Bug #23761 for PAR: pp etc. missing

Mon Dec 04 12:06:32 2006 GAAL [...] cpan.org - Ticket created

Subject:

pp etc. missing - broken distro?

This version of PAR seems to be missing some files, e.g. pp and parl.

Mon Dec 04 13:27:46 2006 smueller [...] cpan.org - Correspondence added

On Mon Dec 04 12:06:32 2006, GAAL wrote: Show quoted text

> This version of PAR seems to be missing some files, e.g. pp and parl.

Hi, Since version 0.970 of PAR, it has been split into two distributions: PAR and PAR-Packer. PAR contains PAR.pm and related code for using .par archives as libraries. It is pure-Perl and does not require a compiler. PAR-Packer contains PAR::Packer, pp, parl and friends and requires a C compiler which was a problem for those users who used PAR in conjunction with .par files as libraries. Please install PAR-Packer-0.970.tar.gz from CPAN. Naturally, PAR cannot have a dependency on PAR-Packer or else the split would be useless. Steffen

Mon Dec 04 13:27:48 2006 The RT System itself - Status changed from 'new' to 'open'

Mon Dec 04 13:27:50 2006 smueller [...] cpan.org - Status changed from 'open' to 'rejected'

Mon Dec 04 13:53:14 2006 gaal [...] forum2.org - Correspondence added

Subject:	Re: [rt.cpan.org #23761] pp etc. missing - broken distro?
Date:	Mon, 4 Dec 2006 20:50:52 +0200
To:	Steffen Müller via RT <bug-PAR [...] rt.cpan.org>
From:	Gaal Yahas <gaal [...] forum2.org>

On Mon, Dec 04, 2006 at 01:27:49PM -0500, Steffen Müller via RT wrote: Show quoted text

> Since version 0.970 of PAR, it has been split into two distributions: > PAR and PAR-Packer. PAR contains PAR.pm and related code for using .par

Thanks for the reply, and sorry for missing this in the documentation. Would you consider adding a 'recommends' directive on PAR-Packer? This will probably cut down on future duplicates to this bug :) If you don't want to confuse cc-less users, the recommends can be predicated upon 'can_cc', similar to the crypto stuff now. -- Gaal Yahas <gaal@forum2.org> http://gaal.livejournal.com/

Mon Dec 04 13:53:15 2006 The RT System itself - Status changed from 'rejected' to 'open'

Tue Dec 05 05:39:53 2006 smueller [...] cpan.org - Correspondence added

On Mon Dec 04 13:53:14 2006, gaal@forum2.org wrote: Show quoted text

> On Mon, Dec 04, 2006 at 01:27:49PM -0500, Steffen Müller via RT wrote:

> > Since version 0.970 of PAR, it has been split into two distributions: > > PAR and PAR-Packer. PAR contains PAR.pm and related code for using .par

> > Thanks for the reply, and sorry for missing this in the documentation.

Don't worry about it. I didn't expect that anyone who upgrades PAR reads the change log. After all, most changes are just bug fixes. I'm sorry the split has caused you trouble. But it's a logical step to take and thus I wanted it to happen before PAR 1.00! Show quoted text

> Would you consider adding a 'recommends' directive on PAR-Packer? This > will probably cut down on future duplicates to this bug :) If you don't > want to confuse cc-less users, the recommends can be predicated upon > 'can_cc', similar to the crypto stuff now.

Okay, I added that to the trunk of PAR but I won't make a release just for that. It's a good idea to ease the transition to two separate distributions. I'm not a huge fan of these "feature" and "recommends" verbs of M::Install because they cause trouble for the automatic installers, but in this case, it's sensible. Steffen

Tue Dec 05 05:39:55 2006 smueller [...] cpan.org - Status changed from 'open' to 'resolved'

Tue Dec 05 05:47:15 2006 smueller [...] cpan.org - Correspondence added

On Mon Dec 04 13:53:14 2006, gaal@forum2.org wrote: Show quoted text

> On Mon, Dec 04, 2006 at 01:27:49PM -0500, Steffen Müller via RT wrote:

> > Since version 0.970 of PAR, it has been split into two distributions: > > PAR and PAR-Packer. PAR contains PAR.pm and related code for using .par

> > Thanks for the reply, and sorry for missing this in the documentation. > > Would you consider adding a 'recommends' directive on PAR-Packer? This > will probably cut down on future duplicates to this bug :) If you don't > want to confuse cc-less users, the recommends can be predicated upon > 'can_cc', similar to the crypto stuff now.

Actually, I just changed my mind about this. It's not possible! PAR::Packer naturally depends on PAR. Thus I cannot possibly depend on PAR::Packer from within PAR's Makefile.PL even if it's just optional. Sorry! Steffen

Tue Dec 05 05:47:17 2006 The RT System itself - Status changed from 'resolved' to 'open'

Wed Dec 06 16:54:56 2006 GAAL [...] cpan.org - Correspondence added

From:

GAAL [...] cpan.org

On Tue Dec 05 05:47:15 2006, SMUELLER wrote: Show quoted text

> Actually, I just changed my mind about this. It's not possible! > PAR::Packer naturally depends on PAR. Thus I cannot possibly depend on > PAR::Packer from within PAR's Makefile.PL even if it's just optional.

You're right! Maybe the solution is to have PAR be a Tasklike package that requires (or just recommends) PAR-Packer, and rename the current PAR to PAR::Core or something? Then the upgrade path is secured, although it introduces another distro.

Thu Dec 07 06:00:22 2006 smueller [...] cpan.org - Correspondence added

From:

SMUELLER [...] cpan.org

On Wed Dec 06 16:54:56 2006, GAAL wrote: Show quoted text

> On Tue Dec 05 05:47:15 2006, SMUELLER wrote:

> > Actually, I just changed my mind about this. It's not possible! > > PAR::Packer naturally depends on PAR. Thus I cannot possibly depend on > > PAR::Packer from within PAR's Makefile.PL even if it's just optional.

> > You're right! > > Maybe the solution is to have PAR be a Tasklike package that requires > (or just recommends) PAR-Packer, and rename the current PAR to PAR::Core > or something? Then the upgrade path is secured, although it introduces > another distro.

That's problematic, too. It's convention to ship a module named equivalent to the distribution with the distribution. I.e. PAR-Packer should have a PAR::Packer module, etc. I know it's not mandatory, but I have been annoyed by authors naming their distributions by a different scheme. It can become *very* confusing. Now, if I adhere to the convention, this wouldn't work or I would have to rename PAR.pm to PAR::Core which is, well, a bad idea! What lets me sleep well although I broke the upgrade path is that PAR::Packer/pp should only affect developers who, before giving up, are much more likely to have a glance at the ChangeLog if something goes awry. If you want me to do this change, please post to par@perl.org so we can get some more input on this. I'm a little reluctant. Steffen

Sat Feb 03 04:47:22 2007 smueller [...] cpan.org - Status changed from 'open' to 'rejected'

Sat Feb 03 04:47:22 2007 smueller [...] cpan.org - Severity Important changed to Normal

Bug #23761 for PAR: pp etc. missing - broken distro?