11:29 p.m.
I have created a patch for fedora-newrpmspec:
http://tkmame.retrogames.com/fedora-extras/fedora-newrpmspec.patch
This works with the current spectemplate-pear file located here:
http://tkmame.retrogames.com/fedora-extras/spectemplate-pear.spec
Can you packaging guys please review this patch to make sure
everything looks okay?
Thanks!
5:34 p.m.
On Thu, 2006-07-06 at 21:29 -0700, Christopher Stone wrote:
I have created a patch for fedora-newrpmspec:
http://tkmame.retrogames.com/fedora-extras/fedora-newrpmspec.patch
This works with the current spectemplate-pear file located here:
http://tkmame.retrogames.com/fedora-extras/spectemplate-pear.spec
Can you packaging guys please review this patch to make sure
everything looks okay?
It's best if you file a bug against fedora-rpmdevtools - that way, it
won't get lost in the shuffle.
David
5:54 p.m.
On 7/7/06, David Lutterkort <dlutter(a)redhat.com> wrote:
On Thu, 2006-07-06 at 21:29 -0700, Christopher Stone wrote:
> I have created a patch for fedora-newrpmspec:
> http://tkmame.retrogames.com/fedora-extras/fedora-newrpmspec.patch
>
> This works with the current spectemplate-pear file located here:
> http://tkmame.retrogames.com/fedora-extras/spectemplate-pear.spec
>
> Can you packaging guys please review this patch to make sure
> everything looks okay?
It's best if you file a bug against fedora-rpmdevtools - that way, it
won't get lost in the shuffle.
Well we currently depend on php-pear to drop in some macros, I would
have to file a bug against that first. But I think a packaging member
should file the bug for php-pear and place it at high priority (should
only take 15 minutes to patch up) because if I file the php-pear bug,
it will never get fixed.
I can make a sample macros.pear file if needed.
6:09 p.m.
On 7/7/06, Christopher Stone <chris.stone(a)gmail.com> wrote:
I can make a sample macros.pear file if needed.
I have put up a sample macros.pear file here:
http://tkmame.retrogames.com/fedora-extras/macros.pear
Tibbs: do these macros look okay? I can file a bug against php-pear
to add this if it looks okay with you, or perhaps you can make any
adjustments necessary and file the bug. Let me know. Thanks.
6:58 p.m.
>>>> "CS" == Christopher Stone
<chris.stone(a)gmail.com> writes:
CS> Tibbs: do these macros look okay?
It took me a while to figure out why pecl-related stuff would be in
those macros, but now I see that both /usr/bin/pear and /usr/bin/pecl
are in the php-pear package.
Everything there looks quite good to me. I just hope that we don't
stall out having to wait for an update to the core php-pear package.
However, there is a php-pear package in updates/testing at the moment;
perhaps we can get this snuck in.
- J<
10:37 p.m.
On 7/7/06, Jason L Tibbitts III <tibbs(a)math.uh.edu> wrote:
>>>>> "CS" == Christopher Stone
<chris.stone(a)gmail.com> writes:
CS> Tibbs: do these macros look okay?
It took me a while to figure out why pecl-related stuff would be in
those macros, but now I see that both /usr/bin/pear and /usr/bin/pecl
are in the php-pear package.
Everything there looks quite good to me. I just hope that we don't
stall out having to wait for an update to the core php-pear package.
However, there is a php-pear package in updates/testing at the moment;
perhaps we can get this snuck in.
That test release was for my bug:
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=196802
Which fixes stuff like:
http://linux.dell.com/files/fedora/FixBuildRequires/mock-results-extras/x...
I was thinking the same thing that these two could be pushed together.
Hopefully Joe is watching this mailing list and knows what we are up
to. :)
11:23 p.m.
On 7/7/06, Jason L Tibbitts III <tibbs(a)math.uh.edu> wrote:
>>>>> "JLT" == Jason L Tibbitts,
<Jason> writes:
JLT> Everything there looks quite good to me.
