Hi,
I did as you requested, and installed and tested
the latest version of festival in unstable:
1.4.3-22.
Then I tested it with the following command
$ counter=0; while echo $counter | festival --language americanenglish -b
--tts ; do counter=$(( counter + 1 )) ; done ; echo Crashed after $(( $counter
-1 )) runs.
It reported
underrun!!! (at least 1192.858 ms long)
and then after running some more
Segmentation fault (core dumped)
Crashed after 55 runs.
Here's the output from the "bt" (back trace)
command in "gdb /usr/bin/festival core"
(gdb) bt
#0 0x00000000 in ?? ()
#1 0xb7bb98e4 in ?? () from /usr/lib/libestbase.so.1.2
#2 0xb7f14364 in ?? () from /usr/lib/libestools.so.1.2
#3 0x00000001 in ?? ()
#4 0x00000000 in ?? ()
#5 0x08142cf4 in ?? ()
#6 0x00000002 in ?? ()
#7 0x081b0b18 in ?? ()
#8 0xb7f14364 in ?? () from /usr/lib/libestools.so.1.2
#9 0xb7018fb0 in ?? ()
#10 0x00000003 in ?? ()
#11 0xbfe465a8 in ?? ()
#12 0xb7ea2405 in load_ngram_cstr_bin () from /usr/lib/libestools.so.1.2
Previous frame identical to this frame (corrupt stack?)
So, it works better than the old version, but
still has at least one bug.
The second time I tested it, it ran 100 times
without error, before I stopped it.
>From out here in the peanut gallery, it looks like
underruns may corrupt the stack.
Thanks for the improvement,
Kingsley
On 10/20/07 23:22, Kartik Mistry wrote:
> Hi,
>
> Should we still keep this bug open? What is result with latest package
> from unstable. Let me know.
>
> The bug report is really informative!
>
> Thanks.
>
> --
> Cheers,
> -----------------------------------------------------------
> Kartik Mistry || GPG: 0xD1028C8D || IRC: kart_
> kartikmistry.org/blog || kartikm.wordpress.com
> ----------------------------------------------------------
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
