On Thu, 26 Feb 2026 16:44:49 +0000 Anatoly Burakov <[email protected]> wrote:
> Add support for parsing floating point numbers in cmdline library, as well > as unit tests for the new functionality. Use C library for parsing. > > Signed-off-by: Anatoly Burakov <[email protected]> > --- Some things AI spotted that I missed. [PATCH v12 2/3] cmdline: add floating point support Error: In float_valid_strs[], the entry {"-2E308", HUGE_VAL} has the wrong sign — it should be {"-2E308", -HUGE_VAL}. The preceding entry {"-2e308", -HUGE_VAL} has the correct sign. This is a copy-paste bug. Warning: In cmdline_parse_num(), the float write path uses if (res && write_float(num_data.type, res, dres)) without the < 0 comparison, while the integer path uses if (res && write_num(num_data.type, res, uintres) < 0). The bare truthiness check happens to work because -1 is truthy in C, but it's inconsistent and fragile — if write_float ever returns a positive error code, this breaks. Should be write_float(...) < 0 to match write_num.
