On 07/29/10 08:23, Benjamin Wolsey wrote:
> These (in all versions) are almost certainly to do with using
> implementation-defined or undefined behaviour when bitshifting or
> when numeric operations overflow. I can probably fix most of them
> without access to the machine.
Classic endian issues... I can test any patch or branch before
checking it in if you want.
> It's nice to have a real case to test against for these cases!
Yeah, in a strange way, I was glad to see the failures. :-) I was also
glad to see the results were better than expected, since none of us have
tried running Gnash big-endian in a long time.
> This is known to be dependent on the timing of the movie load.
I have problems with this on other platforms too. There may not be
many powerpc users left out there, but without Gnash, they can't waste
time viewing YouTube like everyone else...
- rob -
_______________________________________________
Gnash-dev mailing list
Gnash-dev@gnu.org
http://lists.gnu.org/mailman/listinfo/gnash-dev
