develooper Front page | perl.perl5.porters | Postings from December 2016

[perl #130375] Porting/release_managers_guide.pod: need advice renew directories created by CPAN synch

Thread Next
From:
James E Keenan via RT
Date:
December 30, 2016 15:02
Subject:
[perl #130375] Porting/release_managers_guide.pod: need advice renew directories created by CPAN synch
Message ID:
rt-4.0.24-16877-1483110138-951.130375-15-0@perl.org
On Mon, 19 Dec 2016 00:21:28 GMT, jkeen@verizon.net wrote:
> Porting/release_managers_guide.pod needs advise re new directories
> created by CPAN synch
> 
> This is a bug report for perl from jkeenan@cpan.org,
> generated with the help of perlbug 1.40 running under perl 5.24.0.
> 
> 
> -----------------------------------------------------------------
> [Please describe your issue here]
> 
> Over the weekend both BinGOs and I were, in the course of helping to
> prepare for this week's monthly development version "blead-point"
> release, bit by the same problem.
> 
> We were attempting to synchronize the Test-Simple distribution found
> underneath the 'cpan/' directory in the core distribution with the
> latest version of that distro on CPAN.  Specifically, we wanted to go
> from 1.302067 to 1.302071.  We (or, at least, I -- I'm assuming we
> each
> went through the same steps) got to the same point in the instructions
> in Porting/release_managers_guide.pod concerning CPAN-to-blead
> synchronization:
> 
> #####
> =item *
> 
> Run a full configure/build/test cycle.
> 
> =item *
> 
> If everything is ok, commit the changes.
> #####
> 
> Note the absence of guidance concerning what to do if everything is
> *not* ok.  In the course of the final 'make test_harness', I got a
> failure in 't/porting/regen.t' which, when I ran that file by itself,
> appeared as:
> 
> #####
> not ok # Makefile.SH is up to date
> ok # win32/Makefile is up to date
> ok # win32/makefile.mk is up to date
> #####
> 
> This was puzzling, as the 'diff' of the old Test-Simple against the
> new
> showed no changes in Makefile.SH.  I sprinkled 't/porting/regen.t'
> with
> 'print' statements, ran it through the debugger, etc., with no good
> results.  Finally I noticed this line in that test file:
> 
> #####
> require './regen/regen_lib.pl';
> #####
> 
> ... and wondered what would happen when I ran that program by itself
> with the newly build './perl'.  When I did so, I got a whole lot of
> changes in Makefile.SH; see attachment.  I re-ran 'make test_porting'
> and got PASS.  I then re-ran the full configure-build-test cycle and
> got
> PASS.   And today on advice from arc on #p5p, I pushed to blead.  In a
> response to an inquiry from Test-Simple maintainer Exodist, arc said,
> 
> #####
> (03:10:59 PM) arc: Exodist: not afaict — it's a p5-internal thing
> about
> dealing with the fact that the lib/Test2/Event/TAP/ directory is new
> in
> the latest version (containing a file Version.pm)
> (03:11:36 PM) arc: we have to list all cpan dists' directories in the
> makefile, so that "make clean" can clean up the versions that are
> created under lib/ by the build system
> #####
> 
> I ran a couple of 'ack's over the core distribution to try to find any
> discussion of this requirement and couldn't find anything quickly.
> Hence, I infer that the release manager's guide needs updating.  We
> need:
> 
> * explicit documentation of this requirement for Makefile.SH,
> including
> why ('make clean') we need it at all;
> 
> * to decide where in the discussion of CPAN-to-blead synchronization
> guidance should go -- perhaps earlier than the last step;
> 
> * to decide whether 'regen/regen_lib.pl' should be discussed in the
> RLM
> -- it currently is not.
> 
> * discussion as to whether there are any other ways in which a
> CPAN-to-blead synch could go wrong at the very last step and how such
> failures should be handled.
> 
> Thank you very much.
> Jim Keenan
> 

From irc.perl.ort #p5p:

#####
(09:21:11 AM) kid51: arc: Perhaps in the RMG we should (a) mention synch-with-cpan *first*; (b) discuss what we're trying to achieve; then (c) discuss specific distros where synch-with-cpan is likely to be insufficient
(09:25:20 AM) arc: that sounds like a good plan
#####

Aaron, if we decide to go with the approach above, the language attached might be a starting point.

Thank you very much.


-- 
James E Keenan (jkeenan@cpan.org)

---
via perlbug:  queue: perl5 status: new
https://rt.perl.org/Ticket/Display.html?id=130375

Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About