[Bug 1223673] Review Request: llvm35 - The Low Level Virtual Machine

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



https://bugzilla.redhat.com/show_bug.cgi?id=1223673



--- Comment #13 from Jens Petersen <petersen@xxxxxxxxxx> ---
(In reply to Milan Bouchet-Valat from comment #9)
> So what happens when llvm and llvm33/llvm34/llvm35 are installed, could we
> get conflicts about LLVMgold.so, BugpointPasses.so, libLTO.so and
> readline.so? I don't think so, or it wouldn't have worked.

I don't know. I am not aware of any problems, right.
I don't even know which of the programs if any are using them.
I couldn't see anything linking to libLTO for example
(of course something could be dlopen'ing them).

> > So I also feel we could waive this like was done for llvm34.
> Sure, let's investigate this for llvm first.

Yes - bug filed (comment 12).

> > > - Minor point: I realized each subpackage creates its own directory
> > >   under /usr/share/doc/. Since they contain very few files and most apply
> > >   to all subpackages (e.g. LICENSE.txt), wouldn't it be better to put
> > > everthing
> > >   under a common llvm dir?
> > 
> > This is true and also true for many other packages I think.
> > I'd rather just leave it for simplicity - though in principle
> > I agree with you completely, but I feel this is more a deficiency
> > of rpm.
> Not a big deal, but I think RPM handles this if you make the directory owned
> by llvm35-libs, which AFAICT all subpackages depend on.

I happy to take patch if you want to do that.
Sorry I don't really have time/energy/motivation to do it. :)
And again it is true for the main llvm package too I believe, right?

> > Hmm, isn't "runtime" pretty standard?
> Yeah, but I would do anything to shut down an annoying warning. :-)

These warnings are very frequent across many packages.
The spelling dictionaries are really too limited
or don't cover such technical usage.
I feel it is good to stay close to the original llvm.spec
too as far as possible.

> > Good question - perhaps Fedora could grab a patch from Debian?
> > Again I think this should be done first for the llvm package.
> > I opened bug 1258760 for that.
> Or, better, the manpages could be submitted upstream if Debian has some. But
> definitely not an issue for llvm35.

Yep

(In reply to Milan Bouchet-Valat from comment #11)
> Yes, a small comment wouldn't hurt. I think the guidelines say that -03
> should generally be avoided unless you can be sure it really improves
> performance.

Apparently it was added without comment in this commit:

commit be655c46e5d3707531fb8bef5430a9c064653197
Author: Jan Vcelak <jvcelak@xxxxxxxxxxxxxxxxx>
Date:   Tue Nov 12 21:48:50 2013 +0100

    update to 3.3, add compiler-rt and lldb
:    

Not sure how much thought was put into it.

I sent a mail to Jan asking about it.
I opened bug 1272394 to track this too!
We can wait for his reply or go ahead I guess
since I consider this a backport from the llvm package.


Can you live with the current spec file or what do you think
must be changed still for approval? I hope I didn't miss anything. :)

-- 
You are receiving this mail because:
You are on the CC list for the bug.
You are always notified about changes to this product and component
_______________________________________________
package-review mailing list
package-review@xxxxxxxxxxxxxxxxxxxxxxx
https://admin.fedoraproject.org/mailman/listinfo/package-review




[Index of Archives]     [Fedora Legacy]     [Fedora Desktop]     [Fedora SELinux]     [Yosemite News]     [KDE Users]     [Fedora Tools]