[flac-dev] [Flac-dev] Git branch with compiling fixes for win32

[Miroslav Lichvar]

Hi Josh,

nice to see you here again.

On Wed, Apr 25, 2012 at 04:26:05PM -0700, Josh Coalson wrote:
> (Jumping in again, maybe at the wrong point since this doesn't seem
> to involve encoding, but here goes.)
> 
> Miroslav's patches have always been high-quality for sure.  But
> regardless of submitter, any patch that affects encoding must be
> reviewed very carefully, preferably by several other people and
> definitely me.  If there were ever a libFLAC release that had a bug
> and was not always lossless, that would be very damaging to the
> format.

The bitreader patch touches only the rice decoding code which I
believe is very well covered by the test suite and any bugs would be
quickly seen. Also, it has also been included in the Fedora packages
for several years, no bug reports about MD5 mismatch were received
yet :).

It makes the C function faster than the corresponding asm routine, so
if it's included I'd suggest to just drop the asm function to not keep
around more asm code than is necessary.

I'm not sure if anyone is planning to port the asm code to x86_64, I
think that it will be quite a lot of work, perhaps it would be a good
time to reconsider using inline assembly instead of nasm to minimize
the amount of asm code? It would be useful to know how much are the
individual asm routines actually faster, it has been a long time
since I played with it.

-- 
Miroslav Lichvar