NULL; |
|---|
| 2693 | n/a | } |
|---|
| 2694 | n/a | /* should never get here unless if error is set */ |
|---|
| 2695 | n/a | return NULL; |
|---|
| 2696 | n/a | } |
|---|
| 2697 | n/a | |
|---|
| 2698 | n/a | static expr_ty |
|---|
| 2699 | n/a | ast_for_call(struct compiling *c, const node *n, expr_ty func) |
|---|
| 2700 | n/a | { |
|---|
| 2701 | n/a | /* |
|---|
| 2702 | n/a | arglist: argument (',' argument)* [','] |
|---|
| 2703 | n/a | argument: ( test [comp_for] | '*' test | test '=' test | '**' test ) |
|---|
| 2704 | n/a | */ |
|---|
| 2705 | n/a | |
|---|
| 2706 | n/a | int i, nargs, nkeywords, ngens; |
|---|
| 2707 | n/a | int ndoublestars; |
|---|
| 2708 | n/a | asdl_seq *args; |
|---|
| 2709 | n/a | asdl_seq *keywords; |
|---|
| 2710 | n/a | |
|---|
| 2711 | n/a | REQ(n, arglist); |
|---|
| 2712 | n/a | |
|---|
| 2713 | n/a | nargs = 0; |
|---|
| 2714 | n/a | nkeywords = 0; |
|---|
| 2715 | n/a | ngens = 0; |
|---|
| 2716 | n/a | for (i = 0; i < NCH(n); i++) { |
|---|
| 2717 | n/a | node *ch = CHILD(n, i); |
|---|
| 2718 | n/a | if (TYPE(ch) == argument) { |
|---|
| 2719 | n/a | if (NCH(ch) == 1) |
|---|
| 2720 | n/a | nargs++; |
|---|
| 2721 | n/a | else if (TYPE(CHILD(ch, 1)) == comp_for) |
|---|
| 2722 | n/a | ngens++; |
|---|
| 2723 | n/a | else if (TYPE(CHILD(ch, 0)) == STAR) |
|---|
| 2724 | n/a | nargs++; |
|---|
| 2725 | n/a | else |
|---|
| 2726 | n/a | /* TYPE(CHILD(ch, 0)) == DOUBLESTAR or keyword argument */ |
|---|
| 2727 | n/a | nkeywords++; |
|---|
| 2728 | n/a | } |
|---|
| 2729 | n/a | } |
|---|
| 2730 | n/a | if (ngens > 1 || (ngens && (nargs || nkeywords))) { |
|---|
| 2731 | n/a | ast_error(c, n, "Generator expression must be parenthesized " |
|---|
| 2732 | n/a | "if not sole argument"); |
|---|
| 2733 | n/a | return NULL; |
|---|
| 2734 | n/a | } |
|---|
| 2735 | n/a | |
|---|
| 2736 | n/a | args = _Py_asdl_seq_new(nargs + ngens, c->c_arena); |
|---|
| 2737 | n/a | if (!args) |
|---|
| 2738 | n/a | return NULL; |
|---|
| 2739 | n/a | keywords = _Py_asdl_seq_new(nkeywords, c->c_arena); |
|---|
| 2740 | n/a | if (!keywords) |
|---|
| 2741 | n/a | return NULL; |
|---|
| 2742 | n/a | |
|---|
| 2743 | n/a | nargs = 0; /* positional arguments + iterable argument unpackings */ |
|---|
| 2744 | n/a | nkeywords = 0; /* keyword arguments + keyword argument unpackings */ |
|---|
| 2745 | n/a | ndoublestars = 0; /* just keyword argument unpackings */ |
|---|
| 2746 | n/a | for (i = 0; i < NCH(n); i++) { |
|---|
| 2747 | n/a | node *ch = CHILD(n, i); |
|---|
| 2748 | n/a | if (TYPE(ch) == argument) { |
|---|
| 2749 | n/a | expr_ty e; |
|---|
| 2750 | n/a | node *chch = CHILD(ch, 0); |
|---|
| 2751 | n/a | if (NCH(ch) == 1) { |
|---|
| 2752 | n/a | /* a positional argument */ |
|---|
| 2753 | n/a | if (nkeywords) { |
|---|
| 2754 | n/a | if (ndoublestars) { |
|---|
| 2755 | n/a | ast_error(c, chch, |
|---|
| 2756 | n/a | "positional argument follows " |
|---|
| 2757 | n/a | "keyword argument unpacking"); |
|---|
| 2758 | n/a | } |
|---|
| 2759 | n/a | else { |
|---|
| 2760 | n/a | ast_error(c, chch, |
|---|
| 2761 | n/a | "positional argument follows " |
|---|
| 2762 | n/a | "keyword argument"); |
|---|
| 2763 | n/a | } |
|---|
| 2764 | n/a | return NULL; |
|---|
| 2765 | n/a | } |
|---|
| 2766 | n/a | e = ast_for_expr(c, chch); |
|---|
| 2767 | n/a | if (!e) |
|---|
| 2768 | n/a | return NULL; |
|---|
| 2769 | n/a | asdl_seq_SET(args, nargs++, e); |
|---|
| 2770 | n/a | } |
|---|
| 2771 | n/a | else if (TYPE(chch) == STAR) { |
|---|
| 2772 | n/a | /* an iterable argument unpacking */ |
|---|
| 2773 | n/a | expr_ty starred; |
|---|
| 2774 | n/a | if (ndoublestars) { |
|---|
| 2775 | n/a | ast_error(c, chch, |
|---|
| 2776 | n/a | "iterable argument unpacking follows " |
|---|
| 2777 | n/a | "keyword argument unpacking"); |
|---|
| 2778 | n/a | return NULL; |
|---|
| 2779 | n/a | } |
|---|
| 2780 | n/a | e = ast_for_expr(c, CHILD(ch, 1)); |
|---|
| 2781 | n/a | if (!e) |
|---|
| 2782 | n/a | return NULL; |
|---|
| 2783 | n/a | starred = Starred(e, Load, LINENO(chch), |
|---|
| 2784 | n/a | chch->n_col_offset, |
|---|
| 2785 | n/a | c->c_arena); |
|---|
| 2786 | n/a | if (!starred) |
|---|
| 2787 | n/a | return NULL; |
|---|
| 2788 | n/a | asdl_seq_SET(args, nargs++, starred); |
|---|
| 2789 | n/a | |
|---|
| 2790 | n/a | } |
|---|
| 2791 | n/a | else if (TYPE(chch) == DOUBLESTAR) { |
|---|
| 2792 | n/a | /* a keyword argument unpacking */ |
|---|
| 2793 | n/a | keyword_ty kw; |
|---|
| 2794 | n/a | i++; |
|---|
| 2795 | n/a | e = ast_for_expr(c, CHILD(ch, 1)); |
|---|
| 2796 | n/a | if (!e) |
|---|
| 2797 | n/a | return NULL; |
|---|
| 2798 | n/a | kw = keyword(NULL, e, c->c_arena); |
|---|
| 2799 | n/a | asdl_seq_SET(keywords, nkeywords++, kw); |
|---|
| 2800 | n/a | ndoublestars++; |
|---|
| 2801 | n/a | } |
|---|
| 2802 | n/a | else if (TYPE(CHILD(ch, 1)) == comp_for) { |
|---|
| 2803 | n/a | /* the lone generator expression */ |
|---|
| 2804 | n/a | e = ast_for_genexp(c, ch); |
|---|
| 2805 | n/a | if (!e) |
|---|
| 2806 | n/a | return NULL; |
|---|
| 2807 | n/a | asdl_seq_SET(args, nargs++, e); |
|---|
| 2808 | n/a | } |
|---|
| 2809 | n/a | else { |
|---|
| 2810 | n/a | /* a keyword argument */ |
|---|
| 2811 | n/a | keyword_ty kw; |
|---|
| 2812 | n/a | identifier key, tmp; |
|---|
| 2813 | n/a | int k; |
|---|
| 2814 | n/a | |
|---|
| 2815 | n/a | /* chch is test, but must be an identifier? */ |
|---|
| 2816 | n/a | e = ast_for_expr(c, chch); |
|---|
| 2817 | n/a | if (!e) |
|---|
| 2818 | n/a | return NULL; |
|---|
| 2819 | n/a | /* f(lambda x: x[0] = 3) ends up getting parsed with |
|---|
| 2820 | n/a | * LHS test = lambda x: x[0], and RHS test = 3. |
|---|
| 2821 | n/a | * SF bug 132313 points out that complaining about a keyword |
|---|
| 2822 | n/a | * then is very confusing. |
|---|
| 2823 | n/a | */ |
|---|
| 2824 | n/a | if (e->kind == Lambda_kind) { |
|---|
| 2825 | n/a | ast_error(c, chch, |
|---|
| 2826 | n/a | "lambda cannot contain assignment"); |
|---|
| 2827 | n/a | return NULL; |
|---|
| 2828 | n/a | } |
|---|
| 2829 | n/a | else if (e->kind != Name_kind) { |
|---|
| 2830 | n/a | ast_error(c, chch, |
|---|
| 2831 | n/a | "keyword can't be an expression"); |
|---|
| 2832 | n/a | return NULL; |
|---|
| 2833 | n/a | } |
|---|
| 2834 | n/a | else if (forbidden_name(c, e->v.Name.id, ch, 1)) { |
|---|
| 2835 | n/a | return NULL; |
|---|
| 2836 | n/a | } |
|---|
| 2837 | n/a | key = e->v.Name.id; |
|---|
| 2838 | n/a | for (k = 0; k < nkeywords; k++) { |
|---|
| 2839 | n/a | tmp = ((keyword_ty)asdl_seq_GET(keywords, k))->arg; |
|---|
| 2840 | n/a | if (tmp && !PyUnicode_Compare(tmp, key)) { |
|---|
| 2841 | n/a | ast_error(c, chch, |
|---|
| 2842 | n/a | "keyword argument repeated"); |
|---|
| 2843 | n/a | return NULL; |
|---|
| 2844 | n/a | } |
|---|
| 2845 | n/a | } |
|---|
| 2846 | n/a | e = ast_for_expr(c, CHILD(ch, 2)); |
|---|
| 2847 | n/a | if (!e) |
|---|
| 2848 | n/a | return NULL; |
|---|
| 2849 | n/a | kw = keyword(key, e, c->c_arena); |
|---|
| 2850 | n/a | if (!kw) |
|---|
| 2851 | n/a | return NULL; |
|---|
| 2852 | n/a | asdl_seq_SET(keywords, nkeywords++, kw); |
|---|
| 2853 | n/a | } |
|---|
| 2854 | n/a | } |
|---|
| 2855 | n/a | } |
|---|
| 2856 | n/a | |
|---|
| 2857 | n/a | return Call(func, args, keywords, func->lineno, func->col_offset, c->c_arena); |
|---|
| 2858 | n/a | } |
|---|
| 2859 | n/a | |
|---|
| 2860 | n/a | static expr_ty |
|---|
| 2861 | n/a | ast_for_testlist(struct compiling *c, const node* n) |
|---|
| 2862 | n/a | { |
|---|
| 2863 | n/a | /* testlist_comp: test (comp_for | (',' test)* [',']) */ |
|---|
| 2864 | n/a | /* testlist: test (',' test)* [','] */ |
|---|
| 2865 | n/a | assert(NCH(n) > 0); |
|---|
| 2866 | n/a | if (TYPE(n) == testlist_comp) { |
|---|
| 2867 | n/a | if (NCH(n) > 1) |
|---|
| 2868 | n/a | assert(TYPE(CHILD(n, 1)) != comp_for); |
|---|
| 2869 | n/a | } |
|---|
| 2870 | n/a | else { |
|---|
| 2871 | n/a | assert(TYPE(n) == testlist || |
|---|
| 2872 | n/a | TYPE(n) == testlist_star_expr); |
|---|
| 2873 | n/a | } |
|---|
| 2874 | n/a | if (NCH(n) == 1) |
|---|
| 2875 | n/a | return ast_for_expr(c, CHILD(n, 0)); |
|---|
| 2876 | n/a | else { |
|---|
| 2877 | n/a | asdl_seq *tmp = seq_for_testlist(c, n); |
|---|
| 2878 | n/a | if (!tmp) |
|---|
| 2879 | n/a | return NULL; |
|---|
| 2880 | n/a | return Tuple(tmp, Load, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2881 | n/a | } |
|---|
| 2882 | n/a | } |
|---|
| 2883 | n/a | |
|---|
| 2884 | n/a | static stmt_ty |
|---|
| 2885 | n/a | ast_for_expr_stmt(struct compiling *c, const node *n) |
|---|
| 2886 | n/a | { |
|---|
| 2887 | n/a | REQ(n, expr_stmt); |
|---|
| 2888 | n/a | /* expr_stmt: testlist_star_expr (annassign | augassign (yield_expr|testlist) | |
|---|
| 2889 | n/a | ('=' (yield_expr|testlist_star_expr))*) |
|---|
| 2890 | n/a | annassign: ':' test ['=' test] |
|---|
| 2891 | n/a | testlist_star_expr: (test|star_expr) (',' test|star_expr)* [','] |
|---|
| 2892 | n/a | augassign: '+=' | '-=' | '*=' | '@=' | '/=' | '%=' | '&=' | '|=' | '^=' |
|---|
| 2893 | n/a | | '<<=' | '>>=' | '**=' | '//=' |
|---|
| 2894 | n/a | test: ... here starts the operator precedence dance |
|---|
| 2895 | n/a | */ |
|---|
| 2896 | n/a | |
|---|
| 2897 | n/a | if (NCH(n) == 1) { |
|---|
| 2898 | n/a | expr_ty e = ast_for_testlist(c, CHILD(n, 0)); |
|---|
| 2899 | n/a | if (!e) |
|---|
| 2900 | n/a | return NULL; |
|---|
| 2901 | n/a | |
|---|
| 2902 | n/a | return Expr(e, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2903 | n/a | } |
|---|
| 2904 | n/a | else if (TYPE(CHILD(n, 1)) == augassign) { |
|---|
| 2905 | n/a | expr_ty expr1, expr2; |
|---|
| 2906 | n/a | operator_ty newoperator; |
|---|
| 2907 | n/a | node *ch = CHILD(n, 0); |
|---|
| 2908 | n/a | |
|---|
| 2909 | n/a | expr1 = ast_for_testlist(c, ch); |
|---|
| 2910 | n/a | if (!expr1) |
|---|
| 2911 | n/a | return NULL; |
|---|
| 2912 | n/a | if(!set_context(c, expr1, Store, ch)) |
|---|
| 2913 | n/a | return NULL; |
|---|
| 2914 | n/a | /* set_context checks that most expressions are not the left side. |
|---|
| 2915 | n/a | Augmented assignments can only have a name, a subscript, or an |
|---|
| 2916 | n/a | attribute on the left, though, so we have to explicitly check for |
|---|
| 2917 | n/a | those. */ |
|---|
| 2918 | n/a | switch (expr1->kind) { |
|---|
| 2919 | n/a | case Name_kind: |
|---|
| 2920 | n/a | case Attribute_kind: |
|---|
| 2921 | n/a | case Subscript_kind: |
|---|
| 2922 | n/a | break; |
|---|
| 2923 | n/a | default: |
|---|
| 2924 | n/a | ast_error(c, ch, "illegal expression for augmented assignment"); |
|---|
| 2925 | n/a | return NULL; |
|---|
| 2926 | n/a | } |
|---|
| 2927 | n/a | |
|---|
| 2928 | n/a | ch = CHILD(n, 2); |
|---|
| 2929 | n/a | if (TYPE(ch) == testlist) |
|---|
| 2930 | n/a | expr2 = ast_for_testlist(c, ch); |
|---|
| 2931 | n/a | else |
|---|
| 2932 | n/a | expr2 = ast_for_expr(c, ch); |
|---|
| 2933 | n/a | if (!expr2) |
|---|
| 2934 | n/a | return NULL; |
|---|
| 2935 | n/a | |
|---|
| 2936 | n/a | newoperator = ast_for_augassign(c, CHILD(n, 1)); |
|---|
| 2937 | n/a | if (!newoperator) |
|---|
| 2938 | n/a | return NULL; |
|---|
| 2939 | n/a | |
|---|
| 2940 | n/a | return AugAssign(expr1, newoperator, expr2, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 2941 | n/a | } |
|---|
| 2942 | n/a | else if (TYPE(CHILD(n, 1)) == annassign) { |
|---|
| 2943 | n/a | expr_ty expr1, expr2, expr3; |
|---|
| 2944 | n/a | node *ch = CHILD(n, 0); |
|---|
| 2945 | n/a | node *deep, *ann = CHILD(n, 1); |
|---|
| 2946 | n/a | int simple = 1; |
|---|
| 2947 | n/a | |
|---|
| 2948 | n/a | /* we keep track of parens to qualify (x) as expression not name */ |
|---|
| 2949 | n/a | deep = ch; |
|---|
| 2950 | n/a | while (NCH(deep) == 1) { |
|---|
| 2951 | n/a | deep = CHILD(deep, 0); |
|---|
| 2952 | n/a | } |
|---|
| 2953 | n/a | if (NCH(deep) > 0 && TYPE(CHILD(deep, 0)) == LPAR) { |
|---|
| 2954 | n/a | simple = 0; |
|---|
| 2955 | n/a | } |
|---|
| 2956 | n/a | expr1 = ast_for_testlist(c, ch); |
|---|
| 2957 | n/a | if (!expr1) { |
|---|
| 2958 | n/a | return NULL; |
|---|
| 2959 | n/a | } |
|---|
| 2960 | n/a | switch (expr1->kind) { |
|---|
| 2961 | n/a | case Name_kind: |
|---|
| 2962 | n/a | if (forbidden_name(c, expr1->v.Name.id, n, 0)) { |
|---|
| 2963 | n/a | return NULL; |
|---|
| 2964 | n/a | } |
|---|
| 2965 | n/a | expr1->v.Name.ctx = Store; |
|---|
| 2966 | n/a | break; |
|---|
| 2967 | n/a | case Attribute_kind: |
|---|
| 2968 | n/a | if (forbidden_name(c, expr1->v.Attribute.attr, n, 1)) { |
|---|
| 2969 | n/a | return NULL; |
|---|
| 2970 | n/a | } |
|---|
| 2971 | n/a | expr1->v.Attribute.ctx = Store; |
|---|
| 2972 | n/a | break; |
|---|
| 2973 | n/a | case Subscript_kind: |
|---|
| 2974 | n/a | expr1->v.Subscript.ctx = Store; |
|---|
| 2975 | n/a | break; |
|---|
| 2976 | n/a | case List_kind: |
|---|
| 2977 | n/a | ast_error(c, ch, |
|---|
| 2978 | n/a | "only single target (not list) can be annotated"); |
|---|
| 2979 | n/a | return NULL; |
|---|
| 2980 | n/a | case Tuple_kind: |
|---|
| 2981 | n/a | ast_error(c, ch, |
|---|
| 2982 | n/a | "only single target (not tuple) can be annotated"); |
|---|
| 2983 | n/a | return NULL; |
|---|
| 2984 | n/a | default: |
|---|
| 2985 | n/a | ast_error(c, ch, |
|---|
| 2986 | n/a | "illegal target for annotation"); |
|---|
| 2987 | n/a | return NULL; |
|---|
| 2988 | n/a | } |
|---|
| 2989 | n/a | |
|---|
| 2990 | n/a | if (expr1->kind != Name_kind) { |
|---|
| 2991 | n/a | simple = 0; |
|---|
| 2992 | n/a | } |
|---|
| 2993 | n/a | ch = CHILD(ann, 1); |
|---|
| 2994 | n/a | expr2 = ast_for_expr(c, ch); |
|---|
| 2995 | n/a | if (!expr2) { |
|---|
| 2996 | n/a | return NULL; |
|---|
| 2997 | n/a | } |
|---|
| 2998 | n/a | if (NCH(ann) == 2) { |
|---|
| 2999 | n/a | return AnnAssign(expr1, expr2, NULL, simple, |
|---|
| 3000 | n/a | LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3001 | n/a | } |
|---|
| 3002 | n/a | else { |
|---|
| 3003 | n/a | ch = CHILD(ann, 3); |
|---|
| 3004 | n/a | expr3 = ast_for_expr(c, ch); |
|---|
| 3005 | n/a | if (!expr3) { |
|---|
| 3006 | n/a | return NULL; |
|---|
| 3007 | n/a | } |
|---|
| 3008 | n/a | return AnnAssign(expr1, expr2, expr3, simple, |
|---|
| 3009 | n/a | LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3010 | n/a | } |
|---|
| 3011 | n/a | } |
|---|
| 3012 | n/a | else { |
|---|
| 3013 | n/a | int i; |
|---|
| 3014 | n/a | asdl_seq *targets; |
|---|
| 3015 | n/a | node *value; |
|---|
| 3016 | n/a | expr_ty expression; |
|---|
| 3017 | n/a | |
