[Mplayer-cvslog] CVS: main/libmpdemux demux_roq.c,1.4,1.5

[Mike Melanson]

On Sun, 31 Mar 2002, Arpi wrote:

> > and to you personally simply asking a few general questions about how
> > seeking worked. I didn't ask you to jump into the code and fix it for
> you actually did.
> 
> 'i did it in the right way. it doens't work. help me!'

	You certainly read more into that than I intended; there must be
some language barrier. All I asked at the end of the email was, "What am I
missing?" I didn't ask anyone to actually bother looking at the code. I
just wanted to be sure I had the theory straight. Now you say that I
do. Now I'm really lost (but it may be time to look at it again).

> ok, so my general idea for such mails: try to fix the bug.
> i just didn't wanted to flood the lists with such useless mails.

	What are you talking about? Why on earth do you think I'm writing
the email? I want to fix the bug, I truly do; I just get stuck sometimes
and need some advice.

> reading -dev-eng anyway, or just seen that problem with x11 and isn't my bug
> and left to the x11 code maintaienrs.
> don't expect me to fix all bugs.

	Trust me, I expect very little from you, and less and less all the
time. And I sure don't expect anyone else to fix my decoders because I
know that no one else understands them.

> i did it already on vd_rle.c. you just removed it and replaced with your
> buggy one, why do you expect me to write it again? then remove again?
> if you think your code is better - let's hack it yourself...

	If you recall, the old code had a much more serious bug: A
licensing bug.

> ppl really should stop commiting useless features and concentrate to
> bugfixing and restoring stability, if we keep going this way, then

	What do you think I'm doing? I told you that I'm on a strict
no-new-features policy until the I clean up my list of bugs and
outstanding issues. MS RLE was on the list. So was MS IMA ADPCM. So was
RoQ framerate. As it stands, I'm afraid to finish the QT RLE because I
know you're going to get upset for one reason or another.

> > day. The reason is that I don't count on you or anyone else to care about
> > the bugs even though they're in your code. OTOH, I do care about the bugs
> you just expect us to fix bugs in YOUR code... ehh.

	Can you find a single example of where I specifically requested
that other people fix my code? A real example, not one based on your 
warped perception. I definitely don't expect other people to fix my code
because I know I'm largely the only person who understands my own
code. Case in point, there's still an obnoxious bug in the RoQ video
decoder. There's no way I'd even bother asking for help with that because
I know that no one else understands it (except Dr. Tim, but he's busy,
too).

> > cares about that format." No one cares about the MS IMA ADPCM format,
> > either, so why do you keep harping on that? (Even though it does work
> > now.)
> because it worked before - so your commits broke it, so you should fix that.

	See the licensing bug point. Remember my personal goal: Open
source decoders. Binary-only solutions are suboptimal.

-- 
	-Mike Melanson