[FFmpeg-devel] Clang run from 20100402

[Justin Ruggles]

Michael Niedermayer wrote:

> On Sat, Apr 03, 2010 at 11:49:03AM +0200, Benjamin Larsson wrote:
>> http://tranquillity.ath.cx/clang/2010-04-02-1/
>>
>> There are a few real issues. Please take a look into the cases that
>> cover the files you maintain. And Clang does make some mistakes.
> 
> ive looked at 4 in my files and they are all nonsese.
> I dont doubt that some of the issues are real but the signal
> to noise ratio is too poor IMHO
> also unless there is a way to silence false warnings without
> changing the source the effort is futile as in each new
> run the signal to noise ratio would get worse and we would spend
> much more time just rechecking false warnings.
> that said going over the coverity warnings likely already fixed most
> of these static analyzer findable warnings leaving even more noise.
> 
> thus, my humble oppinion, people should rather look at roundup and fix real
> issues until clang manages to reach at least a 50% signal to noise ratio
> 
> and calling it bugs is only true in the sense that they are clang bugs

All of the FLAC and AC-3 reports are false warnings.

At least the clang mistakes seem to be similar in nature.  I'm finding
mostly that they have to do with value bounds.  Where we correctly
assume certain bounds for a variety of reasons, but clang does not
recognize those bounds and takes branch combinations that cannot exist.

-Justin