Crap, did some of those need to be %global instead of %define? All of
them? I don't completely understand the underlying issue.
I think that was if we used the form:
%{!?__php: %define __php /usr/bin/php}
Which we don't do in this case.
The issue in question is layed out here:
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=147238
9:29 a.m.
chris.stone(a)gmail.com ("Christopher Stone") writes:
I have put up a sample macros.pear file here:
http://tkmame.retrogames.com/fedora-extras/macros.pear
How do you want to use these macros? Place them into /etc/rpm/? If so,
you must remove the '%define ' and write only '%'. E.g.
| %__pear %{_bindir}/pear
instead of
| %define __pear %{_bindir}/pear
Then, current practice seems to be, to use a leading underscore for
directory names (e.g. '%_libdir', '%_bindir'). Hence I would prefer
| %_pecl_phpdir
instead of
| %pecl_phpdir
Enrico
9:38 a.m.
Le samedi 08 juillet 2006 à 16:29 +0200, Enrico Scholz a écrit :
Then, current practice seems to be, to use a leading underscore for
directory names (e.g. '%_libdir', '%_bindir'). Hence I would prefer
The current practice is to use a leading underscore before system macro
names to avoid collisions with in-spec definitions I think
So %pecl_phpdir can probably become _pecldir or _php-pecldir
--
Nicolas Mailhot
11:30 a.m.
>>>> "ES" == Enrico Scholz
<enrico.scholz(a)informatik.tu-chemnitz.de> writes:
ES> Then, current practice seems to be, to use a leading underscore
ES> for directory names (e.g. '%_libdir', '%_bindir').
A counterexample, which is what we are copying, from /usr/lib/rpm/macros:
%perl_sitearch %(eval "`%{__perl} -V:installsitearch`"; echo $installsitearch)
%perl_sitelib %(eval "`%{__perl} -V:installsitelib`"; echo $installsitelib)
%perl_vendorarch %(eval "`%{__perl} -V:installvendorarch`"; echo
$installvendorarch)
%perl_vendorlib %(eval "`%{__perl} -V:installvendorlib`"; echo
$installvendorlib)
%perl_archlib %(eval "`%{__perl} -V:installarchlib`"; echo $installarchlib)
%perl_privlib %(eval "`%{__perl} -V:installprivlib`"; echo $installprivlib)
- J<
2:07 p.m.
On 7/8/06, Jason L Tibbitts III <tibbs(a)math.uh.edu> wrote:
>>>>> "ES" == Enrico Scholz
<enrico.scholz(a)informatik.tu-chemnitz.de> writes:
ES> Then, current practice seems to be, to use a leading underscore
ES> for directory names (e.g. '%_libdir', '%_bindir').
A counterexample, which is what we are copying, from /usr/lib/rpm/macros:
%perl_sitearch %(eval "`%{__perl} -V:installsitearch`"; echo
$installsitearch)
%perl_sitelib %(eval "`%{__perl} -V:installsitelib`"; echo $installsitelib)
%perl_vendorarch %(eval "`%{__perl} -V:installvendorarch`"; echo
$installvendorarch)
%perl_vendorlib %(eval "`%{__perl} -V:installvendorlib`"; echo
$installvendorlib)
%perl_archlib %(eval "`%{__perl} -V:installarchlib`"; echo $installarchlib)
%perl_privlib %(eval "`%{__perl} -V:installprivlib`"; echo $installprivlib)
Okay, I have updated the macros file here:
http://tkmame.retrogames.com/fedora-extras/macros.pear
Let me know if this looks okay.
Thanks.
4:59 p.m.