|---|
| 3018 | n/a | /* a normal assignment */ |
|---|
| 3019 | n/a | REQ(CHILD(n, 1), EQUAL); |
|---|
| 3020 | n/a | targets = _Py_asdl_seq_new(NCH(n) / 2, c->c_arena); |
|---|
| 3021 | n/a | if (!targets) |
|---|
| 3022 | n/a | return NULL; |
|---|
| 3023 | n/a | for (i = 0; i < NCH(n) - 2; i += 2) { |
|---|
| 3024 | n/a | expr_ty e; |
|---|
| 3025 | n/a | node *ch = CHILD(n, i); |
|---|
| 3026 | n/a | if (TYPE(ch) == yield_expr) { |
|---|
| 3027 | n/a | ast_error(c, ch, "assignment to yield expression not possible"); |
|---|
| 3028 | n/a | return NULL; |
|---|
| 3029 | n/a | } |
|---|
| 3030 | n/a | e = ast_for_testlist(c, ch); |
|---|
| 3031 | n/a | if (!e) |
|---|
| 3032 | n/a | return NULL; |
|---|
| 3033 | n/a | |
|---|
| 3034 | n/a | /* set context to assign */ |
|---|
| 3035 | n/a | if (!set_context(c, e, Store, CHILD(n, i))) |
|---|
| 3036 | n/a | return NULL; |
|---|
| 3037 | n/a | |
|---|
| 3038 | n/a | asdl_seq_SET(targets, i / 2, e); |
|---|
| 3039 | n/a | } |
|---|
| 3040 | n/a | value = CHILD(n, NCH(n) - 1); |
|---|
| 3041 | n/a | if (TYPE(value) == testlist_star_expr) |
|---|
| 3042 | n/a | expression = ast_for_testlist(c, value); |
|---|
| 3043 | n/a | else |
|---|
| 3044 | n/a | expression = ast_for_expr(c, value); |
|---|
| 3045 | n/a | if (!expression) |
|---|
| 3046 | n/a | return NULL; |
|---|
| 3047 | n/a | return Assign(targets, expression, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3048 | n/a | } |
|---|
| 3049 | n/a | } |
|---|
| 3050 | n/a | |
|---|
| 3051 | n/a | |
|---|
| 3052 | n/a | static asdl_seq * |
|---|
| 3053 | n/a | ast_for_exprlist(struct compiling *c, const node *n, expr_context_ty context) |
|---|
| 3054 | n/a | { |
|---|
| 3055 | n/a | asdl_seq *seq; |
|---|
| 3056 | n/a | int i; |
|---|
| 3057 | n/a | expr_ty e; |
|---|
| 3058 | n/a | |
|---|
| 3059 | n/a | REQ(n, exprlist); |
|---|
| 3060 | n/a | |
|---|
| 3061 | n/a | seq = _Py_asdl_seq_new((NCH(n) + 1) / 2, c->c_arena); |
|---|
| 3062 | n/a | if (!seq) |
|---|
| 3063 | n/a | return NULL; |
|---|
| 3064 | n/a | for (i = 0; i < NCH(n); i += 2) { |
|---|
| 3065 | n/a | e = ast_for_expr(c, CHILD(n, i)); |
|---|
| 3066 | n/a | if (!e) |
|---|
| 3067 | n/a | return NULL; |
|---|
| 3068 | n/a | asdl_seq_SET(seq, i / 2, e); |
|---|
| 3069 | n/a | if (context && !set_context(c, e, context, CHILD(n, i))) |
|---|
| 3070 | n/a | return NULL; |
|---|
| 3071 | n/a | } |
|---|
| 3072 | n/a | return seq; |
|---|
| 3073 | n/a | } |
|---|
| 3074 | n/a | |
|---|
| 3075 | n/a | static stmt_ty |
|---|
| 3076 | n/a | ast_for_del_stmt(struct compiling *c, const node *n) |
|---|
| 3077 | n/a | { |
|---|
| 3078 | n/a | asdl_seq *expr_list; |
|---|
| 3079 | n/a | |
|---|
| 3080 | n/a | /* del_stmt: 'del' exprlist */ |
|---|
| 3081 | n/a | REQ(n, del_stmt); |
|---|
| 3082 | n/a | |
|---|
| 3083 | n/a | expr_list = ast_for_exprlist(c, CHILD(n, 1), Del); |
|---|
| 3084 | n/a | if (!expr_list) |
|---|
| 3085 | n/a | return NULL; |
|---|
| 3086 | n/a | return Delete(expr_list, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3087 | n/a | } |
|---|
| 3088 | n/a | |
|---|
| 3089 | n/a | static stmt_ty |
|---|
| 3090 | n/a | ast_for_flow_stmt(struct compiling *c, const node *n) |
|---|
| 3091 | n/a | { |
|---|
| 3092 | n/a | /* |
|---|
| 3093 | n/a | flow_stmt: break_stmt | continue_stmt | return_stmt | raise_stmt |
|---|
| 3094 | n/a | | yield_stmt |
|---|
| 3095 | n/a | break_stmt: 'break' |
|---|
| 3096 | n/a | continue_stmt: 'continue' |
|---|
| 3097 | n/a | return_stmt: 'return' [testlist] |
|---|
| 3098 | n/a | yield_stmt: yield_expr |
|---|
| 3099 | n/a | yield_expr: 'yield' testlist | 'yield' 'from' test |
|---|
| 3100 | n/a | raise_stmt: 'raise' [test [',' test [',' test]]] |
|---|
| 3101 | n/a | */ |
|---|
| 3102 | n/a | node *ch; |
|---|
| 3103 | n/a | |
|---|
| 3104 | n/a | REQ(n, flow_stmt); |
|---|
| 3105 | n/a | ch = CHILD(n, 0); |
|---|
| 3106 | n/a | switch (TYPE(ch)) { |
|---|
| 3107 | n/a | case break_stmt: |
|---|
| 3108 | n/a | return Break(LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3109 | n/a | case continue_stmt: |
|---|
| 3110 | n/a | return Continue(LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3111 | n/a | case yield_stmt: { /* will reduce to yield_expr */ |
|---|
| 3112 | n/a | expr_ty exp = ast_for_expr(c, CHILD(ch, 0)); |
|---|
| 3113 | n/a | if (!exp) |
|---|
| 3114 | n/a | return NULL; |
|---|
| 3115 | n/a | return Expr(exp, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3116 | n/a | } |
|---|
| 3117 | n/a | case return_stmt: |
|---|
| 3118 | n/a | if (NCH(ch) == 1) |
|---|
| 3119 | n/a | return Return(NULL, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3120 | n/a | else { |
|---|
| 3121 | n/a | expr_ty expression = ast_for_testlist(c, CHILD(ch, 1)); |
|---|
| 3122 | n/a | if (!expression) |
|---|
| 3123 | n/a | return NULL; |
|---|
| 3124 | n/a | return Return(expression, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3125 | n/a | } |
|---|
| 3126 | n/a | case raise_stmt: |
|---|
| 3127 | n/a | if (NCH(ch) == 1) |
|---|
| 3128 | n/a | return Raise(NULL, NULL, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3129 | n/a | else if (NCH(ch) >= 2) { |
|---|
| 3130 | n/a | expr_ty cause = NULL; |
|---|
| 3131 | n/a | expr_ty expression = ast_for_expr(c, CHILD(ch, 1)); |
|---|
| 3132 | n/a | if (!expression) |
|---|
| 3133 | n/a | return NULL; |
|---|
| 3134 | n/a | if (NCH(ch) == 4) { |
|---|
| 3135 | n/a | cause = ast_for_expr(c, CHILD(ch, 3)); |
|---|
| 3136 | n/a | if (!cause) |
|---|
| 3137 | n/a | return NULL; |
|---|
| 3138 | n/a | } |
|---|
| 3139 | n/a | return Raise(expression, cause, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3140 | n/a | } |
|---|
| 3141 | n/a | default: |
|---|
| 3142 | n/a | PyErr_Format(PyExc_SystemError, |
|---|
| 3143 | n/a | "unexpected flow_stmt: %d", TYPE(ch)); |
|---|
| 3144 | n/a | return NULL; |
|---|
| 3145 | n/a | } |
|---|
| 3146 | n/a | } |
|---|
| 3147 | n/a | |
|---|
| 3148 | n/a | static alias_ty |
|---|
| 3149 | n/a | alias_for_import_name(struct compiling *c, const node *n, int store) |
|---|
| 3150 | n/a | { |
|---|
| 3151 | n/a | /* |
|---|
| 3152 | n/a | import_as_name: NAME ['as' NAME] |
|---|
| 3153 | n/a | dotted_as_name: dotted_name ['as' NAME] |
|---|
| 3154 | n/a | dotted_name: NAME ('.' NAME)* |
|---|
| 3155 | n/a | */ |
|---|
| 3156 | n/a | identifier str, name; |
|---|
| 3157 | n/a | |
|---|
| 3158 | n/a | loop: |
|---|
| 3159 | n/a | switch (TYPE(n)) { |
|---|
| 3160 | n/a | case import_as_name: { |
|---|
| 3161 | n/a | node *name_node = CHILD(n, 0); |
|---|
| 3162 | n/a | str = NULL; |
|---|
| 3163 | n/a | name = NEW_IDENTIFIER(name_node); |
|---|
| 3164 | n/a | if (!name) |
|---|
| 3165 | n/a | return NULL; |
|---|
| 3166 | n/a | if (NCH(n) == 3) { |
|---|
| 3167 | n/a | node *str_node = CHILD(n, 2); |
|---|
| 3168 | n/a | str = NEW_IDENTIFIER(str_node); |
|---|
| 3169 | n/a | if (!str) |
|---|
| 3170 | n/a | return NULL; |
|---|
| 3171 | n/a | if (store && forbidden_name(c, str, str_node, 0)) |
|---|
| 3172 | n/a | return NULL; |
|---|
| 3173 | n/a | } |
|---|
| 3174 | n/a | else { |
|---|
| 3175 | n/a | if (forbidden_name(c, name, name_node, 0)) |
|---|
| 3176 | n/a | return NULL; |
|---|
| 3177 | n/a | } |
|---|
| 3178 | n/a | return alias(name, str, c->c_arena); |
|---|
| 3179 | n/a | } |
|---|
| 3180 | n/a | case dotted_as_name: |
|---|
| 3181 | n/a | if (NCH(n) == 1) { |
|---|
| 3182 | n/a | n = CHILD(n, 0); |
|---|
| 3183 | n/a | goto loop; |
|---|
| 3184 | n/a | } |
|---|
| 3185 | n/a | else { |
|---|
| 3186 | n/a | node *asname_node = CHILD(n, 2); |
|---|
| 3187 | n/a | alias_ty a = alias_for_import_name(c, CHILD(n, 0), 0); |
|---|
| 3188 | n/a | if (!a) |
|---|
| 3189 | n/a | return NULL; |
|---|
| 3190 | n/a | assert(!a->asname); |
|---|
| 3191 | n/a | a->asname = NEW_IDENTIFIER(asname_node); |
|---|
| 3192 | n/a | if (!a->asname) |
|---|
| 3193 | n/a | return NULL; |
|---|
| 3194 | n/a | if (forbidden_name(c, a->asname, asname_node, 0)) |
|---|
| 3195 | n/a | return NULL; |
|---|
| 3196 | n/a | return a; |
|---|
| 3197 | n/a | } |
|---|
| 3198 | n/a | break; |
|---|
| 3199 | n/a | case dotted_name: |
|---|
| 3200 | n/a | if (NCH(n) == 1) { |
|---|
| 3201 | n/a | node *name_node = CHILD(n, 0); |
|---|
| 3202 | n/a | name = NEW_IDENTIFIER(name_node); |
|---|
| 3203 | n/a | if (!name) |
|---|
| 3204 | n/a | return NULL; |
|---|
| 3205 | n/a | if (store && forbidden_name(c, name, name_node, 0)) |
|---|
| 3206 | n/a | return NULL; |
|---|
| 3207 | n/a | return alias(name, NULL, c->c_arena); |
|---|
| 3208 | n/a | } |
|---|
| 3209 | n/a | else { |
|---|
| 3210 | n/a | /* Create a string of the form "a.b.c" */ |
|---|
| 3211 | n/a | int i; |
|---|
| 3212 | n/a | size_t len; |
|---|
| 3213 | n/a | char *s; |
|---|
| 3214 | n/a | PyObject *uni; |
|---|
| 3215 | n/a | |
|---|
| 3216 | n/a | len = 0; |
|---|
| 3217 | n/a | for (i = 0; i < NCH(n); i += 2) |
|---|
| 3218 | n/a | /* length of string plus one for the dot */ |
|---|
| 3219 | n/a | len += strlen(STR(CHILD(n, i))) + 1; |
|---|
| 3220 | n/a | len--; /* the last name doesn't have a dot */ |
|---|
| 3221 | n/a | str = PyBytes_FromStringAndSize(NULL, len); |
|---|
| 3222 | n/a | if (!str) |
|---|
| 3223 | n/a | return NULL; |
|---|
| 3224 | n/a | s = PyBytes_AS_STRING(str); |
|---|
| 3225 | n/a | if (!s) |
|---|
| 3226 | n/a | return NULL; |
|---|
| 3227 | n/a | for (i = 0; i < NCH(n); i += 2) { |
|---|
| 3228 | n/a | char *sch = STR(CHILD(n, i)); |
|---|
| 3229 | n/a | strcpy(s, STR(CHILD(n, i))); |
|---|
| 3230 | n/a | s += strlen(sch); |
|---|
| 3231 | n/a | *s++ = '.'; |
|---|
| 3232 | n/a | } |
|---|
| 3233 | n/a | --s; |
|---|
| 3234 | n/a | *s = '\0'; |
|---|
| 3235 | n/a | uni = PyUnicode_DecodeUTF8(PyBytes_AS_STRING(str), |
|---|
| 3236 | n/a | PyBytes_GET_SIZE(str), |
|---|
| 3237 | n/a | NULL); |
|---|
| 3238 | n/a | Py_DECREF(str); |
|---|
| 3239 | n/a | if (!uni) |
|---|
| 3240 | n/a | return NULL; |
|---|
| 3241 | n/a | str = uni; |
|---|
| 3242 | n/a | PyUnicode_InternInPlace(&str); |
|---|
| 3243 | n/a | if (PyArena_AddPyObject(c->c_arena, str) < 0) { |
|---|
| 3244 | n/a | Py_DECREF(str); |
|---|
| 3245 | n/a | return NULL; |
|---|
| 3246 | n/a | } |
|---|
| 3247 | n/a | return alias(str, NULL, c->c_arena); |
|---|
| 3248 | n/a | } |
|---|
| 3249 | n/a | break; |
|---|
| 3250 | n/a | case STAR: |
|---|
| 3251 | n/a | str = PyUnicode_InternFromString("*"); |
|---|
| 3252 | n/a | if (PyArena_AddPyObject(c->c_arena, str) < 0) { |
|---|
| 3253 | n/a | Py_DECREF(str); |
|---|
| 3254 | n/a | return NULL; |
|---|
| 3255 | n/a | } |
|---|
| 3256 | n/a | return alias(str, NULL, c->c_arena); |
|---|
| 3257 | n/a | default: |
|---|
| 3258 | n/a | PyErr_Format(PyExc_SystemError, |
|---|
| 3259 | n/a | "unexpected import name: %d", TYPE(n)); |
|---|
| 3260 | n/a | return NULL; |
|---|
| 3261 | n/a | } |
|---|
| 3262 | n/a | |
|---|
| 3263 | n/a | PyErr_SetString(PyExc_SystemError, "unhandled import name condition"); |
|---|
| 3264 | n/a | return NULL; |
|---|
| 3265 | n/a | } |
|---|
| 3266 | n/a | |
|---|
| 3267 | n/a | static stmt_ty |
|---|
| 3268 | n/a | ast_for_import_stmt(struct compiling *c, const node *n) |
|---|
| 3269 | n/a | { |
|---|
| 3270 | n/a | /* |
|---|
| 3271 | n/a | import_stmt: import_name | import_from |
|---|
| 3272 | n/a | import_name: 'import' dotted_as_names |
|---|
| 3273 | n/a | import_from: 'from' (('.' | '...')* dotted_name | ('.' | '...')+) |
|---|
| 3274 | n/a | 'import' ('*' | '(' import_as_names ')' | import_as_names) |
|---|
| 3275 | n/a | */ |
|---|
| 3276 | n/a | int lineno; |
|---|
| 3277 | n/a | int col_offset; |
|---|
| 3278 | n/a | int i; |
|---|
| 3279 | n/a | asdl_seq *aliases; |
|---|
| 3280 | n/a | |
|---|
| 3281 | n/a | REQ(n, import_stmt); |
|---|
| 3282 | n/a | lineno = LINENO(n); |
|---|
| 3283 | n/a | col_offset = n->n_col_offset; |
|---|
| 3284 | n/a | n = CHILD(n, 0); |
|---|
| 3285 | n/a | if (TYPE(n) == import_name) { |
|---|
| 3286 | n/a | n = CHILD(n, 1); |
|---|
| 3287 | n/a | REQ(n, dotted_as_names); |
|---|
| 3288 | n/a | aliases = _Py_asdl_seq_new((NCH(n) + 1) / 2, c->c_arena); |
|---|
| 3289 | n/a | if (!aliases) |
|---|
| 3290 | n/a | return NULL; |
|---|
| 3291 | n/a | for (i = 0; i < NCH(n); i += 2) { |
|---|
| 3292 | n/a | alias_ty import_alias = alias_for_import_name(c, CHILD(n, i), 1); |
|---|
| 3293 | n/a | if (!import_alias) |
|---|
| 3294 | n/a | return NULL; |
|---|
| 3295 | n/a | asdl_seq_SET(aliases, i / 2, import_alias); |
|---|
| 3296 | n/a | } |
|---|
| 3297 | n/a | return Import(aliases, lineno, col_offset, c->c_arena); |
|---|
| 3298 | n/a | } |
|---|
| 3299 | n/a | else if (TYPE(n) == import_from) { |
|---|
| 3300 | n/a | int n_children; |
|---|
| 3301 | n/a | int idx, ndots = 0; |
|---|
| 3302 | n/a | alias_ty mod = NULL; |
|---|
| 3303 | n/a | identifier modname = NULL; |
|---|
| 3304 | n/a | |
|---|
| 3305 | n/a | /* Count the number of dots (for relative imports) and check for the |
|---|
| 3306 | n/a | optional module name */ |
|---|
| 3307 | n/a | for (idx = 1; idx < NCH(n); idx++) { |
|---|
| 3308 | n/a | if (TYPE(CHILD(n, idx)) == dotted_name) { |
|---|
| 3309 | n/a | mod = alias_for_import_name(c, CHILD(n, idx), 0); |
|---|
| 3310 | n/a | if (!mod) |
|---|
| 3311 | n/a | return NULL; |
|---|
| 3312 | n/a | idx++; |
|---|
| 3313 | n/a | break; |
|---|
| 3314 | n/a | } else if (TYPE(CHILD(n, idx)) == ELLIPSIS) { |
|---|
| 3315 | n/a | /* three consecutive dots are tokenized as one ELLIPSIS */ |
|---|
| 3316 | n/a | ndots += 3; |
|---|
| 3317 | n/a | continue; |
|---|
| 3318 | n/a | } else if (TYPE(CHILD(n, idx)) != DOT) { |
|---|
| 3319 | n/a | break; |
|---|
| 3320 | n/a | } |
|---|
| 3321 | n/a | ndots++; |
|---|
| 3322 | n/a | } |
|---|
| 3323 | n/a | idx++; /* skip over the 'import' keyword */ |
|---|
| 3324 | n/a | switch (TYPE(CHILD(n, idx))) { |
|---|
| 3325 | n/a | case STAR: |
|---|
| 3326 | n/a | /* from ... import * */ |
|---|
| 3327 | n/a | n = CHILD(n, idx); |
|---|
| 3328 | n/a | n_children = 1; |
|---|
| 3329 | n/a | break; |
|---|
| 3330 | n/a | case LPAR: |
|---|
| 3331 | n/a | /* from ... import (x, y, z) */ |
|---|
| 3332 | n/a | n = CHILD(n, idx + 1); |
|---|
| 3333 | n/a | n_children = NCH(n); |
|---|
| 3334 | n/a | break; |
|---|
| 3335 | n/a | case import_as_names: |
|---|
| 3336 | n/a | /* from ... import x, y, z */ |
|---|
| 3337 | n/a | n = CHILD(n, idx); |
|---|
| 3338 | n/a | n_children = NCH(n); |
|---|
| 3339 | n/a | if (n_children % 2 == 0) { |
|---|
| 3340 | n/a | ast_error(c, n, "trailing comma not allowed without" |
