Also in test_decimal, there's a small leak here:


==10040== 24 bytes in 1 blocks are definitely lost in loss record 549 of 5,095
==10040==    at 0x4C2DE56: malloc (vg_replace_malloc.c:299)
==10040==    by 0x643B33: fstring_compile_expr (parse_string.c:594)
==10040==    by 0x643B33: fstring_find_expr (parse_string.c:924)
==10040==    by 0x643B33: fstring_find_literal_and_expr (parse_string.c:1076)
==10040==    by 0x643B33: _PyPegen_FstringParser_ConcatFstring (parse_string.c:1293)
==10040==    by 0x644569: fstring_parse (parse_string.c:1409)
==10040==    by 0x644569: fstring_find_expr (parse_string.c:980)
==10040==    by 0x644569: fstring_find_literal_and_expr (parse_string.c:1076)
==10040==    by 0x644569: _PyPegen_FstringParser_ConcatFstring (parse_string.c:1293)
==10040==    by 0x62CE94: _PyPegen_concatenate_strings (pegen.c:2003)
==10040==    by 0x62EF52: strings_rule (parse.c:10834)
==10040==    by 0x62EF52: atom_rule (parse.c:10674)
==10040==    by 0x6389A2: t_primary_raw (parse.c:14042)
==10040==    by 0x6389A2: t_primary_rule (parse.c:13839)
==10040==    by 0x638D67: star_target_rule (parse.c:12684)
==10040==    by 0x6392FC: star_targets_rule (parse.c:12501)
==10040==    by 0x63BD7B: _tmp_135_rule (parse.c:23255)
==10040==    by 0x63BD7B: _loop1_22_rule (parse.c:16468)
==10040==    by 0x63BD7B: assignment_rule (parse.c:2116)
==10040==    by 0x63BD7B: small_stmt_rule (parse.c:1508)
==10040==    by 0x63DB44: simple_stmt_rule (parse.c:1406)
==10040==    by 0x63F995: statement_rule (parse.c:1240)
==10040==    by 0x63F995: _loop1_11_rule (parse.c:15835)
==10040==    by 0x63F995: statements_rule (parse.c:1175)
==10040==    by 0x63FB49: block_rule (parse.c:6127)