On Sun, Feb 13, 2005 at 10:18:35AM -0000, Paul Volkaerts wrote:
> > I use FC2 but I cannot reproduce this problem (though I use CVS
> > and not the binaries).  If someone can reproduce it, and can send
> > me a backtrace,  or tell me how to reproduce it, I will fix it.
> 
> I found the backtrace on the atrpms-devel mail archive --- thanks Joe.
> 
> This is a problem with festival which has a strange workaround that I have
> never gotten to the bottom of.
> 
> Workround #1 -- Compile without festival!
> 
> Workround #2 -- You need to edit the file speech_tools\siod\slib.cc and do a
> find & replace of the variable backtrace to backtrace1.  For some reason,
> the variable backtrace evaluates to the wrong address at link time.  I have
> greped for backtrace in myth and can find no reference to a variable of this
> name.
> 
> Axel --- unfortunately this is going to affect everyone! Can you please
> rebuild the binaries using one of the two workrounds above?  Festival is no
> good anyway unless you are packaging the speech files with it, so if you are
> not doing that I would go for option #1. Other than this the festival code
> is stable.

I'll go with #1, that will enable building mythphone for x86_64,
too. Festival does come with a few speech files, but some are indeed
not packaged due to licensing issues.

> Anyone any idea why the link fails ?

Adding to the confusion: Why does it fail with 0.17 and not 0.16?
-- 
Axel.Thimm at ATrpms.net

Attachment: pgp8zxEgPhiK7.pgp
Description: PGP signature

_______________________________________________
mythtv-users mailing list
[email protected]
http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users

Reply via email to