|---|
| 3341 | n/a | " surrounding parentheses"); |
|---|
| 3342 | n/a | return NULL; |
|---|
| 3343 | n/a | } |
|---|
| 3344 | n/a | break; |
|---|
| 3345 | n/a | default: |
|---|
| 3346 | n/a | ast_error(c, n, "Unexpected node-type in from-import"); |
|---|
| 3347 | n/a | return NULL; |
|---|
| 3348 | n/a | } |
|---|
| 3349 | n/a | |
|---|
| 3350 | n/a | aliases = _Py_asdl_seq_new((n_children + 1) / 2, c->c_arena); |
|---|
| 3351 | n/a | if (!aliases) |
|---|
| 3352 | n/a | return NULL; |
|---|
| 3353 | n/a | |
|---|
| 3354 | n/a | /* handle "from ... import *" special b/c there's no children */ |
|---|
| 3355 | n/a | if (TYPE(n) == STAR) { |
|---|
| 3356 | n/a | alias_ty import_alias = alias_for_import_name(c, n, 1); |
|---|
| 3357 | n/a | if (!import_alias) |
|---|
| 3358 | n/a | return NULL; |
|---|
| 3359 | n/a | asdl_seq_SET(aliases, 0, import_alias); |
|---|
| 3360 | n/a | } |
|---|
| 3361 | n/a | else { |
|---|
| 3362 | n/a | for (i = 0; i < NCH(n); i += 2) { |
|---|
| 3363 | n/a | alias_ty import_alias = alias_for_import_name(c, CHILD(n, i), 1); |
|---|
| 3364 | n/a | if (!import_alias) |
|---|
| 3365 | n/a | return NULL; |
|---|
| 3366 | n/a | asdl_seq_SET(aliases, i / 2, import_alias); |
|---|
| 3367 | n/a | } |
|---|
| 3368 | n/a | } |
|---|
| 3369 | n/a | if (mod != NULL) |
|---|
| 3370 | n/a | modname = mod->name; |
|---|
| 3371 | n/a | return ImportFrom(modname, aliases, ndots, lineno, col_offset, |
|---|
| 3372 | n/a | c->c_arena); |
|---|
| 3373 | n/a | } |
|---|
| 3374 | n/a | PyErr_Format(PyExc_SystemError, |
|---|
| 3375 | n/a | "unknown import statement: starts with command '%s'", |
|---|
| 3376 | n/a | STR(CHILD(n, 0))); |
|---|
| 3377 | n/a | return NULL; |
|---|
| 3378 | n/a | } |
|---|
| 3379 | n/a | |
|---|
| 3380 | n/a | static stmt_ty |
|---|
| 3381 | n/a | ast_for_global_stmt(struct compiling *c, const node *n) |
|---|
| 3382 | n/a | { |
|---|
| 3383 | n/a | /* global_stmt: 'global' NAME (',' NAME)* */ |
|---|
| 3384 | n/a | identifier name; |
|---|
| 3385 | n/a | asdl_seq *s; |
|---|
| 3386 | n/a | int i; |
|---|
| 3387 | n/a | |
|---|
| 3388 | n/a | REQ(n, global_stmt); |
|---|
| 3389 | n/a | s = _Py_asdl_seq_new(NCH(n) / 2, c->c_arena); |
|---|
| 3390 | n/a | if (!s) |
|---|
| 3391 | n/a | return NULL; |
|---|
| 3392 | n/a | for (i = 1; i < NCH(n); i += 2) { |
|---|
| 3393 | n/a | name = NEW_IDENTIFIER(CHILD(n, i)); |
|---|
| 3394 | n/a | if (!name) |
|---|
| 3395 | n/a | return NULL; |
|---|
| 3396 | n/a | asdl_seq_SET(s, i / 2, name); |
|---|
| 3397 | n/a | } |
|---|
| 3398 | n/a | return Global(s, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3399 | n/a | } |
|---|
| 3400 | n/a | |
|---|
| 3401 | n/a | static stmt_ty |
|---|
| 3402 | n/a | ast_for_nonlocal_stmt(struct compiling *c, const node *n) |
|---|
| 3403 | n/a | { |
|---|
| 3404 | n/a | /* nonlocal_stmt: 'nonlocal' NAME (',' NAME)* */ |
|---|
| 3405 | n/a | identifier name; |
|---|
| 3406 | n/a | asdl_seq *s; |
|---|
| 3407 | n/a | int i; |
|---|
| 3408 | n/a | |
|---|
| 3409 | n/a | REQ(n, nonlocal_stmt); |
|---|
| 3410 | n/a | s = _Py_asdl_seq_new(NCH(n) / 2, c->c_arena); |
|---|
| 3411 | n/a | if (!s) |
|---|
| 3412 | n/a | return NULL; |
|---|
| 3413 | n/a | for (i = 1; i < NCH(n); i += 2) { |
|---|
| 3414 | n/a | name = NEW_IDENTIFIER(CHILD(n, i)); |
|---|
| 3415 | n/a | if (!name) |
|---|
| 3416 | n/a | return NULL; |
|---|
| 3417 | n/a | asdl_seq_SET(s, i / 2, name); |
|---|
| 3418 | n/a | } |
|---|
| 3419 | n/a | return Nonlocal(s, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3420 | n/a | } |
|---|
| 3421 | n/a | |
|---|
| 3422 | n/a | static stmt_ty |
|---|
| 3423 | n/a | ast_for_assert_stmt(struct compiling *c, const node *n) |
|---|
| 3424 | n/a | { |
|---|
| 3425 | n/a | /* assert_stmt: 'assert' test [',' test] */ |
|---|
| 3426 | n/a | REQ(n, assert_stmt); |
|---|
| 3427 | n/a | if (NCH(n) == 2) { |
|---|
| 3428 | n/a | expr_ty expression = ast_for_expr(c, CHILD(n, 1)); |
|---|
| 3429 | n/a | if (!expression) |
|---|
| 3430 | n/a | return NULL; |
|---|
| 3431 | n/a | return Assert(expression, NULL, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3432 | n/a | } |
|---|
| 3433 | n/a | else if (NCH(n) == 4) { |
|---|
| 3434 | n/a | expr_ty expr1, expr2; |
|---|
| 3435 | n/a | |
|---|
| 3436 | n/a | expr1 = ast_for_expr(c, CHILD(n, 1)); |
|---|
| 3437 | n/a | if (!expr1) |
|---|
| 3438 | n/a | return NULL; |
|---|
| 3439 | n/a | expr2 = ast_for_expr(c, CHILD(n, 3)); |
|---|
| 3440 | n/a | if (!expr2) |
|---|
| 3441 | n/a | return NULL; |
|---|
| 3442 | n/a | |
|---|
| 3443 | n/a | return Assert(expr1, expr2, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3444 | n/a | } |
|---|
| 3445 | n/a | PyErr_Format(PyExc_SystemError, |
|---|
| 3446 | n/a | "improper number of parts to 'assert' statement: %d", |
|---|
| 3447 | n/a | NCH(n)); |
|---|
| 3448 | n/a | return NULL; |
|---|
| 3449 | n/a | } |
|---|
| 3450 | n/a | |
|---|
| 3451 | n/a | static asdl_seq * |
|---|
| 3452 | n/a | ast_for_suite(struct compiling *c, const node *n) |
|---|
| 3453 | n/a | { |
|---|
| 3454 | n/a | /* suite: simple_stmt | NEWLINE INDENT stmt+ DEDENT */ |
|---|
| 3455 | n/a | asdl_seq *seq; |
|---|
| 3456 | n/a | stmt_ty s; |
|---|
| 3457 | n/a | int i, total, num, end, pos = 0; |
|---|
| 3458 | n/a | node *ch; |
|---|
| 3459 | n/a | |
|---|
| 3460 | n/a | REQ(n, suite); |
|---|
| 3461 | n/a | |
|---|
| 3462 | n/a | total = num_stmts(n); |
|---|
| 3463 | n/a | seq = _Py_asdl_seq_new(total, c->c_arena); |
|---|
| 3464 | n/a | if (!seq) |
|---|
| 3465 | n/a | return NULL; |
|---|
| 3466 | n/a | if (TYPE(CHILD(n, 0)) == simple_stmt) { |
|---|
| 3467 | n/a | n = CHILD(n, 0); |
|---|
| 3468 | n/a | /* simple_stmt always ends with a NEWLINE, |
|---|
| 3469 | n/a | and may have a trailing SEMI |
|---|
| 3470 | n/a | */ |
|---|
| 3471 | n/a | end = NCH(n) - 1; |
|---|
| 3472 | n/a | if (TYPE(CHILD(n, end - 1)) == SEMI) |
|---|
| 3473 | n/a | end--; |
|---|
| 3474 | n/a | /* loop by 2 to skip semi-colons */ |
|---|
| 3475 | n/a | for (i = 0; i < end; i += 2) { |
|---|
| 3476 | n/a | ch = CHILD(n, i); |
|---|
| 3477 | n/a | s = ast_for_stmt(c, ch); |
|---|
| 3478 | n/a | if (!s) |
|---|
| 3479 | n/a | return NULL; |
|---|
| 3480 | n/a | asdl_seq_SET(seq, pos++, s); |
|---|
| 3481 | n/a | } |
|---|
| 3482 | n/a | } |
|---|
| 3483 | n/a | else { |
|---|
| 3484 | n/a | for (i = 2; i < (NCH(n) - 1); i++) { |
|---|
| 3485 | n/a | ch = CHILD(n, i); |
|---|
| 3486 | n/a | REQ(ch, stmt); |
|---|
| 3487 | n/a | num = num_stmts(ch); |
|---|
| 3488 | n/a | if (num == 1) { |
|---|
| 3489 | n/a | /* small_stmt or compound_stmt with only one child */ |
|---|
| 3490 | n/a | s = ast_for_stmt(c, ch); |
|---|
| 3491 | n/a | if (!s) |
|---|
| 3492 | n/a | return NULL; |
|---|
| 3493 | n/a | asdl_seq_SET(seq, pos++, s); |
|---|
| 3494 | n/a | } |
|---|
| 3495 | n/a | else { |
|---|
| 3496 | n/a | int j; |
|---|
| 3497 | n/a | ch = CHILD(ch, 0); |
|---|
| 3498 | n/a | REQ(ch, simple_stmt); |
|---|
| 3499 | n/a | for (j = 0; j < NCH(ch); j += 2) { |
|---|
| 3500 | n/a | /* statement terminates with a semi-colon ';' */ |
|---|
| 3501 | n/a | if (NCH(CHILD(ch, j)) == 0) { |
|---|
| 3502 | n/a | assert((j + 1) == NCH(ch)); |
|---|
| 3503 | n/a | break; |
|---|
| 3504 | n/a | } |
|---|
| 3505 | n/a | s = ast_for_stmt(c, CHILD(ch, j)); |
|---|
| 3506 | n/a | if (!s) |
|---|
| 3507 | n/a | return NULL; |
|---|
| 3508 | n/a | asdl_seq_SET(seq, pos++, s); |
|---|
| 3509 | n/a | } |
|---|
| 3510 | n/a | } |
|---|
| 3511 | n/a | } |
|---|
| 3512 | n/a | } |
|---|
| 3513 | n/a | assert(pos == seq->size); |
|---|
| 3514 | n/a | return seq; |
|---|
| 3515 | n/a | } |
|---|
| 3516 | n/a | |
|---|
| 3517 | n/a | static stmt_ty |
|---|
| 3518 | n/a | ast_for_if_stmt(struct compiling *c, const node *n) |
|---|
| 3519 | n/a | { |
|---|
| 3520 | n/a | /* if_stmt: 'if' test ':' suite ('elif' test ':' suite)* |
|---|
| 3521 | n/a | ['else' ':' suite] |
|---|
| 3522 | n/a | */ |
|---|
| 3523 | n/a | char *s; |
|---|
| 3524 | n/a | |
|---|
| 3525 | n/a | REQ(n, if_stmt); |
|---|
| 3526 | n/a | |
|---|
| 3527 | n/a | if (NCH(n) == 4) { |
|---|
| 3528 | n/a | expr_ty expression; |
|---|
| 3529 | n/a | asdl_seq *suite_seq; |
|---|
| 3530 | n/a | |
|---|
| 3531 | n/a | expression = ast_for_expr(c, CHILD(n, 1)); |
|---|
| 3532 | n/a | if (!expression) |
|---|
| 3533 | n/a | return NULL; |
|---|
| 3534 | n/a | suite_seq = ast_for_suite(c, CHILD(n, 3)); |
|---|
| 3535 | n/a | if (!suite_seq) |
|---|
| 3536 | n/a | return NULL; |
|---|
| 3537 | n/a | |
|---|
| 3538 | n/a | return If(expression, suite_seq, NULL, LINENO(n), n->n_col_offset, |
|---|
| 3539 | n/a | c->c_arena); |
|---|
| 3540 | n/a | } |
|---|
| 3541 | n/a | |
|---|
| 3542 | n/a | s = STR(CHILD(n, 4)); |
|---|
| 3543 | n/a | /* s[2], the third character in the string, will be |
|---|
| 3544 | n/a | 's' for el_s_e, or |
|---|
| 3545 | n/a | 'i' for el_i_f |
|---|
| 3546 | n/a | */ |
|---|
| 3547 | n/a | if (s[2] == 's') { |
|---|
| 3548 | n/a | expr_ty expression; |
|---|
| 3549 | n/a | asdl_seq *seq1, *seq2; |
|---|
| 3550 | n/a | |
|---|
| 3551 | n/a | expression = ast_for_expr(c, CHILD(n, 1)); |
|---|
| 3552 | n/a | if (!expression) |
|---|
| 3553 | n/a | return NULL; |
|---|
| 3554 | n/a | seq1 = ast_for_suite(c, CHILD(n, 3)); |
|---|
| 3555 | n/a | if (!seq1) |
|---|
| 3556 | n/a | return NULL; |
|---|
| 3557 | n/a | seq2 = ast_for_suite(c, CHILD(n, 6)); |
|---|
| 3558 | n/a | if (!seq2) |
|---|
| 3559 | n/a | return NULL; |
|---|
| 3560 | n/a | |
|---|
| 3561 | n/a | return If(expression, seq1, seq2, LINENO(n), n->n_col_offset, |
|---|
| 3562 | n/a | c->c_arena); |
|---|
| 3563 | n/a | } |
|---|
| 3564 | n/a | else if (s[2] == 'i') { |
|---|
| 3565 | n/a | int i, n_elif, has_else = 0; |
|---|
| 3566 | n/a | expr_ty expression; |
|---|
| 3567 | n/a | asdl_seq *suite_seq; |
|---|
| 3568 | n/a | asdl_seq *orelse = NULL; |
|---|
| 3569 | n/a | n_elif = NCH(n) - 4; |
|---|
| 3570 | n/a | /* must reference the child n_elif+1 since 'else' token is third, |
|---|
| 3571 | n/a | not fourth, child from the end. */ |
|---|
| 3572 | n/a | if (TYPE(CHILD(n, (n_elif + 1))) == NAME |
|---|
| 3573 | n/a | && STR(CHILD(n, (n_elif + 1)))[2] == 's') { |
|---|
| 3574 | n/a | has_else = 1; |
|---|
| 3575 | n/a | n_elif -= 3; |
|---|
| 3576 | n/a | } |
|---|
| 3577 | n/a | n_elif /= 4; |
|---|
| 3578 | n/a | |
|---|
| 3579 | n/a | if (has_else) { |
|---|
| 3580 | n/a | asdl_seq *suite_seq2; |
|---|
| 3581 | n/a | |
|---|
| 3582 | n/a | orelse = _Py_asdl_seq_new(1, c->c_arena); |
|---|
| 3583 | n/a | if (!orelse) |
|---|
| 3584 | n/a | return NULL; |
|---|
| 3585 | n/a | expression = ast_for_expr(c, CHILD(n, NCH(n) - 6)); |
|---|
| 3586 | n/a | if (!expression) |
|---|
| 3587 | n/a | return NULL; |
|---|
| 3588 | n/a | suite_seq = ast_for_suite(c, CHILD(n, NCH(n) - 4)); |
|---|
| 3589 | n/a | if (!suite_seq) |
|---|
| 3590 | n/a | return NULL; |
|---|
| 3591 | n/a | suite_seq2 = ast_for_suite(c, CHILD(n, NCH(n) - 1)); |
|---|
| 3592 | n/a | if (!suite_seq2) |
|---|
| 3593 | n/a | return NULL; |
|---|
| 3594 | n/a | |
|---|
| 3595 | n/a | asdl_seq_SET(orelse, 0, |
|---|
| 3596 | n/a | If(expression, suite_seq, suite_seq2, |
|---|
| 3597 | n/a | LINENO(CHILD(n, NCH(n) - 6)), |
|---|
| 3598 | n/a | CHILD(n, NCH(n) - 6)->n_col_offset, |
|---|
| 3599 | n/a | c->c_arena)); |
|---|
| 3600 | n/a | /* the just-created orelse handled the last elif */ |
|---|
| 3601 | n/a | n_elif--; |
|---|
| 3602 | n/a | } |
|---|
| 3603 | n/a | |
|---|
| 3604 | n/a | for (i = 0; i < n_elif; i++) { |
|---|
| 3605 | n/a | int off = 5 + (n_elif - i - 1) * 4; |
|---|
| 3606 | n/a | asdl_seq *newobj = _Py_asdl_seq_new(1, c->c_arena); |
|---|
| 3607 | n/a | if (!newobj) |
|---|
| 3608 | n/a | return NULL; |
|---|
| 3609 | n/a | expression = ast_for_expr(c, CHILD(n, off)); |
|---|
| 3610 | n/a | if (!expression) |
|---|
| 3611 | n/a | return NULL; |
|---|
| 3612 | n/a | suite_seq = ast_for_suite(c, CHILD(n, off + 2)); |
|---|
| 3613 | n/a | if (!suite_seq) |
|---|
| 3614 | n/a | return NULL; |
|---|
| 3615 | n/a | |
|---|
| 3616 | n/a | asdl_seq_SET(newobj, 0, |
|---|
| 3617 | n/a | If(expression, suite_seq, orelse, |
|---|
| 3618 | n/a | LINENO(CHILD(n, off)), |
|---|
| 3619 | n/a | CHILD(n, off)->n_col_offset, c->c_arena)); |
|---|
| 3620 | n/a | orelse = newobj; |
|---|
| 3621 | n/a | } |
|---|
| 3622 | n/a | expression = ast_for_expr(c, CHILD(n, 1)); |
|---|
| 3623 | n/a | if (!expression) |
|---|
| 3624 | n/a | return NULL; |
|---|
| 3625 | n/a | suite_seq = ast_for_suite(c, CHILD(n, 3)); |
|---|
| 3626 | n/a | if (!suite_seq) |
|---|
| 3627 | n/a | return NULL; |
|---|
| 3628 | n/a | return If(expression, suite_seq, orelse, |
|---|
| 3629 | n/a | LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3630 | n/a | } |
|---|
| 3631 | n/a | |
|---|
| 3632 | n/a | PyErr_Format(PyExc_SystemError, |
|---|
| 3633 | n/a | "unexpected token in 'if' statement: %s", s); |
|---|
| 3634 | n/a | return NULL; |
|---|
| 3635 | n/a | } |
|---|
| 3636 | n/a | |
|---|
| 3637 | n/a | static stmt_ty |
|---|
| 3638 | n/a | ast_for_while_stmt(struct compiling *c, const node *n) |
|---|
| 3639 | n/a | { |
|---|
| 3640 | n/a | /* while_stmt: 'while' test ':' suite ['else' ':' suite] */ |
|---|
| 3641 | n/a | REQ(n, while_stmt); |
|---|
| 3642 | n/a | |
|---|
| 3643 | n/a | if (NCH(n) == 4) { |
|---|
| 3644 | n/a | expr_ty expression; |
|---|
| 3645 | n/a | asdl_seq *suite_seq; |
|---|
| 3646 | n/a | |
|---|
| 3647 | n/a | expression = ast_for_expr(c, CHILD(n, 1)); |
|---|
| 3648 | n/a | if (!expression) |
|---|
| 3649 | n/a | return NULL; |
|---|
| 3650 | n/a | suite_seq = ast_for_suite(c, CHILD(n, 3)); |
|---|
| 3651 | n/a | if (!suite_seq) |
|---|
| 3652 | n/a | return NULL; |
|---|
| 3653 | n/a | return While(expression, suite_seq, NULL, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3654 | n/a | } |
|---|
| 3655 | n/a | else if (NCH(n) == 7) { |
|---|
| 3656 | n/a | expr_ty expression; |
|---|
| 3657 | n/a | asdl_seq *seq1, *seq2; |
|---|
| 3658 | n/a | |
|---|
| 3659 | n/a | expression = ast_for_expr(c, CHILD(n, 1)); |
|---|
| 3660 | n/a | if (!expression) |
|---|
| 3661 | n/a | return NULL; |
|---|
| 3662 | n/a | seq1 = ast_for_suite(c, CHILD(n, 3)); |
|---|
| 3663 | n/a | if (!seq1) |
|---|
| 3664 | n/a | return NULL; |
|---|
| 3665 | n/a | seq2 = ast_for_suite(c, CHILD(n, 6)); |
|---|
| 3666 | n/a | if (!seq2) |
|---|
| 3667 | n/a | return NULL; |
|---|
| 3668 | n/a | |
|---|
| 3669 | n/a | return While(expression, seq1, seq2, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3670 | n/a | } |
