Static analysis reveals that `tok_mode->last_expr_buffer[i] != '\0' && i < input_length` is not safe

[sobolevn]

Bug report

This code can be improved:

cpython/Parser/lexer/lexer.c

Line 143 in 132b6bc

while (tok_mode->last_expr_buffer[i] != '\0' && i < input_length) {

It would be safer to first check the value of i and then try to access [i] index. This way it is harder to get read out of bounds.

Found by PVS-Studio in https://habr.com/ru/companies/pvs-studio/articles/902048/

Linked PRs

• gh-132769: Refactor possible read-out-of-bounds in lexer.c #132770
• [3.13] gh-132769: Refactor possible read-out-of-bounds in lexer.c (GH-132770) #132788