On Sat, 2006-07-08 at 12:07 -0700, Christopher Stone wrote:
On 7/8/06, Jason L Tibbitts III <tibbs(a)math.uh.edu> wrote:
> %perl_sitearch %(eval "`%{__perl} -V:installsitearch`"; echo
$installsitearch)
> %perl_sitelib %(eval "`%{__perl} -V:installsitelib`"; echo
$installsitelib)
> %perl_vendorarch %(eval "`%{__perl} -V:installvendorarch`"; echo
$installvendorarch)
> %perl_vendorlib %(eval "`%{__perl} -V:installvendorlib`"; echo
$installvendorlib)
> %perl_archlib %(eval "`%{__perl} -V:installarchlib`"; echo
$installarchlib)
> %perl_privlib %(eval "`%{__perl} -V:installprivlib`"; echo
$installprivlib)
Okay, I have updated the macros file here:
http://tkmame.retrogames.com/fedora-extras/macros.pear
Let me know if this looks okay.
We have leading-underscoreless %perl_* in rpm, %python_* in spec
templates and upstream rpm, and %ruby_* in the forthcoming ruby spec
template and ruby packaging guidelines; any reason for pear/pecl to be
different?
5:33 p.m.
On 7/9/06, Ville Skyttä <ville.skytta(a)iki.fi> wrote:
On Sat, 2006-07-08 at 12:07 -0700, Christopher Stone wrote:
> On 7/8/06, Jason L Tibbitts III <tibbs(a)math.uh.edu> wrote:
>
> > %perl_sitearch %(eval "`%{__perl} -V:installsitearch`"; echo
$installsitearch)
> > %perl_sitelib %(eval "`%{__perl} -V:installsitelib`"; echo
$installsitelib)
> > %perl_vendorarch %(eval "`%{__perl} -V:installvendorarch`"; echo
$installvendorarch)
> > %perl_vendorlib %(eval "`%{__perl} -V:installvendorlib`"; echo
$installvendorlib)
> > %perl_archlib %(eval "`%{__perl} -V:installarchlib`"; echo
$installarchlib)
> > %perl_privlib %(eval "`%{__perl} -V:installprivlib`"; echo
$installprivlib)
>
> Okay, I have updated the macros file here:
>
> http://tkmame.retrogames.com/fedora-extras/macros.pear
>
> Let me know if this looks okay.
We have leading-underscoreless %perl_* in rpm, %python_* in spec
templates and upstream rpm, and %ruby_* in the forthcoming ruby spec
template and ruby packaging guidelines; any reason for pear/pecl to be
different?
I put the underscores in because Enrico said the opposite:
ES> Then, current practice seems to be, to use a leading underscore for
ES> directory names (e.g. '%_libdir', '%_bindir'). Hence I would
prefer
ES>
ES> | %_pecl_phpdir
ES>
ES> instead of
ES>
ES> | %pecl_phpdir
So I guess I will remove the underscores to be in line with the other
language directory definitions.
4:58 a.m.
On Sun, Jul 09, 2006 at 03:33:13PM -0700, Christopher Stone wrote:
So I guess I will remove the underscores to be in line with the
other
language directory definitions.
I've submitted a new php-pear to FC-5 testing which includes this latest
macros.pear; can you test this out? Thanks a lot for your work here.
joe
3:36 p.m.
On 7/10/06, Joe Orton <jorton(a)redhat.com> wrote:
On Sun, Jul 09, 2006 at 03:33:13PM -0700, Christopher Stone wrote:
> So I guess I will remove the underscores to be in line with the other
> language directory definitions.
I've submitted a new php-pear to FC-5 testing which includes this latest
macros.pear; can you test this out? Thanks a lot for your work here.
I rebuilt all my pear packages with the new php-pear and its macros
and everything worked great.
6:35 p.m.
>>>> "CS" == Christopher Stone
<chris.stone(a)gmail.com> writes:
CS> I rebuilt all my pear packages with the new php-pear and its
CS> macros and everything worked great.
Good. So where are we at? We have nice macros which should be in FC5
soon. Which leaves:
* FC4. Will these macros make it there, or do we just not target that
release?
* Guidelines. They need to be altered to take into account the
existence of these new macros.