|---|
| 3671 | n/a | |
|---|
| 3672 | n/a | PyErr_Format(PyExc_SystemError, |
|---|
| 3673 | n/a | "wrong number of tokens for 'while' statement: %d", |
|---|
| 3674 | n/a | NCH(n)); |
|---|
| 3675 | n/a | return NULL; |
|---|
| 3676 | n/a | } |
|---|
| 3677 | n/a | |
|---|
| 3678 | n/a | static stmt_ty |
|---|
| 3679 | n/a | ast_for_for_stmt(struct compiling *c, const node *n, int is_async) |
|---|
| 3680 | n/a | { |
|---|
| 3681 | n/a | asdl_seq *_target, *seq = NULL, *suite_seq; |
|---|
| 3682 | n/a | expr_ty expression; |
|---|
| 3683 | n/a | expr_ty target, first; |
|---|
| 3684 | n/a | const node *node_target; |
|---|
| 3685 | n/a | /* for_stmt: 'for' exprlist 'in' testlist ':' suite ['else' ':' suite] */ |
|---|
| 3686 | n/a | REQ(n, for_stmt); |
|---|
| 3687 | n/a | |
|---|
| 3688 | n/a | if (NCH(n) == 9) { |
|---|
| 3689 | n/a | seq = ast_for_suite(c, CHILD(n, 8)); |
|---|
| 3690 | n/a | if (!seq) |
|---|
| 3691 | n/a | return NULL; |
|---|
| 3692 | n/a | } |
|---|
| 3693 | n/a | |
|---|
| 3694 | n/a | node_target = CHILD(n, 1); |
|---|
| 3695 | n/a | _target = ast_for_exprlist(c, node_target, Store); |
|---|
| 3696 | n/a | if (!_target) |
|---|
| 3697 | n/a | return NULL; |
|---|
| 3698 | n/a | /* Check the # of children rather than the length of _target, since |
|---|
| 3699 | n/a | for x, in ... has 1 element in _target, but still requires a Tuple. */ |
|---|
| 3700 | n/a | first = (expr_ty)asdl_seq_GET(_target, 0); |
|---|
| 3701 | n/a | if (NCH(node_target) == 1) |
|---|
| 3702 | n/a | target = first; |
|---|
| 3703 | n/a | else |
|---|
| 3704 | n/a | target = Tuple(_target, Store, first->lineno, first->col_offset, c->c_arena); |
|---|
| 3705 | n/a | |
|---|
| 3706 | n/a | expression = ast_for_testlist(c, CHILD(n, 3)); |
|---|
| 3707 | n/a | if (!expression) |
|---|
| 3708 | n/a | return NULL; |
|---|
| 3709 | n/a | suite_seq = ast_for_suite(c, CHILD(n, 5)); |
|---|
| 3710 | n/a | if (!suite_seq) |
|---|
| 3711 | n/a | return NULL; |
|---|
| 3712 | n/a | |
|---|
| 3713 | n/a | if (is_async) |
|---|
| 3714 | n/a | return AsyncFor(target, expression, suite_seq, seq, |
|---|
| 3715 | n/a | LINENO(n), n->n_col_offset, |
|---|
| 3716 | n/a | c->c_arena); |
|---|
| 3717 | n/a | else |
|---|
| 3718 | n/a | return For(target, expression, suite_seq, seq, |
|---|
| 3719 | n/a | LINENO(n), n->n_col_offset, |
|---|
| 3720 | n/a | c->c_arena); |
|---|
| 3721 | n/a | } |
|---|
| 3722 | n/a | |
|---|
| 3723 | n/a | static excepthandler_ty |
|---|
| 3724 | n/a | ast_for_except_clause(struct compiling *c, const node *exc, node *body) |
|---|
| 3725 | n/a | { |
|---|
| 3726 | n/a | /* except_clause: 'except' [test ['as' test]] */ |
|---|
| 3727 | n/a | REQ(exc, except_clause); |
|---|
| 3728 | n/a | REQ(body, suite); |
|---|
| 3729 | n/a | |
|---|
| 3730 | n/a | if (NCH(exc) == 1) { |
|---|
| 3731 | n/a | asdl_seq *suite_seq = ast_for_suite(c, body); |
|---|
| 3732 | n/a | if (!suite_seq) |
|---|
| 3733 | n/a | return NULL; |
|---|
| 3734 | n/a | |
|---|
| 3735 | n/a | return ExceptHandler(NULL, NULL, suite_seq, LINENO(exc), |
|---|
| 3736 | n/a | exc->n_col_offset, c->c_arena); |
|---|
| 3737 | n/a | } |
|---|
| 3738 | n/a | else if (NCH(exc) == 2) { |
|---|
| 3739 | n/a | expr_ty expression; |
|---|
| 3740 | n/a | asdl_seq *suite_seq; |
|---|
| 3741 | n/a | |
|---|
| 3742 | n/a | expression = ast_for_expr(c, CHILD(exc, 1)); |
|---|
| 3743 | n/a | if (!expression) |
|---|
| 3744 | n/a | return NULL; |
|---|
| 3745 | n/a | suite_seq = ast_for_suite(c, body); |
|---|
| 3746 | n/a | if (!suite_seq) |
|---|
| 3747 | n/a | return NULL; |
|---|
| 3748 | n/a | |
|---|
| 3749 | n/a | return ExceptHandler(expression, NULL, suite_seq, LINENO(exc), |
|---|
| 3750 | n/a | exc->n_col_offset, c->c_arena); |
|---|
| 3751 | n/a | } |
|---|
| 3752 | n/a | else if (NCH(exc) == 4) { |
|---|
| 3753 | n/a | asdl_seq *suite_seq; |
|---|
| 3754 | n/a | expr_ty expression; |
|---|
| 3755 | n/a | identifier e = NEW_IDENTIFIER(CHILD(exc, 3)); |
|---|
| 3756 | n/a | if (!e) |
|---|
| 3757 | n/a | return NULL; |
|---|
| 3758 | n/a | if (forbidden_name(c, e, CHILD(exc, 3), 0)) |
|---|
| 3759 | n/a | return NULL; |
|---|
| 3760 | n/a | expression = ast_for_expr(c, CHILD(exc, 1)); |
|---|
| 3761 | n/a | if (!expression) |
|---|
| 3762 | n/a | return NULL; |
|---|
| 3763 | n/a | suite_seq = ast_for_suite(c, body); |
|---|
| 3764 | n/a | if (!suite_seq) |
|---|
| 3765 | n/a | return NULL; |
|---|
| 3766 | n/a | |
|---|
| 3767 | n/a | return ExceptHandler(expression, e, suite_seq, LINENO(exc), |
|---|
| 3768 | n/a | exc->n_col_offset, c->c_arena); |
|---|
| 3769 | n/a | } |
|---|
| 3770 | n/a | |
|---|
| 3771 | n/a | PyErr_Format(PyExc_SystemError, |
|---|
| 3772 | n/a | "wrong number of children for 'except' clause: %d", |
|---|
| 3773 | n/a | NCH(exc)); |
|---|
| 3774 | n/a | return NULL; |
|---|
| 3775 | n/a | } |
|---|
| 3776 | n/a | |
|---|
| 3777 | n/a | static stmt_ty |
|---|
| 3778 | n/a | ast_for_try_stmt(struct compiling *c, const node *n) |
|---|
| 3779 | n/a | { |
|---|
| 3780 | n/a | const int nch = NCH(n); |
|---|
| 3781 | n/a | int n_except = (nch - 3)/3; |
|---|
| 3782 | n/a | asdl_seq *body, *handlers = NULL, *orelse = NULL, *finally = NULL; |
|---|
| 3783 | n/a | |
|---|
| 3784 | n/a | REQ(n, try_stmt); |
|---|
| 3785 | n/a | |
|---|
| 3786 | n/a | body = ast_for_suite(c, CHILD(n, 2)); |
|---|
| 3787 | n/a | if (body == NULL) |
|---|
| 3788 | n/a | return NULL; |
|---|
| 3789 | n/a | |
|---|
| 3790 | n/a | if (TYPE(CHILD(n, nch - 3)) == NAME) { |
|---|
| 3791 | n/a | if (strcmp(STR(CHILD(n, nch - 3)), "finally") == 0) { |
|---|
| 3792 | n/a | if (nch >= 9 && TYPE(CHILD(n, nch - 6)) == NAME) { |
|---|
| 3793 | n/a | /* we can assume it's an "else", |
|---|
| 3794 | n/a | because nch >= 9 for try-else-finally and |
|---|
| 3795 | n/a | it would otherwise have a type of except_clause */ |
|---|
| 3796 | n/a | orelse = ast_for_suite(c, CHILD(n, nch - 4)); |
|---|
| 3797 | n/a | if (orelse == NULL) |
|---|
| 3798 | n/a | return NULL; |
|---|
| 3799 | n/a | n_except--; |
|---|
| 3800 | n/a | } |
|---|
| 3801 | n/a | |
|---|
| 3802 | n/a | finally = ast_for_suite(c, CHILD(n, nch - 1)); |
|---|
| 3803 | n/a | if (finally == NULL) |
|---|
| 3804 | n/a | return NULL; |
|---|
| 3805 | n/a | n_except--; |
|---|
| 3806 | n/a | } |
|---|
| 3807 | n/a | else { |
|---|
| 3808 | n/a | /* we can assume it's an "else", |
|---|
| 3809 | n/a | otherwise it would have a type of except_clause */ |
|---|
| 3810 | n/a | orelse = ast_for_suite(c, CHILD(n, nch - 1)); |
|---|
| 3811 | n/a | if (orelse == NULL) |
|---|
| 3812 | n/a | return NULL; |
|---|
| 3813 | n/a | n_except--; |
|---|
| 3814 | n/a | } |
|---|
| 3815 | n/a | } |
|---|
| 3816 | n/a | else if (TYPE(CHILD(n, nch - 3)) != except_clause) { |
|---|
| 3817 | n/a | ast_error(c, n, "malformed 'try' statement"); |
|---|
| 3818 | n/a | return NULL; |
|---|
| 3819 | n/a | } |
|---|
| 3820 | n/a | |
|---|
| 3821 | n/a | if (n_except > 0) { |
|---|
| 3822 | n/a | int i; |
|---|
| 3823 | n/a | /* process except statements to create a try ... except */ |
|---|
| 3824 | n/a | handlers = _Py_asdl_seq_new(n_except, c->c_arena); |
|---|
| 3825 | n/a | if (handlers == NULL) |
|---|
| 3826 | n/a | return NULL; |
|---|
| 3827 | n/a | |
|---|
| 3828 | n/a | for (i = 0; i < n_except; i++) { |
|---|
| 3829 | n/a | excepthandler_ty e = ast_for_except_clause(c, CHILD(n, 3 + i * 3), |
|---|
| 3830 | n/a | CHILD(n, 5 + i * 3)); |
|---|
| 3831 | n/a | if (!e) |
|---|
| 3832 | n/a | return NULL; |
|---|
| 3833 | n/a | asdl_seq_SET(handlers, i, e); |
|---|
| 3834 | n/a | } |
|---|
| 3835 | n/a | } |
|---|
| 3836 | n/a | |
|---|
| 3837 | n/a | assert(finally != NULL || asdl_seq_LEN(handlers)); |
|---|
| 3838 | n/a | return Try(body, handlers, orelse, finally, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3839 | n/a | } |
|---|
| 3840 | n/a | |
|---|
| 3841 | n/a | /* with_item: test ['as' expr] */ |
|---|
| 3842 | n/a | static withitem_ty |
|---|
| 3843 | n/a | ast_for_with_item(struct compiling *c, const node *n) |
|---|
| 3844 | n/a | { |
|---|
| 3845 | n/a | expr_ty context_expr, optional_vars = NULL; |
|---|
| 3846 | n/a | |
|---|
| 3847 | n/a | REQ(n, with_item); |
|---|
| 3848 | n/a | context_expr = ast_for_expr(c, CHILD(n, 0)); |
|---|
| 3849 | n/a | if (!context_expr) |
|---|
| 3850 | n/a | return NULL; |
|---|
| 3851 | n/a | if (NCH(n) == 3) { |
|---|
| 3852 | n/a | optional_vars = ast_for_expr(c, CHILD(n, 2)); |
|---|
| 3853 | n/a | |
|---|
| 3854 | n/a | if (!optional_vars) { |
|---|
| 3855 | n/a | return NULL; |
|---|
| 3856 | n/a | } |
|---|
| 3857 | n/a | if (!set_context(c, optional_vars, Store, n)) { |
|---|
| 3858 | n/a | return NULL; |
|---|
| 3859 | n/a | } |
|---|
| 3860 | n/a | } |
|---|
| 3861 | n/a | |
|---|
| 3862 | n/a | return withitem(context_expr, optional_vars, c->c_arena); |
|---|
| 3863 | n/a | } |
|---|
| 3864 | n/a | |
|---|
| 3865 | n/a | /* with_stmt: 'with' with_item (',' with_item)* ':' suite */ |
|---|
| 3866 | n/a | static stmt_ty |
|---|
| 3867 | n/a | ast_for_with_stmt(struct compiling *c, const node *n, int is_async) |
|---|
| 3868 | n/a | { |
|---|
| 3869 | n/a | int i, n_items; |
|---|
| 3870 | n/a | asdl_seq *items, *body; |
|---|
| 3871 | n/a | |
|---|
| 3872 | n/a | REQ(n, with_stmt); |
|---|
| 3873 | n/a | |
|---|
| 3874 | n/a | n_items = (NCH(n) - 2) / 2; |
|---|
| 3875 | n/a | items = _Py_asdl_seq_new(n_items, c->c_arena); |
|---|
| 3876 | n/a | if (!items) |
|---|
| 3877 | n/a | return NULL; |
|---|
| 3878 | n/a | for (i = 1; i < NCH(n) - 2; i += 2) { |
|---|
| 3879 | n/a | withitem_ty item = ast_for_with_item(c, CHILD(n, i)); |
|---|
| 3880 | n/a | if (!item) |
|---|
| 3881 | n/a | return NULL; |
|---|
| 3882 | n/a | asdl_seq_SET(items, (i - 1) / 2, item); |
|---|
| 3883 | n/a | } |
|---|
| 3884 | n/a | |
|---|
| 3885 | n/a | body = ast_for_suite(c, CHILD(n, NCH(n) - 1)); |
|---|
| 3886 | n/a | if (!body) |
|---|
| 3887 | n/a | return NULL; |
|---|
| 3888 | n/a | |
|---|
| 3889 | n/a | if (is_async) |
|---|
| 3890 | n/a | return AsyncWith(items, body, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3891 | n/a | else |
|---|
| 3892 | n/a | return With(items, body, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3893 | n/a | } |
|---|
| 3894 | n/a | |
|---|
| 3895 | n/a | static stmt_ty |
|---|
| 3896 | n/a | ast_for_classdef(struct compiling *c, const node *n, asdl_seq *decorator_seq) |
|---|
| 3897 | n/a | { |
|---|
| 3898 | n/a | /* classdef: 'class' NAME ['(' arglist ')'] ':' suite */ |
|---|
| 3899 | n/a | PyObject *classname; |
|---|
| 3900 | n/a | asdl_seq *s; |
|---|
| 3901 | n/a | expr_ty call; |
|---|
| 3902 | n/a | |
|---|
| 3903 | n/a | REQ(n, classdef); |
|---|
| 3904 | n/a | |
|---|
| 3905 | n/a | if (NCH(n) == 4) { /* class NAME ':' suite */ |
|---|
| 3906 | n/a | s = ast_for_suite(c, CHILD(n, 3)); |
|---|
| 3907 | n/a | if (!s) |
|---|
| 3908 | n/a | return NULL; |
|---|
| 3909 | n/a | classname = NEW_IDENTIFIER(CHILD(n, 1)); |
|---|
| 3910 | n/a | if (!classname) |
|---|
| 3911 | n/a | return NULL; |
|---|
| 3912 | n/a | if (forbidden_name(c, classname, CHILD(n, 3), 0)) |
|---|
| 3913 | n/a | return NULL; |
|---|
| 3914 | n/a | return ClassDef(classname, NULL, NULL, s, decorator_seq, LINENO(n), |
|---|
| 3915 | n/a | n->n_col_offset, c->c_arena); |
|---|
| 3916 | n/a | } |
|---|
| 3917 | n/a | |
|---|
| 3918 | n/a | if (TYPE(CHILD(n, 3)) == RPAR) { /* class NAME '(' ')' ':' suite */ |
|---|
| 3919 | n/a | s = ast_for_suite(c, CHILD(n,5)); |
|---|
| 3920 | n/a | if (!s) |
|---|
| 3921 | n/a | return NULL; |
|---|
| 3922 | n/a | classname = NEW_IDENTIFIER(CHILD(n, 1)); |
|---|
| 3923 | n/a | if (!classname) |
|---|
| 3924 | n/a | return NULL; |
|---|
| 3925 | n/a | if (forbidden_name(c, classname, CHILD(n, 3), 0)) |
|---|
| 3926 | n/a | return NULL; |
|---|
| 3927 | n/a | return ClassDef(classname, NULL, NULL, s, decorator_seq, LINENO(n), |
|---|
| 3928 | n/a | n->n_col_offset, c->c_arena); |
|---|
| 3929 | n/a | } |
|---|
| 3930 | n/a | |
|---|
| 3931 | n/a | /* class NAME '(' arglist ')' ':' suite */ |
|---|
| 3932 | n/a | /* build up a fake Call node so we can extract its pieces */ |
|---|
| 3933 | n/a | { |
|---|
| 3934 | n/a | PyObject *dummy_name; |
|---|
| 3935 | n/a | expr_ty dummy; |
|---|
| 3936 | n/a | dummy_name = NEW_IDENTIFIER(CHILD(n, 1)); |
|---|
| 3937 | n/a | if (!dummy_name) |
|---|
| 3938 | n/a | return NULL; |
|---|
| 3939 | n/a | dummy = Name(dummy_name, Load, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3940 | n/a | call = ast_for_call(c, CHILD(n, 3), dummy); |
|---|
| 3941 | n/a | if (!call) |
|---|
| 3942 | n/a | return NULL; |
|---|
| 3943 | n/a | } |
|---|
| 3944 | n/a | s = ast_for_suite(c, CHILD(n, 6)); |
|---|
| 3945 | n/a | if (!s) |
|---|
| 3946 | n/a | return NULL; |
|---|
| 3947 | n/a | classname = NEW_IDENTIFIER(CHILD(n, 1)); |
|---|
| 3948 | n/a | if (!classname) |
|---|
| 3949 | n/a | return NULL; |
|---|
| 3950 | n/a | if (forbidden_name(c, classname, CHILD(n, 1), 0)) |
|---|
| 3951 | n/a | return NULL; |
|---|
| 3952 | n/a | |
|---|
| 3953 | n/a | return ClassDef(classname, call->v.Call.args, call->v.Call.keywords, s, |
|---|
| 3954 | n/a | decorator_seq, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3955 | n/a | } |
|---|
| 3956 | n/a | |
|---|
| 3957 | n/a | static stmt_ty |
|---|
| 3958 | n/a | ast_for_stmt(struct compiling *c, const node *n) |
|---|
| 3959 | n/a | { |
|---|
| 3960 | n/a | if (TYPE(n) == stmt) { |
|---|
| 3961 | n/a | assert(NCH(n) == 1); |
|---|
| 3962 | n/a | n = CHILD(n, 0); |
|---|
| 3963 | n/a | } |
|---|
| 3964 | n/a | if (TYPE(n) == simple_stmt) { |
|---|
| 3965 | n/a | assert(num_stmts(n) == 1); |
|---|
| 3966 | n/a | n = CHILD(n, 0); |
|---|
| 3967 | n/a | } |
|---|
| 3968 | n/a | if (TYPE(n) == small_stmt) { |
|---|
| 3969 | n/a | n = CHILD(n, 0); |
|---|
| 3970 | n/a | /* small_stmt: expr_stmt | del_stmt | pass_stmt | flow_stmt |
|---|
| 3971 | n/a | | import_stmt | global_stmt | nonlocal_stmt | assert_stmt |
|---|
| 3972 | n/a | */ |
|---|
| 3973 | n/a | switch (TYPE(n)) { |
|---|
| 3974 | n/a | case expr_stmt: |
|---|
| 3975 | n/a | return ast_for_expr_stmt(c, n); |
|---|
| 3976 | n/a | case del_stmt: |
|---|
| 3977 | n/a | return ast_for_del_stmt(c, n); |
|---|
| 3978 | n/a | case pass_stmt: |
|---|
| 3979 | n/a | return Pass(LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 3980 | n/a | case flow_stmt: |
|---|
| 3981 | n/a | return ast_for_flow_stmt(c, n); |
|---|
| 3982 | n/a | case import_stmt: |
