[Bug 441] Review request: traverso - Multitrack Audio Recording and Editing Suite

[RPM Fusion Bugzilla]

http://bugzilla.rpmfusion.org/show_bug.cgi?id=441





--- Comment #29 from Orcan Ogetbil <oget.fedora at gmail.com>  2009-04-30 21:29:51 ---
(In reply to comment #28)
> (In reply to comment #26)
> > Jack using applications won't stop working, if you are asking that. My own
> > username priority is set to 90 from almost the beginning, and I never had
> > trouble with jack.
> So, that's an information, indeed. But why jack isn't set as 99 then since all
> is about been happier ?
> 

Good question. I don't know the answer. I don't maintain jack.

Which way do you prefer? A README.Fedora file or a limits file? I prefer the
latter one but I'll be OK with both.

> 
> > > Then if the values needs manual installation, it would be better to leave 
> > > the choice to the users, and have it documented in a README.Fedora.
> > > (along with a better warning message).
> > > 
> > 
> > I lost the count of how many times I asked you this question about
> > README.Fedora. You don't give direct answers.
> I have no directive to give you, you will remain responsible for the package.
> All is about knowing how things behave, then one can tell how things need to > be packaged.
> 

kwizart, I've been trying to guess what you actually mean by what you are
saying since the beginning and obviously I don't guess it the right way always.

We can avoid passing the ball back and forth, and avoid the possibility of
misinterpretations by minimal ambiguity.

> 
> > What is "x86 specific code"?
> This file : 'src/common/fpu.cc' is conditionalized with #ifndef ARCH_X86 which
> has nothing to deal with been SSE or not.
> I do think this will be a problem on ppc (or ia64) and it should have be:
>  #ifndef ARCH_X86
>         return;
> -#endif
> +#else
> ...
> +#endif (somewhere)
> 

Yes, thanks for pointing this out. As the author told me, this part of the code
is directly taken from ardour. It looks like they made a typo when they were
transferring (#else needs to be used instead of #endif), and they are going to
fix this in the next release. I'll update my sseopt patch to incorporate this
change.

> Now i don't see much difference between USE_XMMINTRIN and SSE
> at least traverso.cpp use that:
> #ifdef __SSE__
> #include <xmmintrin.h>
> #endif
> Which suggest one or the other should be avoided.
> 

As for USE_XMMINTRIN, again this is the way done in ardour. XMMINTRIN implies
SSE and vice versa both in ardour and in traverso. I know it is redundant but
there is no need to modify the code just to change the variable names.


-- 
Configure bugmail: http://bugzilla.rpmfusion.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.