diff options
Diffstat (limited to 'Python')
| -rw-r--r-- | Python/ast.c | 1 | ||||
| -rw-r--r-- | Python/peephole.c | 18 |
2 files changed, 19 insertions, 0 deletions
diff --git a/Python/ast.c b/Python/ast.c index 590bc90..d97e951 100644 --- a/Python/ast.c +++ b/Python/ast.c @@ -469,6 +469,7 @@ set_context(struct compiling *c, expr_ty e, expr_context_ty ctx, const node *n) case Set_kind: case Num_kind: case Str_kind: + case Bytes_kind: expr_name = "literal"; break; case Ellipsis_kind: diff --git a/Python/peephole.c b/Python/peephole.c index 06e7fe6..c27878c 100644 --- a/Python/peephole.c +++ b/Python/peephole.c @@ -124,6 +124,24 @@ fold_binops_on_constants(unsigned char *codestr, PyObject *consts) break; case BINARY_SUBSCR: newconst = PyObject_GetItem(v, w); + /* #5057: if v is unicode, there might be differences between + wide and narrow builds in cases like '\U00012345'[0]. + Wide builds will return a non-BMP char, whereas narrow builds + will return a surrogate. In both the cases skip the + optimization in order to produce compatible pycs. + */ + if (newconst != NULL && + PyUnicode_Check(v) && PyUnicode_Check(newconst)) { + Py_UNICODE ch = PyUnicode_AS_UNICODE(newconst)[0]; +#ifdef Py_UNICODE_WIDE + if (ch > 0xFFFF) { +#else + if (ch >= 0xD800 && ch <= 0xDFFF) { +#endif + Py_DECREF(newconst); + return 0; + } + } break; case BINARY_LSHIFT: newconst = PyNumber_Lshift(v, w); |