|---|
| 3983 | n/a | return ast_for_import_stmt(c, n); |
|---|
| 3984 | n/a | case global_stmt: |
|---|
| 3985 | n/a | return ast_for_global_stmt(c, n); |
|---|
| 3986 | n/a | case nonlocal_stmt: |
|---|
| 3987 | n/a | return ast_for_nonlocal_stmt(c, n); |
|---|
| 3988 | n/a | case assert_stmt: |
|---|
| 3989 | n/a | return ast_for_assert_stmt(c, n); |
|---|
| 3990 | n/a | default: |
|---|
| 3991 | n/a | PyErr_Format(PyExc_SystemError, |
|---|
| 3992 | n/a | "unhandled small_stmt: TYPE=%d NCH=%d\n", |
|---|
| 3993 | n/a | TYPE(n), NCH(n)); |
|---|
| 3994 | n/a | return NULL; |
|---|
| 3995 | n/a | } |
|---|
| 3996 | n/a | } |
|---|
| 3997 | n/a | else { |
|---|
| 3998 | n/a | /* compound_stmt: if_stmt | while_stmt | for_stmt | try_stmt |
|---|
| 3999 | n/a | | funcdef | classdef | decorated | async_stmt |
|---|
| 4000 | n/a | */ |
|---|
| 4001 | n/a | node *ch = CHILD(n, 0); |
|---|
| 4002 | n/a | REQ(n, compound_stmt); |
|---|
| 4003 | n/a | switch (TYPE(ch)) { |
|---|
| 4004 | n/a | case if_stmt: |
|---|
| 4005 | n/a | return ast_for_if_stmt(c, ch); |
|---|
| 4006 | n/a | case while_stmt: |
|---|
| 4007 | n/a | return ast_for_while_stmt(c, ch); |
|---|
| 4008 | n/a | case for_stmt: |
|---|
| 4009 | n/a | return ast_for_for_stmt(c, ch, 0); |
|---|
| 4010 | n/a | case try_stmt: |
|---|
| 4011 | n/a | return ast_for_try_stmt(c, ch); |
|---|
| 4012 | n/a | case with_stmt: |
|---|
| 4013 | n/a | return ast_for_with_stmt(c, ch, 0); |
|---|
| 4014 | n/a | case funcdef: |
|---|
| 4015 | n/a | return ast_for_funcdef(c, ch, NULL); |
|---|
| 4016 | n/a | case classdef: |
|---|
| 4017 | n/a | return ast_for_classdef(c, ch, NULL); |
|---|
| 4018 | n/a | case decorated: |
|---|
| 4019 | n/a | return ast_for_decorated(c, ch); |
|---|
| 4020 | n/a | case async_stmt: |
|---|
| 4021 | n/a | return ast_for_async_stmt(c, ch); |
|---|
| 4022 | n/a | default: |
|---|
| 4023 | n/a | PyErr_Format(PyExc_SystemError, |
|---|
| 4024 | n/a | "unhandled small_stmt: TYPE=%d NCH=%d\n", |
|---|
| 4025 | n/a | TYPE(n), NCH(n)); |
|---|
| 4026 | n/a | return NULL; |
|---|
| 4027 | n/a | } |
|---|
| 4028 | n/a | } |
|---|
| 4029 | n/a | } |
|---|
| 4030 | n/a | |
|---|
| 4031 | n/a | static PyObject * |
|---|
| 4032 | n/a | parsenumber_raw(struct compiling *c, const char *s) |
|---|
| 4033 | n/a | { |
|---|
| 4034 | n/a | const char *end; |
|---|
| 4035 | n/a | long x; |
|---|
| 4036 | n/a | double dx; |
|---|
| 4037 | n/a | Py_complex compl; |
|---|
| 4038 | n/a | int imflag; |
|---|
| 4039 | n/a | |
|---|
| 4040 | n/a | assert(s != NULL); |
|---|
| 4041 | n/a | errno = 0; |
|---|
| 4042 | n/a | end = s + strlen(s) - 1; |
|---|
| 4043 | n/a | imflag = *end == 'j' || *end == 'J'; |
|---|
| 4044 | n/a | if (s[0] == '0') { |
|---|
| 4045 | n/a | x = (long) PyOS_strtoul(s, (char **)&end, 0); |
|---|
| 4046 | n/a | if (x < 0 && errno == 0) { |
|---|
| 4047 | n/a | return PyLong_FromString(s, (char **)0, 0); |
|---|
| 4048 | n/a | } |
|---|
| 4049 | n/a | } |
|---|
| 4050 | n/a | else |
|---|
| 4051 | n/a | x = PyOS_strtol(s, (char **)&end, 0); |
|---|
| 4052 | n/a | if (*end == '\0') { |
|---|
| 4053 | n/a | if (errno != 0) |
|---|
| 4054 | n/a | return PyLong_FromString(s, (char **)0, 0); |
|---|
| 4055 | n/a | return PyLong_FromLong(x); |
|---|
| 4056 | n/a | } |
|---|
| 4057 | n/a | /* XXX Huge floats may silently fail */ |
|---|
| 4058 | n/a | if (imflag) { |
|---|
| 4059 | n/a | compl.real = 0.; |
|---|
| 4060 | n/a | compl.imag = PyOS_string_to_double(s, (char **)&end, NULL); |
|---|
| 4061 | n/a | if (compl.imag == -1.0 && PyErr_Occurred()) |
|---|
| 4062 | n/a | return NULL; |
|---|
| 4063 | n/a | return PyComplex_FromCComplex(compl); |
|---|
| 4064 | n/a | } |
|---|
| 4065 | n/a | else |
|---|
| 4066 | n/a | { |
|---|
| 4067 | n/a | dx = PyOS_string_to_double(s, NULL, NULL); |
|---|
| 4068 | n/a | if (dx == -1.0 && PyErr_Occurred()) |
|---|
| 4069 | n/a | return NULL; |
|---|
| 4070 | n/a | return PyFloat_FromDouble(dx); |
|---|
| 4071 | n/a | } |
|---|
| 4072 | n/a | } |
|---|
| 4073 | n/a | |
|---|
| 4074 | n/a | static PyObject * |
|---|
| 4075 | n/a | parsenumber(struct compiling *c, const char *s) |
|---|
| 4076 | n/a | { |
|---|
| 4077 | n/a | char *dup, *end; |
|---|
| 4078 | n/a | PyObject *res = NULL; |
|---|
| 4079 | n/a | |
|---|
| 4080 | n/a | assert(s != NULL); |
|---|
| 4081 | n/a | |
|---|
| 4082 | n/a | if (strchr(s, '_') == NULL) { |
|---|
| 4083 | n/a | return parsenumber_raw(c, s); |
|---|
| 4084 | n/a | } |
|---|
| 4085 | n/a | /* Create a duplicate without underscores. */ |
|---|
| 4086 | n/a | dup = PyMem_Malloc(strlen(s) + 1); |
|---|
| 4087 | n/a | end = dup; |
|---|
| 4088 | n/a | for (; *s; s++) { |
|---|
| 4089 | n/a | if (*s != '_') { |
|---|
| 4090 | n/a | *end++ = *s; |
|---|
| 4091 | n/a | } |
|---|
| 4092 | n/a | } |
|---|
| 4093 | n/a | *end = '\0'; |
|---|
| 4094 | n/a | res = parsenumber_raw(c, dup); |
|---|
| 4095 | n/a | PyMem_Free(dup); |
|---|
| 4096 | n/a | return res; |
|---|
| 4097 | n/a | } |
|---|
| 4098 | n/a | |
|---|
| 4099 | n/a | static PyObject * |
|---|
| 4100 | n/a | decode_utf8(struct compiling *c, const char **sPtr, const char *end) |
|---|
| 4101 | n/a | { |
|---|
| 4102 | n/a | const char *s, *t; |
|---|
| 4103 | n/a | t = s = *sPtr; |
|---|
| 4104 | n/a | /* while (s < end && *s != '\\') s++; */ /* inefficient for u".." */ |
|---|
| 4105 | n/a | while (s < end && (*s & 0x80)) s++; |
|---|
| 4106 | n/a | *sPtr = s; |
|---|
| 4107 | n/a | return PyUnicode_DecodeUTF8(t, s - t, NULL); |
|---|
| 4108 | n/a | } |
|---|
| 4109 | n/a | |
|---|
| 4110 | n/a | static int |
|---|
| 4111 | n/a | warn_invalid_escape_sequence(struct compiling *c, const node *n, |
|---|
| 4112 | n/a | char first_invalid_escape_char) |
|---|
| 4113 | n/a | { |
|---|
| 4114 | n/a | PyObject *msg = PyUnicode_FromFormat("invalid escape sequence \\%c", |
|---|
| 4115 | n/a | first_invalid_escape_char); |
|---|
| 4116 | n/a | if (msg == NULL) { |
|---|
| 4117 | n/a | return -1; |
|---|
| 4118 | n/a | } |
|---|
| 4119 | n/a | if (PyErr_WarnExplicitObject(PyExc_DeprecationWarning, msg, |
|---|
| 4120 | n/a | c->c_filename, LINENO(n), |
|---|
| 4121 | n/a | NULL, NULL) < 0 && |
|---|
| 4122 | n/a | PyErr_ExceptionMatches(PyExc_DeprecationWarning)) |
|---|
| 4123 | n/a | { |
|---|
| 4124 | n/a | const char *s; |
|---|
| 4125 | n/a | |
|---|
| 4126 | n/a | /* Replace the DeprecationWarning exception with a SyntaxError |
|---|
| 4127 | n/a | to get a more accurate error report */ |
|---|
| 4128 | n/a | PyErr_Clear(); |
|---|
| 4129 | n/a | |
|---|
| 4130 | n/a | s = PyUnicode_AsUTF8(msg); |
|---|
| 4131 | n/a | if (s != NULL) { |
|---|
| 4132 | n/a | ast_error(c, n, s); |
|---|
| 4133 | n/a | } |
|---|
| 4134 | n/a | Py_DECREF(msg); |
|---|
| 4135 | n/a | return -1; |
|---|
| 4136 | n/a | } |
|---|
| 4137 | n/a | Py_DECREF(msg); |
|---|
| 4138 | n/a | return 0; |
|---|
| 4139 | n/a | } |
|---|
| 4140 | n/a | |
|---|
| 4141 | n/a | static PyObject * |
|---|
| 4142 | n/a | decode_unicode_with_escapes(struct compiling *c, const node *n, const char *s, |
|---|
| 4143 | n/a | size_t len) |
|---|
| 4144 | n/a | { |
|---|
| 4145 | n/a | PyObject *v, *u; |
|---|
| 4146 | n/a | char *buf; |
|---|
| 4147 | n/a | char *p; |
|---|
| 4148 | n/a | const char *end; |
|---|
| 4149 | n/a | |
|---|
| 4150 | n/a | /* check for integer overflow */ |
|---|
| 4151 | n/a | if (len > SIZE_MAX / 6) |
|---|
| 4152 | n/a | return NULL; |
|---|
| 4153 | n/a | /* "รยค" (2 bytes) may become "\U000000E4" (10 bytes), or 1:5 |
|---|
| 4154 | n/a | "\รยค" (3 bytes) may become "\u005c\U000000E4" (16 bytes), or ~1:6 */ |
|---|
| 4155 | n/a | u = PyBytes_FromStringAndSize((char *)NULL, len * 6); |
|---|
| 4156 | n/a | if (u == NULL) |
|---|
| 4157 | n/a | return NULL; |
|---|
| 4158 | n/a | p = buf = PyBytes_AsString(u); |
|---|
| 4159 | n/a | end = s + len; |
|---|
| 4160 | n/a | while (s < end) { |
|---|
| 4161 | n/a | if (*s == '\\') { |
|---|
| 4162 | n/a | *p++ = *s++; |
|---|
| 4163 | n/a | if (*s & 0x80) { |
|---|
| 4164 | n/a | strcpy(p, "u005c"); |
|---|
| 4165 | n/a | p += 5; |
|---|
| 4166 | n/a | } |
|---|
| 4167 | n/a | } |
|---|
| 4168 | n/a | if (*s & 0x80) { /* XXX inefficient */ |
|---|
| 4169 | n/a | PyObject *w; |
|---|
| 4170 | n/a | int kind; |
|---|
| 4171 | n/a | void *data; |
|---|
| 4172 | n/a | Py_ssize_t len, i; |
|---|
| 4173 | n/a | w = decode_utf8(c, &s, end); |
|---|
| 4174 | n/a | if (w == NULL) { |
|---|
| 4175 | n/a | Py_DECREF(u); |
|---|
| 4176 | n/a | return NULL; |
|---|
| 4177 | n/a | } |
|---|
| 4178 | n/a | kind = PyUnicode_KIND(w); |
|---|
| 4179 | n/a | data = PyUnicode_DATA(w); |
|---|
| 4180 | n/a | len = PyUnicode_GET_LENGTH(w); |
|---|
| 4181 | n/a | for (i = 0; i < len; i++) { |
|---|
| 4182 | n/a | Py_UCS4 chr = PyUnicode_READ(kind, data, i); |
|---|
| 4183 | n/a | sprintf(p, "\\U%08x", chr); |
|---|
| 4184 | n/a | p += 10; |
|---|
| 4185 | n/a | } |
|---|
| 4186 | n/a | /* Should be impossible to overflow */ |
|---|
| 4187 | n/a | assert(p - buf <= Py_SIZE(u)); |
|---|
| 4188 | n/a | Py_DECREF(w); |
|---|
| 4189 | n/a | } else { |
|---|
| 4190 | n/a | *p++ = *s++; |
|---|
| 4191 | n/a | } |
|---|
| 4192 | n/a | } |
|---|
| 4193 | n/a | len = p - buf; |
|---|
| 4194 | n/a | s = buf; |
|---|
| 4195 | n/a | |
|---|
| 4196 | n/a | const char *first_invalid_escape; |
|---|
| 4197 | n/a | v = _PyUnicode_DecodeUnicodeEscape(s, len, NULL, &first_invalid_escape); |
|---|
| 4198 | n/a | |
|---|
| 4199 | n/a | if (v != NULL && first_invalid_escape != NULL) { |
|---|
| 4200 | n/a | if (warn_invalid_escape_sequence(c, n, *first_invalid_escape) < 0) { |
|---|
| 4201 | n/a | /* We have not decref u before because first_invalid_escape points |
|---|
| 4202 | n/a | inside u. */ |
|---|
| 4203 | n/a | Py_XDECREF(u); |
|---|
| 4204 | n/a | Py_DECREF(v); |
|---|
| 4205 | n/a | return NULL; |
|---|
| 4206 | n/a | } |
|---|
| 4207 | n/a | } |
|---|
| 4208 | n/a | Py_XDECREF(u); |
|---|
| 4209 | n/a | return v; |
|---|
| 4210 | n/a | } |
|---|
| 4211 | n/a | |
|---|
| 4212 | n/a | static PyObject * |
|---|
| 4213 | n/a | decode_bytes_with_escapes(struct compiling *c, const node *n, const char *s, |
|---|
| 4214 | n/a | size_t len) |
|---|
| 4215 | n/a | { |
|---|
| 4216 | n/a | const char *first_invalid_escape; |
|---|
| 4217 | n/a | PyObject *result = _PyBytes_DecodeEscape(s, len, NULL, 0, NULL, |
|---|
| 4218 | n/a | &first_invalid_escape); |
|---|
| 4219 | n/a | if (result == NULL) |
|---|
| 4220 | n/a | return NULL; |
|---|
| 4221 | n/a | |
|---|
| 4222 | n/a | if (first_invalid_escape != NULL) { |
|---|
| 4223 | n/a | if (warn_invalid_escape_sequence(c, n, *first_invalid_escape) < 0) { |
|---|
| 4224 | n/a | Py_DECREF(result); |
|---|
| 4225 | n/a | return NULL; |
|---|
| 4226 | n/a | } |
|---|
| 4227 | n/a | } |
|---|
| 4228 | n/a | return result; |
|---|
| 4229 | n/a | } |
|---|
| 4230 | n/a | |
|---|
| 4231 | n/a | /* Compile this expression in to an expr_ty. Add parens around the |
|---|
| 4232 | n/a | expression, in order to allow leading spaces in the expression. */ |
|---|
| 4233 | n/a | static expr_ty |
|---|
| 4234 | n/a | fstring_compile_expr(const char *expr_start, const char *expr_end, |
|---|
| 4235 | n/a | struct compiling *c, const node *n) |
|---|
| 4236 | n/a | |
|---|
| 4237 | n/a | { |
|---|
| 4238 | n/a | int all_whitespace = 1; |
|---|
| 4239 | n/a | int kind; |
|---|
| 4240 | n/a | void *data; |
|---|
| 4241 | n/a | PyCompilerFlags cf; |
|---|
| 4242 | n/a | mod_ty mod; |
|---|
| 4243 | n/a | char *str; |
|---|
| 4244 | n/a | PyObject *o; |
|---|
| 4245 | n/a | Py_ssize_t len; |
|---|
| 4246 | n/a | Py_ssize_t i; |
|---|
| 4247 | n/a | |
|---|
| 4248 | n/a | assert(expr_end >= expr_start); |
|---|
| 4249 | n/a | assert(*(expr_start-1) == '{'); |
|---|
| 4250 | n/a | assert(*expr_end == '}' || *expr_end == '!' || *expr_end == ':'); |
|---|
| 4251 | n/a | |
|---|
| 4252 | n/a | /* We know there are no escapes here, because backslashes are not allowed, |
|---|
| 4253 | n/a | and we know it's utf-8 encoded (per PEP 263). But, in order to check |
|---|
| 4254 | n/a | that each char is not whitespace, we need to decode it to unicode. |
|---|
| 4255 | n/a | Which is unfortunate, but such is life. */ |
|---|
| 4256 | n/a | |
|---|
| 4257 | n/a | /* If the substring is all whitespace, it's an error. We need to catch |
|---|
| 4258 | n/a | this here, and not when we call PyParser_ASTFromString, because turning |
|---|
| 4259 | n/a | the expression '' in to '()' would go from being invalid to valid. */ |
|---|
| 4260 | n/a | /* Note that this code says an empty string is all whitespace. That's |
|---|
| 4261 | n/a | important. There's a test for it: f'{}'. */ |
|---|
| 4262 | n/a | o = PyUnicode_DecodeUTF8(expr_start, expr_end-expr_start, NULL); |
|---|
| 4263 | n/a | if (o == NULL) |
|---|
| 4264 | n/a | return NULL; |
|---|
| 4265 | n/a | len = PyUnicode_GET_LENGTH(o); |
|---|
| 4266 | n/a | kind = PyUnicode_KIND(o); |
|---|
| 4267 | n/a | data = PyUnicode_DATA(o); |
|---|
| 4268 | n/a | for (i = 0; i < len; i++) { |
|---|
| 4269 | n/a | if (!Py_UNICODE_ISSPACE(PyUnicode_READ(kind, data, i))) { |
|---|
| 4270 | n/a | all_whitespace = 0; |
|---|
| 4271 | n/a | break; |
|---|
| 4272 | n/a | } |
|---|
| 4273 | n/a | } |
|---|
| 4274 | n/a | Py_DECREF(o); |
|---|
| 4275 | n/a | if (all_whitespace) { |
|---|
| 4276 | n/a | ast_error(c, n, "f-string: empty expression not allowed"); |
|---|
| 4277 | n/a | return NULL; |
|---|
| 4278 | n/a | } |
|---|
| 4279 | n/a | |
|---|
| 4280 | n/a | /* Reuse len to be the length of the utf-8 input string. */ |
|---|
| 4281 | n/a | len = expr_end - expr_start; |
|---|
| 4282 | n/a | /* Allocate 3 extra bytes: open paren, close paren, null byte. */ |
|---|
| 4283 | n/a | str = PyMem_RawMalloc(len + 3); |
|---|
| 4284 | n/a | if (str == NULL) |
|---|
| 4285 | n/a | return NULL; |
|---|
| 4286 | n/a | |
|---|
| 4287 | n/a | str[0] = '('; |
|---|
| 4288 | n/a | memcpy(str+1, expr_start, len); |
|---|
| 4289 | n/a | str[len+1] = ')'; |
|---|
| 4290 | n/a | str[len+2] = 0; |
|---|
| 4291 | n/a | |
|---|
| 4292 | n/a | cf.cf_flags = PyCF_ONLY_AST; |
|---|
| 4293 | n/a | mod = PyParser_ASTFromString(str, "<fstring>", |
|---|
| 4294 | n/a | Py_eval_input, &cf, c->c_arena); |
|---|
| 4295 | n/a | PyMem_RawFree(str); |
|---|
| 4296 | n/a | if (!mod) |
|---|
| 4297 | n/a | return NULL; |
|---|
| 4298 | n/a | return mod->v.Expression.body; |
|---|
| 4299 | n/a | } |
|---|
| 4300 | n/a | |
|---|
| 4301 | n/a | /* Return -1 on error. |
|---|
| 4302 | n/a | |
|---|
