[XviD-devel] [In progress] devapi4 -- mpeg matrices

[Edouard Gomez]

Michael Militzer (michael at xvid.org) wrote:
> You said it yourself: 'take every part that doesn't work and make _it_
> working' - that's  what bugfixing is all about,  that's what we agreed
> on to do and that's what I expect you to do.

Grrrr.

Sorry, but i don't like the final "you" in your sentence. No one told me
you're the  boss and that  sysKin and I  owe you obediance in  every all
aspect of your decisions.

First your "so  polished code" proved to be wrong  for xmm/3dne for very
long time[1].  How many users noticed  ?  0. How many devs noticed ?  1,
sysKin.  This is  not the only remanent problem of  legacy code in xvid.
Lumimasking was  never designed to  work for bframes, 2pass  was working
because of the moon phases, ME was unmaintainable/monolithic. The age of
code has  never been a  proof of stability.   It only tells us  that the
code behaves  the same, and  that if there  are bugs, then we  are lucky
because they don't hurt on the end.

Now concerning  the matrix  fixing, you know  I tried the  small hacking
way[2], and  i failed.  So  perhaps i'm not  that good in  assembler and
fixing so many  functions is out of my skills...  but  at least i tried.
Then despite my  call for help[3], i got no  answer, probably people was
busy.   So I  proposed to  simply replace  the code  with code  that has
always been  designed to be safe,  and once i  started happily replacing
code, you appear telling me it would be better to fix it the "small hack
way".  The flame war started  mostly because your mail was despising the
efforts that syskin and I, are putting in XviD.

The  only  acceptable  answer  i  can  get  from  you  now,  is  a  code
contribution.

<in crude words>
Contribute or shut up
</in crude words>

This should be the rule #0, especialy for people who contributed so much
and  has the  sufficient knowledge  to do  it. People  who has  not this
knowledge is excusable, but you aren't.

For me it's  the end of this counter productive  thread. I'm waiting for
your fixes ASAP.

[1] DC offset bug
http://www.xvid.org/cvs/chora/co.php/xvidcore/src/quant/x86_asm/quantize4_xmm.asm?login=2&r=1.2.2.1

[2] The small hack way report
http://list.xvid.org/pipermail/xvid-devel/2003-November/003723.html

[2] The Call for help
http://list.xvid.org/pipermail/xvid-devel/2003-November/003724.html

-- 
Edouard Gomez