[Bug 836835] New: RFE: Rebuild perl-Math-Pari with the bundled pari-2.3 and update to pari-2.5

[Red Hat Bugzilla]

https://bugzilla.redhat.com/show_bug.cgi?id=836835

Bug ID: 836835
        QA Contact: [e-mail]
          Severity: unspecified
           Version: rawhide
          Priority: unspecified
                CC: [e-mail], [e-mail]
          Assignee: [e-mail]
           Summary: RFE: Rebuild perl-Math-Pari with the bundled pari-2.3
                    and update to pari-2.5
        Regression: ---
      Story Points: ---
    Classification: Fedora
                OS: Unspecified
          Reporter: [e-mail]
              Type: Bug
     Documentation: ---
          Hardware: Unspecified
        Mount Type: ---
            Status: NEW
         Component: perl-Math-Pari
           Product: Fedora

I found some bug reports after some googling about issues with pari-2.5, and,
when attempting to convert Pari.xs to pari-2.5 api it got too much confusing
due to internal api changes.

Since perl-Math-Pari should be the only showstopper of pari-2.5 update, I
believe it should be ok to build perl's Pari.so with an internal pari
implementation without depending on system pari.

I agree this is somewhat hackish, but I believe better than creating a pari23
package.

Spec URL: http://fedorapeople.org/~pcpa/perl-Math-Pari.spec
SRPM URL: http://fedorapeople.org/~pcpa/perl-Math-Pari-2.010806-14.fc18.src.rpm

I am sending a mail to the perl-Math-Pari author as specified in
http://search.cpan.org/~ilyaz/Math-Pari-2.01080605/Pari.pm as bugzilla does not
allow adding someone not in database to CC...

[Bug 836835] RFE: Rebuild perl-Math-Pari with the bundled pari-2.3 and update to pari-2.5

[Red Hat Bugzilla]

https://bugzilla.redhat.com/show_bug.cgi?id=836835

--- Comment #1 from pcpa  ---
For openness, the email I sent to the perl-Math-Pari author is:

-%<-
para ilyaz
 Hi,

I am working on updating and making update requests for fedora
rawhide with a goal of packaging sagemath for fedora 18, and one
pending issue is perl-Math-Pari that does not build with pari-2.5.

I suppose it should be easier for the author to come with a patch,
as I failed after some time trying, and am just proposing to link
perl's Pari.so module with a private pari-2.3 build, but this
is agains't fedora policies.

I reported the problem at https://bugzilla.redhat.com/show_bug.cgi?id=836835

Many Thanks,
Paulo
-%<-

[Bug 836835] RFE: Rebuild perl-Math-Pari with the bundled pari-2.3 and update to pari-2.5

[Red Hat Bugzilla]

https://bugzilla.redhat.com/show_bug.cgi?id=836835

--- Comment #2 from Paul Howarth  ---
(In reply to comment #0)
> I found some bug reports after some googling about issues with pari-2.5,
> and, when attempting to convert Pari.xs to pari-2.5 api it got too much
> confusing due to internal api changes.
>
> Since perl-Math-Pari should be the only showstopper of pari-2.5 update, I
> believe it should be ok to build perl's Pari.so with an internal pari
> implementation without depending on system pari.
>
> I agree this is somewhat hackish, but I believe better than creating a
> pari23 package.

On the other hand, that would require a library bundling exemption from FESCO,
which might not be granted. The libpari23 package I have ready would not
require that, you could review it and once it's in place we could then update
pari to 2.5.1 without breaking anything.

If I submit the libpari23 package for review, will you review it?

[Bug 836835] RFE: Rebuild perl-Math-Pari with the bundled pari-2.3 and update to pari-2.5

[Red Hat Bugzilla]

https://bugzilla.redhat.com/show_bug.cgi?id=836835

--- Comment #3 from pcpa  ---
(In reply to comment #2)
> (In reply to comment #0)
> > I found some bug reports after some googling about issues with pari-2.5,
> > and, when attempting to convert Pari.xs to pari-2.5 api it got too much
> > confusing due to internal api changes.
> >
> > Since perl-Math-Pari should be the only showstopper of pari-2.5 update, I
> > believe it should be ok to build perl's Pari.so with an internal pari
> > implementation without depending on system pari.
> >
> > I agree this is somewhat hackish, but I believe better than creating a
> > pari23 package.
>
> On the other hand, that would require a library bundling exemption from
> FESCO, which might not be granted. The libpari23 package I have ready would
> not require that, you could review it and once it's in place we could then
> update pari to 2.5.1 without breaking anything.
>
> If I submit the libpari23 package for review, will you review it?

Yes, I will review it!

[Bug 836835] RFE: Rebuild perl-Math-Pari with the bundled pari-2.3 and update to pari-2.5

[Red Hat Bugzilla]

https://bugzilla.redhat.com/show_bug.cgi?id=836835

--- Comment #4 from Paul Howarth  ---
libpari23 review request submitted:
https://bugzilla.redhat.com/show_bug.cgi?id=837004

[Bug 836835] RFE: Rebuild perl-Math-Pari with the bundled pari-2.3 and update to pari-2.5

[Red Hat Bugzilla]

https://bugzilla.redhat.com/show_bug.cgi?id=836835

Paul Howarth  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |CLOSED
   Fixed In Version|                            |perl-Math-Pari-2.010806-14.
                   |                            |fc18
         Resolution|---                         |RAWHIDE
        Last Closed|                            |2012-07-02 15:29:53

--- Comment #5 from Paul Howarth  ---
perl-Math-Pari is now build against libpari23 in Rawhide (f18-perl)