| 4303 | n/a | Return 0 if we reached the end of the literal. |
|---|
| 4304 | n/a | |
|---|
| 4305 | n/a | Return 1 if we haven't reached the end of the literal, but we want |
|---|
| 4306 | n/a | the caller to process the literal up to this point. Used for |
|---|
| 4307 | n/a | doubled braces. |
|---|
| 4308 | n/a | */ |
|---|
| 4309 | n/a | static int |
|---|
| 4310 | n/a | fstring_find_literal(const char **str, const char *end, int raw, |
|---|
| 4311 | n/a | PyObject **literal, int recurse_lvl, |
|---|
| 4312 | n/a | struct compiling *c, const node *n) |
|---|
| 4313 | n/a | { |
|---|
| 4314 | n/a | /* Get any literal string. It ends when we hit an un-doubled left |
|---|
| 4315 | n/a | brace (which isn't part of a unicode name escape such as |
|---|
| 4316 | n/a | "\N{EULER CONSTANT}"), or the end of the string. */ |
|---|
| 4317 | n/a | |
|---|
| 4318 | n/a | const char *literal_start = *str; |
|---|
| 4319 | n/a | const char *literal_end; |
|---|
| 4320 | n/a | int in_named_escape = 0; |
|---|
| 4321 | n/a | int result = 0; |
|---|
| 4322 | n/a | |
|---|
| 4323 | n/a | assert(*literal == NULL); |
|---|
| 4324 | n/a | for (; *str < end; (*str)++) { |
|---|
| 4325 | n/a | char ch = **str; |
|---|
| 4326 | n/a | if (!in_named_escape && ch == '{' && (*str)-literal_start >= 2 && |
|---|
| 4327 | n/a | *(*str-2) == '\\' && *(*str-1) == 'N') { |
|---|
| 4328 | n/a | in_named_escape = 1; |
|---|
| 4329 | n/a | } else if (in_named_escape && ch == '}') { |
|---|
| 4330 | n/a | in_named_escape = 0; |
|---|
| 4331 | n/a | } else if (ch == '{' || ch == '}') { |
|---|
| 4332 | n/a | /* Check for doubled braces, but only at the top level. If |
|---|
| 4333 | n/a | we checked at every level, then f'{0:{3}}' would fail |
|---|
| 4334 | n/a | with the two closing braces. */ |
|---|
| 4335 | n/a | if (recurse_lvl == 0) { |
|---|
| 4336 | n/a | if (*str+1 < end && *(*str+1) == ch) { |
|---|
| 4337 | n/a | /* We're going to tell the caller that the literal ends |
|---|
| 4338 | n/a | here, but that they should continue scanning. But also |
|---|
| 4339 | n/a | skip over the second brace when we resume scanning. */ |
|---|
| 4340 | n/a | literal_end = *str+1; |
|---|
| 4341 | n/a | *str += 2; |
|---|
| 4342 | n/a | result = 1; |
|---|
| 4343 | n/a | goto done; |
|---|
| 4344 | n/a | } |
|---|
| 4345 | n/a | |
|---|
| 4346 | n/a | /* Where a single '{' is the start of a new expression, a |
|---|
| 4347 | n/a | single '}' is not allowed. */ |
|---|
| 4348 | n/a | if (ch == '}') { |
|---|
| 4349 | n/a | ast_error(c, n, "f-string: single '}' is not allowed"); |
|---|
| 4350 | n/a | return -1; |
|---|
| 4351 | n/a | } |
|---|
| 4352 | n/a | } |
|---|
| 4353 | n/a | /* We're either at a '{', which means we're starting another |
|---|
| 4354 | n/a | expression; or a '}', which means we're at the end of this |
|---|
| 4355 | n/a | f-string (for a nested format_spec). */ |
|---|
| 4356 | n/a | break; |
|---|
| 4357 | n/a | } |
|---|
| 4358 | n/a | } |
|---|
| 4359 | n/a | literal_end = *str; |
|---|
| 4360 | n/a | assert(*str <= end); |
|---|
| 4361 | n/a | assert(*str == end || **str == '{' || **str == '}'); |
|---|
| 4362 | n/a | done: |
|---|
| 4363 | n/a | if (literal_start != literal_end) { |
|---|
| 4364 | n/a | if (raw) |
|---|
| 4365 | n/a | *literal = PyUnicode_DecodeUTF8Stateful(literal_start, |
|---|
| 4366 | n/a | literal_end-literal_start, |
|---|
| 4367 | n/a | NULL, NULL); |
|---|
| 4368 | n/a | else |
|---|
| 4369 | n/a | *literal = decode_unicode_with_escapes(c, n, literal_start, |
|---|
| 4370 | n/a | literal_end-literal_start); |
|---|
| 4371 | n/a | if (!*literal) |
|---|
| 4372 | n/a | return -1; |
|---|
| 4373 | n/a | } |
|---|
| 4374 | n/a | return result; |
|---|
| 4375 | n/a | } |
|---|
| 4376 | n/a | |
|---|
| 4377 | n/a | /* Forward declaration because parsing is recursive. */ |
|---|
| 4378 | n/a | static expr_ty |
|---|
| 4379 | n/a | fstring_parse(const char **str, const char *end, int raw, int recurse_lvl, |
|---|
| 4380 | n/a | struct compiling *c, const node *n); |
|---|
| 4381 | n/a | |
|---|
| 4382 | n/a | /* Parse the f-string at *str, ending at end. We know *str starts an |
|---|
| 4383 | n/a | expression (so it must be a '{'). Returns the FormattedValue node, |
|---|
| 4384 | n/a | which includes the expression, conversion character, and |
|---|
| 4385 | n/a | format_spec expression. |
|---|
| 4386 | n/a | |
|---|
| 4387 | n/a | Note that I don't do a perfect job here: I don't make sure that a |
|---|
| 4388 | n/a | closing brace doesn't match an opening paren, for example. It |
|---|
| 4389 | n/a | doesn't need to error on all invalid expressions, just correctly |
|---|
| 4390 | n/a | find the end of all valid ones. Any errors inside the expression |
|---|
| 4391 | n/a | will be caught when we parse it later. */ |
|---|
| 4392 | n/a | static int |
|---|
| 4393 | n/a | fstring_find_expr(const char **str, const char *end, int raw, int recurse_lvl, |
|---|
| 4394 | n/a | expr_ty *expression, struct compiling *c, const node *n) |
|---|
| 4395 | n/a | { |
|---|
| 4396 | n/a | /* Return -1 on error, else 0. */ |
|---|
| 4397 | n/a | |
|---|
| 4398 | n/a | const char *expr_start; |
|---|
| 4399 | n/a | const char *expr_end; |
|---|
| 4400 | n/a | expr_ty simple_expression; |
|---|
| 4401 | n/a | expr_ty format_spec = NULL; /* Optional format specifier. */ |
|---|
| 4402 | n/a | int conversion = -1; /* The conversion char. -1 if not specified. */ |
|---|
| 4403 | n/a | |
|---|
| 4404 | n/a | /* 0 if we're not in a string, else the quote char we're trying to |
|---|
| 4405 | n/a | match (single or double quote). */ |
|---|
| 4406 | n/a | char quote_char = 0; |
|---|
| 4407 | n/a | |
|---|
| 4408 | n/a | /* If we're inside a string, 1=normal, 3=triple-quoted. */ |
|---|
| 4409 | n/a | int string_type = 0; |
|---|
| 4410 | n/a | |
|---|
| 4411 | n/a | /* Keep track of nesting level for braces/parens/brackets in |
|---|
| 4412 | n/a | expressions. */ |
|---|
| 4413 | n/a | Py_ssize_t nested_depth = 0; |
|---|
| 4414 | n/a | |
|---|
| 4415 | n/a | /* Can only nest one level deep. */ |
|---|
| 4416 | n/a | if (recurse_lvl >= 2) { |
|---|
| 4417 | n/a | ast_error(c, n, "f-string: expressions nested too deeply"); |
|---|
| 4418 | n/a | return -1; |
|---|
| 4419 | n/a | } |
|---|
| 4420 | n/a | |
|---|
| 4421 | n/a | /* The first char must be a left brace, or we wouldn't have gotten |
|---|
| 4422 | n/a | here. Skip over it. */ |
|---|
| 4423 | n/a | assert(**str == '{'); |
|---|
| 4424 | n/a | *str += 1; |
|---|
| 4425 | n/a | |
|---|
| 4426 | n/a | expr_start = *str; |
|---|
| 4427 | n/a | for (; *str < end; (*str)++) { |
|---|
| 4428 | n/a | char ch; |
|---|
| 4429 | n/a | |
|---|
| 4430 | n/a | /* Loop invariants. */ |
|---|
| 4431 | n/a | assert(nested_depth >= 0); |
|---|
| 4432 | n/a | assert(*str >= expr_start && *str < end); |
|---|
| 4433 | n/a | if (quote_char) |
|---|
| 4434 | n/a | assert(string_type == 1 || string_type == 3); |
|---|
| 4435 | n/a | else |
|---|
| 4436 | n/a | assert(string_type == 0); |
|---|
| 4437 | n/a | |
|---|
| 4438 | n/a | ch = **str; |
|---|
| 4439 | n/a | /* Nowhere inside an expression is a backslash allowed. */ |
|---|
| 4440 | n/a | if (ch == '\\') { |
|---|
| 4441 | n/a | /* Error: can't include a backslash character, inside |
|---|
| 4442 | n/a | parens or strings or not. */ |
|---|
| 4443 | n/a | ast_error(c, n, "f-string expression part " |
|---|
| 4444 | n/a | "cannot include a backslash"); |
|---|
| 4445 | n/a | return -1; |
|---|
| 4446 | n/a | } |
|---|
| 4447 | n/a | if (quote_char) { |
|---|
| 4448 | n/a | /* We're inside a string. See if we're at the end. */ |
|---|
| 4449 | n/a | /* This code needs to implement the same non-error logic |
|---|
| 4450 | n/a | as tok_get from tokenizer.c, at the letter_quote |
|---|
| 4451 | n/a | label. To actually share that code would be a |
|---|
| 4452 | n/a | nightmare. But, it's unlikely to change and is small, |
|---|
| 4453 | n/a | so duplicate it here. Note we don't need to catch all |
|---|
| 4454 | n/a | of the errors, since they'll be caught when parsing the |
|---|
| 4455 | n/a | expression. We just need to match the non-error |
|---|
| 4456 | n/a | cases. Thus we can ignore \n in single-quoted strings, |
|---|
| 4457 | n/a | for example. Or non-terminated strings. */ |
|---|
| 4458 | n/a | if (ch == quote_char) { |
|---|
| 4459 | n/a | /* Does this match the string_type (single or triple |
|---|
| 4460 | n/a | quoted)? */ |
|---|
| 4461 | n/a | if (string_type == 3) { |
|---|
| 4462 | n/a | if (*str+2 < end && *(*str+1) == ch && *(*str+2) == ch) { |
|---|
| 4463 | n/a | /* We're at the end of a triple quoted string. */ |
|---|
| 4464 | n/a | *str += 2; |
|---|
| 4465 | n/a | string_type = 0; |
|---|
| 4466 | n/a | quote_char = 0; |
|---|
| 4467 | n/a | continue; |
|---|
| 4468 | n/a | } |
|---|
| 4469 | n/a | } else { |
|---|
| 4470 | n/a | /* We're at the end of a normal string. */ |
|---|
| 4471 | n/a | quote_char = 0; |
|---|
| 4472 | n/a | string_type = 0; |
|---|
| 4473 | n/a | continue; |
|---|
| 4474 | n/a | } |
|---|
| 4475 | n/a | } |
|---|
| 4476 | n/a | } else if (ch == '\'' || ch == '"') { |
|---|
| 4477 | n/a | /* Is this a triple quoted string? */ |
|---|
| 4478 | n/a | if (*str+2 < end && *(*str+1) == ch && *(*str+2) == ch) { |
|---|
| 4479 | n/a | string_type = 3; |
|---|
| 4480 | n/a | *str += 2; |
|---|
| 4481 | n/a | } else { |
|---|
| 4482 | n/a | /* Start of a normal string. */ |
|---|
| 4483 | n/a | string_type = 1; |
|---|
| 4484 | n/a | } |
|---|
| 4485 | n/a | /* Start looking for the end of the string. */ |
|---|
| 4486 | n/a | quote_char = ch; |
|---|
| 4487 | n/a | } else if (ch == '[' || ch == '{' || ch == '(') { |
|---|
| 4488 | n/a | nested_depth++; |
|---|
| 4489 | n/a | } else if (nested_depth != 0 && |
|---|
| 4490 | n/a | (ch == ']' || ch == '}' || ch == ')')) { |
|---|
| 4491 | n/a | nested_depth--; |
|---|
| 4492 | n/a | } else if (ch == '#') { |
|---|
| 4493 | n/a | /* Error: can't include a comment character, inside parens |
|---|
| 4494 | n/a | or not. */ |
|---|
| 4495 | n/a | ast_error(c, n, "f-string expression part cannot include '#'"); |
|---|
| 4496 | n/a | return -1; |
|---|
| 4497 | n/a | } else if (nested_depth == 0 && |
|---|
| 4498 | n/a | (ch == '!' || ch == ':' || ch == '}')) { |
|---|
| 4499 | n/a | /* First, test for the special case of "!=". Since '=' is |
|---|
| 4500 | n/a | not an allowed conversion character, nothing is lost in |
|---|
| 4501 | n/a | this test. */ |
|---|
| 4502 | n/a | if (ch == '!' && *str+1 < end && *(*str+1) == '=') { |
|---|
| 4503 | n/a | /* This isn't a conversion character, just continue. */ |
|---|
| 4504 | n/a | continue; |
|---|
| 4505 | n/a | } |
|---|
| 4506 | n/a | /* Normal way out of this loop. */ |
|---|
| 4507 | n/a | break; |
|---|
| 4508 | n/a | } else { |
|---|
| 4509 | n/a | /* Just consume this char and loop around. */ |
|---|
| 4510 | n/a | } |
|---|
| 4511 | n/a | } |
|---|
| 4512 | n/a | expr_end = *str; |
|---|
| 4513 | n/a | /* If we leave this loop in a string or with mismatched parens, we |
|---|
| 4514 | n/a | don't care. We'll get a syntax error when compiling the |
|---|
| 4515 | n/a | expression. But, we can produce a better error message, so |
|---|
| 4516 | n/a | let's just do that.*/ |
|---|
| 4517 | n/a | if (quote_char) { |
|---|
| 4518 | n/a | ast_error(c, n, "f-string: unterminated string"); |
|---|
| 4519 | n/a | return -1; |
|---|
| 4520 | n/a | } |
|---|
| 4521 | n/a | if (nested_depth) { |
|---|
| 4522 | n/a | ast_error(c, n, "f-string: mismatched '(', '{', or '['"); |
|---|
| 4523 | n/a | return -1; |
|---|
| 4524 | n/a | } |
|---|
| 4525 | n/a | |
|---|
| 4526 | n/a | if (*str >= end) |
|---|
| 4527 | n/a | goto unexpected_end_of_string; |
|---|
| 4528 | n/a | |
|---|
| 4529 | n/a | /* Compile the expression as soon as possible, so we show errors |
|---|
| 4530 | n/a | related to the expression before errors related to the |
|---|
| 4531 | n/a | conversion or format_spec. */ |
|---|
| 4532 | n/a | simple_expression = fstring_compile_expr(expr_start, expr_end, c, n); |
|---|
| 4533 | n/a | if (!simple_expression) |
|---|
| 4534 | n/a | return -1; |
|---|
| 4535 | n/a | |
|---|
| 4536 | n/a | /* Check for a conversion char, if present. */ |
|---|
| 4537 | n/a | if (**str == '!') { |
|---|
| 4538 | n/a | *str += 1; |
|---|
| 4539 | n/a | if (*str >= end) |
|---|
| 4540 | n/a | goto unexpected_end_of_string; |
|---|
| 4541 | n/a | |
|---|
| 4542 | n/a | conversion = **str; |
|---|
| 4543 | n/a | *str += 1; |
|---|
| 4544 | n/a | |
|---|
| 4545 | n/a | /* Validate the conversion. */ |
|---|
| 4546 | n/a | if (!(conversion == 's' || conversion == 'r' |
|---|
| 4547 | n/a | || conversion == 'a')) { |
|---|
| 4548 | n/a | ast_error(c, n, "f-string: invalid conversion character: " |
|---|
| 4549 | n/a | "expected 's', 'r', or 'a'"); |
|---|
| 4550 | n/a | return -1; |
|---|
| 4551 | n/a | } |
|---|
| 4552 | n/a | } |
|---|
| 4553 | n/a | |
|---|
| 4554 | n/a | /* Check for the format spec, if present. */ |
|---|
| 4555 | n/a | if (*str >= end) |
|---|
| 4556 | n/a | goto unexpected_end_of_string; |
|---|
| 4557 | n/a | if (**str == ':') { |
|---|
| 4558 | n/a | *str += 1; |
|---|
| 4559 | n/a | if (*str >= end) |
|---|
| 4560 | n/a | goto unexpected_end_of_string; |
|---|
| 4561 | n/a | |
|---|
| 4562 | n/a | /* Parse the format spec. */ |
|---|
| 4563 | n/a | format_spec = fstring_parse(str, end, raw, recurse_lvl+1, c, n); |
|---|
| 4564 | n/a | if (!format_spec) |
|---|
| 4565 | n/a | return -1; |
|---|
| 4566 | n/a | } |
|---|
| 4567 | n/a | |
|---|
| 4568 | n/a | if (*str >= end || **str != '}') |
|---|
| 4569 | n/a | goto unexpected_end_of_string; |
|---|
| 4570 | n/a | |
|---|
| 4571 | n/a | /* We're at a right brace. Consume it. */ |
|---|
| 4572 | n/a | assert(*str < end); |
|---|
| 4573 | n/a | assert(**str == '}'); |
|---|
| 4574 | n/a | *str += 1; |
|---|
| 4575 | n/a | |
|---|
| 4576 | n/a | /* And now create the FormattedValue node that represents this |
|---|
| 4577 | n/a | entire expression with the conversion and format spec. */ |
|---|
| 4578 | n/a | *expression = FormattedValue(simple_expression, conversion, |
|---|
| 4579 | n/a | format_spec, LINENO(n), n->n_col_offset, |
|---|
| 4580 | n/a | c->c_arena); |
|---|
| 4581 | n/a | if (!*expression) |
|---|
| 4582 | n/a | return -1; |
|---|
| 4583 | n/a | |
|---|
| 4584 | n/a | return 0; |
|---|
| 4585 | n/a | |
|---|
| 4586 | n/a | unexpected_end_of_string: |
|---|
| 4587 | n/a | ast_error(c, n, "f-string: expecting '}'"); |
|---|
