[vorbis] Compression Artifacts at -q 5. Help!

[Nemo -earth native-]

On Sun, Jun 15, 2003 at 05:25:08PM -0400, Ross Vandegrift did utter:
> 
> Well, it's also a fact of life that PC hardware is flaky.  Should free
> software developers support my software if I have broken RAM?  I think
> you have to draw the line somewhere, and the way I see it, the compiler
> needs to be assumed to work, just like the hardware.

And when the hardware is known flaky, the software takes it into account. 
(look at all the manual workaround in the linux kernel config for
example... the pentium division bug springs to mind). 

Sure, nobody can predict what bugs will be in the next round of hardware
of compilers, but we do know what problems exist in previous round. So
when an existing/old compiler is known to be flaky, isn't it prudent
to do the same?  Do a basic check for compiler version, and if it's a
*specific* version that is known to be a problem, then warn the user in
big bold letters.

I'm not a fan of generalised "gcc3 and above only" though. 

As a side thought - it's the complex floating point that causes problems,
right? ...How does the integer encoder (tremor?) do on the old buggy gcc?

.../Nemo

-- 
  ------------------------------------------ --------------------------
                                                    earth native
--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to 'vorbis-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.