* Modules neither PEAR nor PECL. We still need guidelines for this
class of packages, if for no other reason than it would let me
finish reviewing php-shout.
* Specfile templates. We either need a template or fedora-newrpmspec
should call into the specfile generator module.
- J<
6:45 p.m.
On 7/12/06, Jason L Tibbitts III <tibbs(a)math.uh.edu> wrote:
>>>>> "CS" == Christopher Stone
<chris.stone(a)gmail.com> writes:
CS> I rebuilt all my pear packages with the new php-pear and its
CS> macros and everything worked great.
Good. So where are we at? We have nice macros which should be in FC5
soon. Which leaves:
...
* Specfile templates. We either need a template or
fedora-newrpmspec
should call into the specfile generator module.
I have created a bug for the php-pear-* template:
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=198706
2:59 p.m.
Trying to resurrect the PHP discussion so that we can finally put this
to bed.
>>>> "JLT" == Jason L Tibbitts, <Jason>
writes:
JLT> We have nice macros which should be in FC5 soon.
As far as I can tell, the updated php-pear package isn't released yet.
JLT> * FC4. Will these macros make it there, or do we just not target
JLT> that release?
I'm going to take this as a no. Is it problematic that we're not
going to target FC4?
JLT> * Guidelines. They need to be altered to take into account the
JLT> existence of these new macros.
I have done a first pass cleanup of what we have at
http://fedoraproject.org/wiki/PackagingDrafts/PHP
Still needed there:
A block of Provides and Requires for PECL modules analogous to those
for PEAR modules.
Scriptlets for PECL module registration. Or maybe not; I have no idea
if PECL modules need to be registered in the same way that PEAR
modules are.
More comments about non-PEAR-PECL modules.
JLT> * Specfile templates. We either need a template or
JLT> fedora-newrpmspec should call into the specfile generator module.
What happened with this? Chris Stone had a patch; did it get reported
upstream?
Is there something analogous that we could do for PECL modules?
- J<
3:36 p.m.
On Wed, 2006-07-19 at 14:59 -0500, Jason L Tibbitts III wrote:
>>>>> "JLT" == Jason L Tibbitts,
<Jason> writes:
JLT> We have nice macros which should be in FC5 soon.
As far as I can tell, the updated php-pear package isn't released yet.
JLT> * FC4. Will these macros make it there, or do we just not target
JLT> that release?
I'm going to take this as a no. Is it problematic that we're not
going to target FC4?
See also https://bugzilla.redhat.com/198706#c4 and the rest of the
comments there: IMHO with pretty nonintrusive changes to the
suggested/submitted template, no new external macros at all would be
needed.
6:20 p.m.
>>>> "VS" == Ville Skytt <Ville> writes:
VS> IMHO with pretty nonintrusive changes to the suggested/submitted
VS> template, no new external macros at all would be needed.
I didn't realize that a ticket had been opened for that. Honestly I
don't care what is accepted; I just think we need to make some forward
progress. But after a quick look at your template, it does seem
really nice, especially the file list generation. I wonder if we
couldn't do that for Python, where the complexity involving .pyo files
is often needlessly confusing. Another topic, I guess.
Obviously it would be nice to not have to wait for an update to the
core php-pear package before we move ahead. On the other hand, Joe
was nice enough to add the macros for us; we should at least be sure
not to conflict with them.
So what to do? We can't really proceed with the guidelines (at least
for the PEAR bits) without knowing what's going into
fedora-newrpmspec.
We also really need to decide whether PECL modules need a template.
I'm guessing they do, but I haven't actually looked at one.
Finally, to answer a question in that ticket, all PEAR modules are
indeed noarch and all PECL modules are arch-specific, by definition as
I understand it.
- J<
6491
days inactive
6503
days old
packaging@lists.fedoraproject.org
20 comments
7 participants
participants (7)
-
Christopher Stone -
David Lutterkort -
Enrico Scholz -
Jason L Tibbitts III -
Joe Orton -
Nicolas Mailhot -
Ville Skyttä