| 4588 | n/a | return -1; |
|---|
| 4589 | n/a | } |
|---|
| 4590 | n/a | |
|---|
| 4591 | n/a | /* Return -1 on error. |
|---|
| 4592 | n/a | |
|---|
| 4593 | n/a | Return 0 if we have a literal (possible zero length) and an |
|---|
| 4594 | n/a | expression (zero length if at the end of the string. |
|---|
| 4595 | n/a | |
|---|
| 4596 | n/a | Return 1 if we have a literal, but no expression, and we want the |
|---|
| 4597 | n/a | caller to call us again. This is used to deal with doubled |
|---|
| 4598 | n/a | braces. |
|---|
| 4599 | n/a | |
|---|
| 4600 | n/a | When called multiple times on the string 'a{{b{0}c', this function |
|---|
| 4601 | n/a | will return: |
|---|
| 4602 | n/a | |
|---|
| 4603 | n/a | 1. the literal 'a{' with no expression, and a return value |
|---|
| 4604 | n/a | of 1. Despite the fact that there's no expression, the return |
|---|
| 4605 | n/a | value of 1 means we're not finished yet. |
|---|
| 4606 | n/a | |
|---|
| 4607 | n/a | 2. the literal 'b' and the expression '0', with a return value of |
|---|
| 4608 | n/a | 0. The fact that there's an expression means we're not finished. |
|---|
| 4609 | n/a | |
|---|
| 4610 | n/a | 3. literal 'c' with no expression and a return value of 0. The |
|---|
| 4611 | n/a | combination of the return value of 0 with no expression means |
|---|
| 4612 | n/a | we're finished. |
|---|
| 4613 | n/a | */ |
|---|
| 4614 | n/a | static int |
|---|
| 4615 | n/a | fstring_find_literal_and_expr(const char **str, const char *end, int raw, |
|---|
| 4616 | n/a | int recurse_lvl, PyObject **literal, |
|---|
| 4617 | n/a | expr_ty *expression, |
|---|
| 4618 | n/a | struct compiling *c, const node *n) |
|---|
| 4619 | n/a | { |
|---|
| 4620 | n/a | int result; |
|---|
| 4621 | n/a | |
|---|
| 4622 | n/a | assert(*literal == NULL && *expression == NULL); |
|---|
| 4623 | n/a | |
|---|
| 4624 | n/a | /* Get any literal string. */ |
|---|
| 4625 | n/a | result = fstring_find_literal(str, end, raw, literal, recurse_lvl, c, n); |
|---|
| 4626 | n/a | if (result < 0) |
|---|
| 4627 | n/a | goto error; |
|---|
| 4628 | n/a | |
|---|
| 4629 | n/a | assert(result == 0 || result == 1); |
|---|
| 4630 | n/a | |
|---|
| 4631 | n/a | if (result == 1) |
|---|
| 4632 | n/a | /* We have a literal, but don't look at the expression. */ |
|---|
| 4633 | n/a | return 1; |
|---|
| 4634 | n/a | |
|---|
| 4635 | n/a | if (*str >= end || **str == '}') |
|---|
| 4636 | n/a | /* We're at the end of the string or the end of a nested |
|---|
| 4637 | n/a | f-string: no expression. The top-level error case where we |
|---|
| 4638 | n/a | expect to be at the end of the string but we're at a '}' is |
|---|
| 4639 | n/a | handled later. */ |
|---|
| 4640 | n/a | return 0; |
|---|
| 4641 | n/a | |
|---|
| 4642 | n/a | /* We must now be the start of an expression, on a '{'. */ |
|---|
| 4643 | n/a | assert(**str == '{'); |
|---|
| 4644 | n/a | |
|---|
| 4645 | n/a | if (fstring_find_expr(str, end, raw, recurse_lvl, expression, c, n) < 0) |
|---|
| 4646 | n/a | goto error; |
|---|
| 4647 | n/a | |
|---|
| 4648 | n/a | return 0; |
|---|
| 4649 | n/a | |
|---|
| 4650 | n/a | error: |
|---|
| 4651 | n/a | Py_CLEAR(*literal); |
|---|
| 4652 | n/a | return -1; |
|---|
| 4653 | n/a | } |
|---|
| 4654 | n/a | |
|---|
| 4655 | n/a | #define EXPRLIST_N_CACHED 64 |
|---|
| 4656 | n/a | |
|---|
| 4657 | n/a | typedef struct { |
|---|
| 4658 | n/a | /* Incrementally build an array of expr_ty, so be used in an |
|---|
| 4659 | n/a | asdl_seq. Cache some small but reasonably sized number of |
|---|
| 4660 | n/a | expr_ty's, and then after that start dynamically allocating, |
|---|
| 4661 | n/a | doubling the number allocated each time. Note that the f-string |
|---|
| 4662 | n/a | f'{0}a{1}' contains 3 expr_ty's: 2 FormattedValue's, and one |
|---|
| 4663 | n/a | Str for the literal 'a'. So you add expr_ty's about twice as |
|---|
| 4664 | n/a | fast as you add exressions in an f-string. */ |
|---|
| 4665 | n/a | |
|---|
| 4666 | n/a | Py_ssize_t allocated; /* Number we've allocated. */ |
|---|
| 4667 | n/a | Py_ssize_t size; /* Number we've used. */ |
|---|
| 4668 | n/a | expr_ty *p; /* Pointer to the memory we're actually |
|---|
| 4669 | n/a | using. Will point to 'data' until we |
|---|
| 4670 | n/a | start dynamically allocating. */ |
|---|
| 4671 | n/a | expr_ty data[EXPRLIST_N_CACHED]; |
|---|
| 4672 | n/a | } ExprList; |
|---|
| 4673 | n/a | |
|---|
| 4674 | n/a | #ifdef NDEBUG |
|---|
| 4675 | n/a | #define ExprList_check_invariants(l) |
|---|
| 4676 | n/a | #else |
|---|
| 4677 | n/a | static void |
|---|
| 4678 | n/a | ExprList_check_invariants(ExprList *l) |
|---|
| 4679 | n/a | { |
|---|
| 4680 | n/a | /* Check our invariants. Make sure this object is "live", and |
|---|
| 4681 | n/a | hasn't been deallocated. */ |
|---|
| 4682 | n/a | assert(l->size >= 0); |
|---|
| 4683 | n/a | assert(l->p != NULL); |
|---|
| 4684 | n/a | if (l->size <= EXPRLIST_N_CACHED) |
|---|
| 4685 | n/a | assert(l->data == l->p); |
|---|
| 4686 | n/a | } |
|---|
| 4687 | n/a | #endif |
|---|
| 4688 | n/a | |
|---|
| 4689 | n/a | static void |
|---|
| 4690 | n/a | ExprList_Init(ExprList *l) |
|---|
| 4691 | n/a | { |
|---|
| 4692 | n/a | l->allocated = EXPRLIST_N_CACHED; |
|---|
| 4693 | n/a | l->size = 0; |
|---|
| 4694 | n/a | |
|---|
| 4695 | n/a | /* Until we start allocating dynamically, p points to data. */ |
|---|
| 4696 | n/a | l->p = l->data; |
|---|
| 4697 | n/a | |
|---|
| 4698 | n/a | ExprList_check_invariants(l); |
|---|
| 4699 | n/a | } |
|---|
| 4700 | n/a | |
|---|
| 4701 | n/a | static int |
|---|
| 4702 | n/a | ExprList_Append(ExprList *l, expr_ty exp) |
|---|
| 4703 | n/a | { |
|---|
| 4704 | n/a | ExprList_check_invariants(l); |
|---|
| 4705 | n/a | if (l->size >= l->allocated) { |
|---|
| 4706 | n/a | /* We need to alloc (or realloc) the memory. */ |
|---|
| 4707 | n/a | Py_ssize_t new_size = l->allocated * 2; |
|---|
| 4708 | n/a | |
|---|
| 4709 | n/a | /* See if we've ever allocated anything dynamically. */ |
|---|
| 4710 | n/a | if (l->p == l->data) { |
|---|
| 4711 | n/a | Py_ssize_t i; |
|---|
| 4712 | n/a | /* We're still using the cached data. Switch to |
|---|
| 4713 | n/a | alloc-ing. */ |
|---|
| 4714 | n/a | l->p = PyMem_RawMalloc(sizeof(expr_ty) * new_size); |
|---|
| 4715 | n/a | if (!l->p) |
|---|
| 4716 | n/a | return -1; |
|---|
| 4717 | n/a | /* Copy the cached data into the new buffer. */ |
|---|
| 4718 | n/a | for (i = 0; i < l->size; i++) |
|---|
| 4719 | n/a | l->p[i] = l->data[i]; |
|---|
| 4720 | n/a | } else { |
|---|
| 4721 | n/a | /* Just realloc. */ |
|---|
| 4722 | n/a | expr_ty *tmp = PyMem_RawRealloc(l->p, sizeof(expr_ty) * new_size); |
|---|
| 4723 | n/a | if (!tmp) { |
|---|
| 4724 | n/a | PyMem_RawFree(l->p); |
|---|
| 4725 | n/a | l->p = NULL; |
|---|
| 4726 | n/a | return -1; |
|---|
| 4727 | n/a | } |
|---|
| 4728 | n/a | l->p = tmp; |
|---|
| 4729 | n/a | } |
|---|
| 4730 | n/a | |
|---|
| 4731 | n/a | l->allocated = new_size; |
|---|
| 4732 | n/a | assert(l->allocated == 2 * l->size); |
|---|
| 4733 | n/a | } |
|---|
| 4734 | n/a | |
|---|
| 4735 | n/a | l->p[l->size++] = exp; |
|---|
| 4736 | n/a | |
|---|
| 4737 | n/a | ExprList_check_invariants(l); |
|---|
| 4738 | n/a | return 0; |
|---|
| 4739 | n/a | } |
|---|
| 4740 | n/a | |
|---|
| 4741 | n/a | static void |
|---|
| 4742 | n/a | ExprList_Dealloc(ExprList *l) |
|---|
| 4743 | n/a | { |
|---|
| 4744 | n/a | ExprList_check_invariants(l); |
|---|
| 4745 | n/a | |
|---|
| 4746 | n/a | /* If there's been an error, or we've never dynamically allocated, |
|---|
| 4747 | n/a | do nothing. */ |
|---|
| 4748 | n/a | if (!l->p || l->p == l->data) { |
|---|
| 4749 | n/a | /* Do nothing. */ |
|---|
| 4750 | n/a | } else { |
|---|
| 4751 | n/a | /* We have dynamically allocated. Free the memory. */ |
|---|
| 4752 | n/a | PyMem_RawFree(l->p); |
|---|
| 4753 | n/a | } |
|---|
| 4754 | n/a | l->p = NULL; |
|---|
| 4755 | n/a | l->size = -1; |
|---|
| 4756 | n/a | } |
|---|
| 4757 | n/a | |
|---|
| 4758 | n/a | static asdl_seq * |
|---|
| 4759 | n/a | ExprList_Finish(ExprList *l, PyArena *arena) |
|---|
| 4760 | n/a | { |
|---|
| 4761 | n/a | asdl_seq *seq; |
|---|
| 4762 | n/a | |
|---|
| 4763 | n/a | ExprList_check_invariants(l); |
|---|
| 4764 | n/a | |
|---|
| 4765 | n/a | /* Allocate the asdl_seq and copy the expressions in to it. */ |
|---|
| 4766 | n/a | seq = _Py_asdl_seq_new(l->size, arena); |
|---|
| 4767 | n/a | if (seq) { |
|---|
| 4768 | n/a | Py_ssize_t i; |
|---|
| 4769 | n/a | for (i = 0; i < l->size; i++) |
|---|
| 4770 | n/a | asdl_seq_SET(seq, i, l->p[i]); |
|---|
| 4771 | n/a | } |
|---|
| 4772 | n/a | ExprList_Dealloc(l); |
|---|
| 4773 | n/a | return seq; |
|---|
| 4774 | n/a | } |
|---|
| 4775 | n/a | |
|---|
| 4776 | n/a | /* The FstringParser is designed to add a mix of strings and |
|---|
| 4777 | n/a | f-strings, and concat them together as needed. Ultimately, it |
|---|
| 4778 | n/a | generates an expr_ty. */ |
|---|
| 4779 | n/a | typedef struct { |
|---|
| 4780 | n/a | PyObject *last_str; |
|---|
| 4781 | n/a | ExprList expr_list; |
|---|
| 4782 | n/a | int fmode; |
|---|
| 4783 | n/a | } FstringParser; |
|---|
| 4784 | n/a | |
|---|
| 4785 | n/a | #ifdef NDEBUG |
|---|
| 4786 | n/a | #define FstringParser_check_invariants(state) |
|---|
| 4787 | n/a | #else |
|---|
| 4788 | n/a | static void |
|---|
| 4789 | n/a | FstringParser_check_invariants(FstringParser *state) |
|---|
| 4790 | n/a | { |
|---|
| 4791 | n/a | if (state->last_str) |
|---|
| 4792 | n/a | assert(PyUnicode_CheckExact(state->last_str)); |
|---|
| 4793 | n/a | ExprList_check_invariants(&state->expr_list); |
|---|
| 4794 | n/a | } |
|---|
| 4795 | n/a | #endif |
|---|
| 4796 | n/a | |
|---|
| 4797 | n/a | static void |
|---|
| 4798 | n/a | FstringParser_Init(FstringParser *state) |
|---|
| 4799 | n/a | { |
|---|
| 4800 | n/a | state->last_str = NULL; |
|---|
| 4801 | n/a | state->fmode = 0; |
|---|
| 4802 | n/a | ExprList_Init(&state->expr_list); |
|---|
| 4803 | n/a | FstringParser_check_invariants(state); |
|---|
| 4804 | n/a | } |
|---|
| 4805 | n/a | |
|---|
| 4806 | n/a | static void |
|---|
| 4807 | n/a | FstringParser_Dealloc(FstringParser *state) |
|---|
| 4808 | n/a | { |
|---|
| 4809 | n/a | FstringParser_check_invariants(state); |
|---|
| 4810 | n/a | |
|---|
| 4811 | n/a | Py_XDECREF(state->last_str); |
|---|
| 4812 | n/a | ExprList_Dealloc(&state->expr_list); |
|---|
| 4813 | n/a | } |
|---|
| 4814 | n/a | |
|---|
| 4815 | n/a | /* Make a Str node, but decref the PyUnicode object being added. */ |
|---|
| 4816 | n/a | static expr_ty |
|---|
| 4817 | n/a | make_str_node_and_del(PyObject **str, struct compiling *c, const node* n) |
|---|
| 4818 | n/a | { |
|---|
| 4819 | n/a | PyObject *s = *str; |
|---|
| 4820 | n/a | *str = NULL; |
|---|
| 4821 | n/a | assert(PyUnicode_CheckExact(s)); |
|---|
| 4822 | n/a | if (PyArena_AddPyObject(c->c_arena, s) < 0) { |
|---|
| 4823 | n/a | Py_DECREF(s); |
|---|
| 4824 | n/a | return NULL; |
|---|
| 4825 | n/a | } |
|---|
| 4826 | n/a | return Str(s, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 4827 | n/a | } |
|---|
| 4828 | n/a | |
|---|
| 4829 | n/a | /* Add a non-f-string (that is, a regular literal string). str is |
|---|
| 4830 | n/a | decref'd. */ |
|---|
| 4831 | n/a | static int |
|---|
| 4832 | n/a | FstringParser_ConcatAndDel(FstringParser *state, PyObject *str) |
|---|
| 4833 | n/a | { |
|---|
| 4834 | n/a | FstringParser_check_invariants(state); |
|---|
| 4835 | n/a | |
|---|
| 4836 | n/a | assert(PyUnicode_CheckExact(str)); |
|---|
| 4837 | n/a | |
|---|
| 4838 | n/a | if (PyUnicode_GET_LENGTH(str) == 0) { |
|---|
| 4839 | n/a | Py_DECREF(str); |
|---|
| 4840 | n/a | return 0; |
|---|
| 4841 | n/a | } |
|---|
| 4842 | n/a | |
|---|
| 4843 | n/a | if (!state->last_str) { |
|---|
| 4844 | n/a | /* We didn't have a string before, so just remember this one. */ |
|---|
| 4845 | n/a | state->last_str = str; |
|---|
| 4846 | n/a | } else { |
|---|
| 4847 | n/a | /* Concatenate this with the previous string. */ |
|---|
| 4848 | n/a | PyUnicode_AppendAndDel(&state->last_str, str); |
|---|
| 4849 | n/a | if (!state->last_str) |
|---|
| 4850 | n/a | return -1; |
|---|
| 4851 | n/a | } |
|---|
| 4852 | n/a | FstringParser_check_invariants(state); |
|---|
| 4853 | n/a | return 0; |
|---|
| 4854 | n/a | } |
|---|
| 4855 | n/a | |
|---|
| 4856 | n/a | /* Parse an f-string. The f-string is in *str to end, with no |
|---|
| 4857 | n/a | 'f' or quotes. */ |
|---|
| 4858 | n/a | static int |
|---|
| 4859 | n/a | FstringParser_ConcatFstring(FstringParser *state, const char **str, |
|---|
| 4860 | n/a | const char *end, int raw, int recurse_lvl, |
|---|
| 4861 | n/a | struct compiling *c, const node *n) |
|---|
| 4862 | n/a | { |
|---|
| 4863 | n/a | FstringParser_check_invariants(state); |
|---|
| 4864 | n/a | state->fmode = 1; |
|---|
| 4865 | n/a | |
|---|
| 4866 | n/a | /* Parse the f-string. */ |
|---|
| 4867 | n/a | while (1) { |
|---|
| 4868 | n/a | PyObject *literal = NULL; |
|---|
| 4869 | n/a | expr_ty expression = NULL; |
|---|
| 4870 | n/a | |
|---|
| 4871 | n/a | /* If there's a zero length literal in front of the |
|---|
| 4872 | n/a | expression, literal will be NULL. If we're at the end of |
|---|
| 4873 | n/a | the f-string, expression will be NULL (unless result == 1, |
|---|
| 4874 | n/a | see below). */ |
|---|
| 4875 | n/a | int result = fstring_find_literal_and_expr(str, end, raw, recurse_lvl, |
|---|
| 4876 | n/a | &literal, &expression, |
|---|
| 4877 | n/a | c, n); |
|---|
| 4878 | n/a | if (result < 0) |
|---|
| 4879 | n/a | return -1; |
|---|
| 4880 | n/a | |
|---|
| 4881 | n/a | /* Add the literal, if any. */ |
|---|
| 4882 | n/a | if (!literal) { |
|---|
| 4883 | n/a | /* Do nothing. Just leave last_str alone (and possibly |
|---|
| 4884 | n/a | NULL). */ |
|---|
| 4885 | n/a | } else if (!state->last_str) { |
|---|
| 4886 | n/a | state->last_str = literal; |
|---|
| 4887 | n/a | literal = NULL; |
|---|
| 4888 | n/a | } else { |
|---|
| 4889 | n/a | /* We have a literal, concatenate it. */ |
|---|
| 4890 | n/a | assert(PyUnicode_GET_LENGTH(literal) != 0); |
|---|
| 4891 | n/a | if (FstringParser_ConcatAndDel(state, literal) < 0) |
|---|
| 4892 | n/a | return -1; |
|---|
| 4893 | n/a | literal = NULL; |
|---|
| 4894 | n/a | } |
|---|
| 4895 | n/a | assert(!state->last_str || |
|---|
| 4896 | n/a | PyUnicode_GET_LENGTH(state->last_str) != 0); |
|---|
| 4897 | n/a | |
|---|
| 4898 | n/a | /* We've dealt with the literal now. It can't be leaked on further |
