On 14 Jan, J.D. wrote:

> GCC 2.96 is not an official release version of GCC, so
> weird bugs spawned by the compiler should not be
> surprising.  However, the GCC 3.0.3 release also has given
> me similar trouble...
> 
> Internal buffer inconsistency. flushbits <> ResvSizeInternal buffer
> inconsistency. flushbits <> ResvSizebit reservoir error:
> 
> To work around this, compile with "-O" instead of "-O3"
> with GCC 3.0.3, or fall back to a 2.95 version.  (With an
> Athlon, GCC 3.0.3 with "-O" seems to yield better LAME
> performance than GCC 2.95 with "-O3".)

I was under the impression, I already fixed it in the CVS tree.

I also have an additional commit pending, I need to get time to evaluate
some benchmark results first.

> Meanwhile, someone should isolate the problem and file a
> GCC bug report, if this optimization bug has not yet been
> fully characterized.  Hopefully, the GCC developers get
> around to hammering the bugs out of version 3 someday...

The GCC developers usually want to have a short example whit the
breakage, so the submitter has at least to find the file which needs to
be compiled with "-O3 -funroll-loops" to reveal the bug. Better would be
to compare the output after the assembly stage and to construct a small
example which reproduces the bug.

As I didn't have gcc 3.x, this task is up to someone else.

Bye,
Alexander.

-- 
Failure is not an option. It comes bundled with your Microsoft product.

http://www.Leidinger.net                       Alexander @ Leidinger.net
  GPG fingerprint = C518 BC70 E67F 143F BE91  3365 79E2 9C60 B006 3FE7

_______________________________________________
mp3encoder mailing list
[EMAIL PROTECTED]
http://minnie.tuhs.org/mailman/listinfo/mp3encoder

Reply via email to