[flac-dev] question about 0bea5fb commit

[Erik de Castro Lopo]

lvqcl wrote:

> The patch http://git.xiph.org/?p=flac.git;a=commitdiff;h=0bea5fb96436588b4e78c5be79bf174b9be7a202
> changes the type of t value from FLAC__int32 to uint32_t,
> but only for 24-bit signed input samples. 24-bit *un*signed
> input still uses FLAC__int32 t.

These UB fixes were driven purely by UBSan warnings generated while running
the test suite. The fact that no warning was gneerated by 24 bit unsigned
input suggests that the test suite doesn't cover 24 bit unsigned input.

That's probably something that should be fixed.

> So I wonder - does it makes sense to change this type too?
> The patch is attached.

That looks ok, but a test is needed first. I'll see if I can come up with
something.

Erik
-- 
----------------------------------------------------------------------
Erik de Castro Lopo
http://www.mega-nerd.com/