|---|
| 4899 | n/a | errors. */ |
|---|
| 4900 | n/a | assert(literal == NULL); |
|---|
| 4901 | n/a | |
|---|
| 4902 | n/a | /* See if we should just loop around to get the next literal |
|---|
| 4903 | n/a | and expression, while ignoring the expression this |
|---|
| 4904 | n/a | time. This is used for un-doubling braces, as an |
|---|
| 4905 | n/a | optimization. */ |
|---|
| 4906 | n/a | if (result == 1) |
|---|
| 4907 | n/a | continue; |
|---|
| 4908 | n/a | |
|---|
| 4909 | n/a | if (!expression) |
|---|
| 4910 | n/a | /* We're done with this f-string. */ |
|---|
| 4911 | n/a | break; |
|---|
| 4912 | n/a | |
|---|
| 4913 | n/a | /* We know we have an expression. Convert any existing string |
|---|
| 4914 | n/a | to a Str node. */ |
|---|
| 4915 | n/a | if (!state->last_str) { |
|---|
| 4916 | n/a | /* Do nothing. No previous literal. */ |
|---|
| 4917 | n/a | } else { |
|---|
| 4918 | n/a | /* Convert the existing last_str literal to a Str node. */ |
|---|
| 4919 | n/a | expr_ty str = make_str_node_and_del(&state->last_str, c, n); |
|---|
| 4920 | n/a | if (!str || ExprList_Append(&state->expr_list, str) < 0) |
|---|
| 4921 | n/a | return -1; |
|---|
| 4922 | n/a | } |
|---|
| 4923 | n/a | |
|---|
| 4924 | n/a | if (ExprList_Append(&state->expr_list, expression) < 0) |
|---|
| 4925 | n/a | return -1; |
|---|
| 4926 | n/a | } |
|---|
| 4927 | n/a | |
|---|
| 4928 | n/a | /* If recurse_lvl is zero, then we must be at the end of the |
|---|
| 4929 | n/a | string. Otherwise, we must be at a right brace. */ |
|---|
| 4930 | n/a | |
|---|
| 4931 | n/a | if (recurse_lvl == 0 && *str < end-1) { |
|---|
| 4932 | n/a | ast_error(c, n, "f-string: unexpected end of string"); |
|---|
| 4933 | n/a | return -1; |
|---|
| 4934 | n/a | } |
|---|
| 4935 | n/a | if (recurse_lvl != 0 && **str != '}') { |
|---|
| 4936 | n/a | ast_error(c, n, "f-string: expecting '}'"); |
|---|
| 4937 | n/a | return -1; |
|---|
| 4938 | n/a | } |
|---|
| 4939 | n/a | |
|---|
| 4940 | n/a | FstringParser_check_invariants(state); |
|---|
| 4941 | n/a | return 0; |
|---|
| 4942 | n/a | } |
|---|
| 4943 | n/a | |
|---|
| 4944 | n/a | /* Convert the partial state reflected in last_str and expr_list to an |
|---|
| 4945 | n/a | expr_ty. The expr_ty can be a Str, or a JoinedStr. */ |
|---|
| 4946 | n/a | static expr_ty |
|---|
| 4947 | n/a | FstringParser_Finish(FstringParser *state, struct compiling *c, |
|---|
| 4948 | n/a | const node *n) |
|---|
| 4949 | n/a | { |
|---|
| 4950 | n/a | asdl_seq *seq; |
|---|
| 4951 | n/a | |
|---|
| 4952 | n/a | FstringParser_check_invariants(state); |
|---|
| 4953 | n/a | |
|---|
| 4954 | n/a | /* If we're just a constant string with no expressions, return |
|---|
| 4955 | n/a | that. */ |
|---|
| 4956 | n/a | if (!state->fmode) { |
|---|
| 4957 | n/a | assert(!state->expr_list.size); |
|---|
| 4958 | n/a | if (!state->last_str) { |
|---|
| 4959 | n/a | /* Create a zero length string. */ |
|---|
| 4960 | n/a | state->last_str = PyUnicode_FromStringAndSize(NULL, 0); |
|---|
| 4961 | n/a | if (!state->last_str) |
|---|
| 4962 | n/a | goto error; |
|---|
| 4963 | n/a | } |
|---|
| 4964 | n/a | return make_str_node_and_del(&state->last_str, c, n); |
|---|
| 4965 | n/a | } |
|---|
| 4966 | n/a | |
|---|
| 4967 | n/a | /* Create a Str node out of last_str, if needed. It will be the |
|---|
| 4968 | n/a | last node in our expression list. */ |
|---|
| 4969 | n/a | if (state->last_str) { |
|---|
| 4970 | n/a | expr_ty str = make_str_node_and_del(&state->last_str, c, n); |
|---|
| 4971 | n/a | if (!str || ExprList_Append(&state->expr_list, str) < 0) |
|---|
| 4972 | n/a | goto error; |
|---|
| 4973 | n/a | } |
|---|
| 4974 | n/a | /* This has already been freed. */ |
|---|
| 4975 | n/a | assert(state->last_str == NULL); |
|---|
| 4976 | n/a | |
|---|
| 4977 | n/a | seq = ExprList_Finish(&state->expr_list, c->c_arena); |
|---|
| 4978 | n/a | if (!seq) |
|---|
| 4979 | n/a | goto error; |
|---|
| 4980 | n/a | |
|---|
| 4981 | n/a | return JoinedStr(seq, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 4982 | n/a | |
|---|
| 4983 | n/a | error: |
|---|
| 4984 | n/a | FstringParser_Dealloc(state); |
|---|
| 4985 | n/a | return NULL; |
|---|
| 4986 | n/a | } |
|---|
| 4987 | n/a | |
|---|
| 4988 | n/a | /* Given an f-string (with no 'f' or quotes) that's in *str and ends |
|---|
| 4989 | n/a | at end, parse it into an expr_ty. Return NULL on error. Adjust |
|---|
| 4990 | n/a | str to point past the parsed portion. */ |
|---|
| 4991 | n/a | static expr_ty |
|---|
| 4992 | n/a | fstring_parse(const char **str, const char *end, int raw, int recurse_lvl, |
|---|
| 4993 | n/a | struct compiling *c, const node *n) |
|---|
| 4994 | n/a | { |
|---|
| 4995 | n/a | FstringParser state; |
|---|
| 4996 | n/a | |
|---|
| 4997 | n/a | FstringParser_Init(&state); |
|---|
| 4998 | n/a | if (FstringParser_ConcatFstring(&state, str, end, raw, recurse_lvl, |
|---|
| 4999 | n/a | c, n) < 0) { |
|---|
| 5000 | n/a | FstringParser_Dealloc(&state); |
|---|
| 5001 | n/a | return NULL; |
|---|
| 5002 | n/a | } |
|---|
| 5003 | n/a | |
|---|
| 5004 | n/a | return FstringParser_Finish(&state, c, n); |
|---|
| 5005 | n/a | } |
|---|
| 5006 | n/a | |
|---|
| 5007 | n/a | /* n is a Python string literal, including the bracketing quote |
|---|
| 5008 | n/a | characters, and r, b, u, &/or f prefixes (if any), and embedded |
|---|
| 5009 | n/a | escape sequences (if any). parsestr parses it, and sets *result to |
|---|
| 5010 | n/a | decoded Python string object. If the string is an f-string, set |
|---|
| 5011 | n/a | *fstr and *fstrlen to the unparsed string object. Return 0 if no |
|---|
| 5012 | n/a | errors occurred. |
|---|
| 5013 | n/a | */ |
|---|
| 5014 | n/a | static int |
|---|
| 5015 | n/a | parsestr(struct compiling *c, const node *n, int *bytesmode, int *rawmode, |
|---|
| 5016 | n/a | PyObject **result, const char **fstr, Py_ssize_t *fstrlen) |
|---|
| 5017 | n/a | { |
|---|
| 5018 | n/a | size_t len; |
|---|
| 5019 | n/a | const char *s = STR(n); |
|---|
| 5020 | n/a | int quote = Py_CHARMASK(*s); |
|---|
| 5021 | n/a | int fmode = 0; |
|---|
| 5022 | n/a | *bytesmode = 0; |
|---|
| 5023 | n/a | *rawmode = 0; |
|---|
| 5024 | n/a | *result = NULL; |
|---|
| 5025 | n/a | *fstr = NULL; |
|---|
| 5026 | n/a | if (Py_ISALPHA(quote)) { |
|---|
| 5027 | n/a | while (!*bytesmode || !*rawmode) { |
|---|
| 5028 | n/a | if (quote == 'b' || quote == 'B') { |
|---|
| 5029 | n/a | quote = *++s; |
|---|
| 5030 | n/a | *bytesmode = 1; |
|---|
| 5031 | n/a | } |
|---|
| 5032 | n/a | else if (quote == 'u' || quote == 'U') { |
|---|
| 5033 | n/a | quote = *++s; |
|---|
| 5034 | n/a | } |
|---|
| 5035 | n/a | else if (quote == 'r' || quote == 'R') { |
|---|
| 5036 | n/a | quote = *++s; |
|---|
| 5037 | n/a | *rawmode = 1; |
|---|
| 5038 | n/a | } |
|---|
| 5039 | n/a | else if (quote == 'f' || quote == 'F') { |
|---|
| 5040 | n/a | quote = *++s; |
|---|
| 5041 | n/a | fmode = 1; |
|---|
| 5042 | n/a | } |
|---|
| 5043 | n/a | else { |
|---|
| 5044 | n/a | break; |
|---|
| 5045 | n/a | } |
|---|
| 5046 | n/a | } |
|---|
| 5047 | n/a | } |
|---|
| 5048 | n/a | if (fmode && *bytesmode) { |
|---|
| 5049 | n/a | PyErr_BadInternalCall(); |
|---|
| 5050 | n/a | return -1; |
|---|
| 5051 | n/a | } |
|---|
| 5052 | n/a | if (quote != '\'' && quote != '\"') { |
|---|
| 5053 | n/a | PyErr_BadInternalCall(); |
|---|
| 5054 | n/a | return -1; |
|---|
| 5055 | n/a | } |
|---|
| 5056 | n/a | /* Skip the leading quote char. */ |
|---|
| 5057 | n/a | s++; |
|---|
| 5058 | n/a | len = strlen(s); |
|---|
| 5059 | n/a | if (len > INT_MAX) { |
|---|
| 5060 | n/a | PyErr_SetString(PyExc_OverflowError, |
|---|
| 5061 | n/a | "string to parse is too long"); |
|---|
| 5062 | n/a | return -1; |
|---|
| 5063 | n/a | } |
|---|
| 5064 | n/a | if (s[--len] != quote) { |
|---|
| 5065 | n/a | /* Last quote char must match the first. */ |
|---|
| 5066 | n/a | PyErr_BadInternalCall(); |
|---|
| 5067 | n/a | return -1; |
|---|
| 5068 | n/a | } |
|---|
| 5069 | n/a | if (len >= 4 && s[0] == quote && s[1] == quote) { |
|---|
| 5070 | n/a | /* A triple quoted string. We've already skipped one quote at |
|---|
| 5071 | n/a | the start and one at the end of the string. Now skip the |
|---|
| 5072 | n/a | two at the start. */ |
|---|
| 5073 | n/a | s += 2; |
|---|
| 5074 | n/a | len -= 2; |
|---|
| 5075 | n/a | /* And check that the last two match. */ |
|---|
| 5076 | n/a | if (s[--len] != quote || s[--len] != quote) { |
|---|
| 5077 | n/a | PyErr_BadInternalCall(); |
|---|
| 5078 | n/a | return -1; |
|---|
| 5079 | n/a | } |
|---|
| 5080 | n/a | } |
|---|
| 5081 | n/a | |
|---|
| 5082 | n/a | if (fmode) { |
|---|
| 5083 | n/a | /* Just return the bytes. The caller will parse the resulting |
|---|
| 5084 | n/a | string. */ |
|---|
| 5085 | n/a | *fstr = s; |
|---|
| 5086 | n/a | *fstrlen = len; |
|---|
| 5087 | n/a | return 0; |
|---|
| 5088 | n/a | } |
|---|
| 5089 | n/a | |
|---|
| 5090 | n/a | /* Not an f-string. */ |
|---|
| 5091 | n/a | /* Avoid invoking escape decoding routines if possible. */ |
|---|
| 5092 | n/a | *rawmode = *rawmode || strchr(s, '\\') == NULL; |
|---|
| 5093 | n/a | if (*bytesmode) { |
|---|
| 5094 | n/a | /* Disallow non-ASCII characters. */ |
|---|
| 5095 | n/a | const char *ch; |
|---|
| 5096 | n/a | for (ch = s; *ch; ch++) { |
|---|
| 5097 | n/a | if (Py_CHARMASK(*ch) >= 0x80) { |
|---|
| 5098 | n/a | ast_error(c, n, "bytes can only contain ASCII " |
|---|
| 5099 | n/a | "literal characters."); |
|---|
| 5100 | n/a | return -1; |
|---|
| 5101 | n/a | } |
|---|
| 5102 | n/a | } |
|---|
| 5103 | n/a | if (*rawmode) |
|---|
| 5104 | n/a | *result = PyBytes_FromStringAndSize(s, len); |
|---|
| 5105 | n/a | else |
|---|
| 5106 | n/a | *result = decode_bytes_with_escapes(c, n, s, len); |
|---|
| 5107 | n/a | } else { |
|---|
| 5108 | n/a | if (*rawmode) |
|---|
| 5109 | n/a | *result = PyUnicode_DecodeUTF8Stateful(s, len, NULL, NULL); |
|---|
| 5110 | n/a | else |
|---|
| 5111 | n/a | *result = decode_unicode_with_escapes(c, n, s, len); |
|---|
| 5112 | n/a | } |
|---|
| 5113 | n/a | return *result == NULL ? -1 : 0; |
|---|
| 5114 | n/a | } |
|---|
| 5115 | n/a | |
|---|
| 5116 | n/a | /* Accepts a STRING+ atom, and produces an expr_ty node. Run through |
|---|
| 5117 | n/a | each STRING atom, and process it as needed. For bytes, just |
|---|
| 5118 | n/a | concatenate them together, and the result will be a Bytes node. For |
|---|
| 5119 | n/a | normal strings and f-strings, concatenate them together. The result |
|---|
| 5120 | n/a | will be a Str node if there were no f-strings; a FormattedValue |
|---|
| 5121 | n/a | node if there's just an f-string (with no leading or trailing |
|---|
| 5122 | n/a | literals), or a JoinedStr node if there are multiple f-strings or |
|---|
| 5123 | n/a | any literals involved. */ |
|---|
| 5124 | n/a | static expr_ty |
|---|
| 5125 | n/a | parsestrplus(struct compiling *c, const node *n) |
|---|
| 5126 | n/a | { |
|---|
| 5127 | n/a | int bytesmode = 0; |
|---|
| 5128 | n/a | PyObject *bytes_str = NULL; |
|---|
| 5129 | n/a | int i; |
|---|
| 5130 | n/a | |
|---|
| 5131 | n/a | FstringParser state; |
|---|
| 5132 | n/a | FstringParser_Init(&state); |
|---|
| 5133 | n/a | |
|---|
| 5134 | n/a | for (i = 0; i < NCH(n); i++) { |
|---|
| 5135 | n/a | int this_bytesmode; |
|---|
| 5136 | n/a | int this_rawmode; |
|---|
| 5137 | n/a | PyObject *s; |
|---|
| 5138 | n/a | const char *fstr; |
|---|
| 5139 | n/a | Py_ssize_t fstrlen = -1; /* Silence a compiler warning. */ |
|---|
| 5140 | n/a | |
|---|
| 5141 | n/a | REQ(CHILD(n, i), STRING); |
|---|
| 5142 | n/a | if (parsestr(c, CHILD(n, i), &this_bytesmode, &this_rawmode, &s, |
|---|
| 5143 | n/a | &fstr, &fstrlen) != 0) |
|---|
| 5144 | n/a | goto error; |
|---|
| 5145 | n/a | |
|---|
| 5146 | n/a | /* Check that we're not mixing bytes with unicode. */ |
|---|
| 5147 | n/a | if (i != 0 && bytesmode != this_bytesmode) { |
|---|
| 5148 | n/a | ast_error(c, n, "cannot mix bytes and nonbytes literals"); |
|---|
| 5149 | n/a | /* s is NULL if the current string part is an f-string. */ |
|---|
| 5150 | n/a | Py_XDECREF(s); |
|---|
| 5151 | n/a | goto error; |
|---|
| 5152 | n/a | } |
|---|
| 5153 | n/a | bytesmode = this_bytesmode; |
|---|
| 5154 | n/a | |
|---|
| 5155 | n/a | if (fstr != NULL) { |
|---|
| 5156 | n/a | int result; |
|---|
| 5157 | n/a | assert(s == NULL && !bytesmode); |
|---|
| 5158 | n/a | /* This is an f-string. Parse and concatenate it. */ |
|---|
| 5159 | n/a | result = FstringParser_ConcatFstring(&state, &fstr, fstr+fstrlen, |
|---|
| 5160 | n/a | this_rawmode, 0, c, n); |
|---|
| 5161 | n/a | if (result < 0) |
|---|
| 5162 | n/a | goto error; |
|---|
| 5163 | n/a | } else { |
|---|
| 5164 | n/a | /* A string or byte string. */ |
|---|
| 5165 | n/a | assert(s != NULL && fstr == NULL); |
|---|
| 5166 | n/a | |
|---|
| 5167 | n/a | assert(bytesmode ? PyBytes_CheckExact(s) : |
|---|
| 5168 | n/a | PyUnicode_CheckExact(s)); |
|---|
| 5169 | n/a | |
|---|
| 5170 | n/a | if (bytesmode) { |
|---|
| 5171 | n/a | /* For bytes, concat as we go. */ |
|---|
| 5172 | n/a | if (i == 0) { |
|---|
| 5173 | n/a | /* First time, just remember this value. */ |
|---|
| 5174 | n/a | bytes_str = s; |
|---|
| 5175 | n/a | } else { |
|---|
| 5176 | n/a | PyBytes_ConcatAndDel(&bytes_str, s); |
|---|
| 5177 | n/a | if (!bytes_str) |
|---|
| 5178 | n/a | goto error; |
|---|
| 5179 | n/a | } |
|---|
| 5180 | n/a | } else { |
|---|
| 5181 | n/a | /* This is a regular string. Concatenate it. */ |
|---|
| 5182 | n/a | if (FstringParser_ConcatAndDel(&state, s) < 0) |
|---|
| 5183 | n/a | goto error; |
|---|
| 5184 | n/a | } |
|---|
| 5185 | n/a | } |
|---|
| 5186 | n/a | } |
|---|
| 5187 | n/a | if (bytesmode) { |
|---|
| 5188 | n/a | /* Just return the bytes object and we're done. */ |
|---|
| 5189 | n/a | if (PyArena_AddPyObject(c->c_arena, bytes_str) < 0) |
|---|
| 5190 | n/a | goto error; |
|---|
| 5191 | n/a | return Bytes(bytes_str, LINENO(n), n->n_col_offset, c->c_arena); |
|---|
| 5192 | n/a | } |
|---|
| 5193 | n/a | |
|---|
| 5194 | n/a | /* We're not a bytes string, bytes_str should never have been set. */ |
|---|
| 5195 | n/a | assert(bytes_str == NULL); |
|---|
| 5196 | n/a | |
|---|
| 5197 | n/a | return FstringParser_Finish(&state, c, n); |
|---|
| 5198 | n/a | |
|---|
| 5199 | n/a | error: |
|---|
| 5200 | n/a | Py_XDECREF(bytes_str); |
|---|
| 5201 | n/a | FstringParser_Dealloc(&state); |
|---|
| 5202 | n/a | return NULL; |
|---|
| 5203 | n/a | } |
|---|