Re: [patch] advice regarding make regen products
[email protected] (Jim Cromie)
| Newsgroups | perl.perl5.porters |
|---|---|
| Message-ID | <[email protected]> |
On Mon, Jan 4, 2010 at 3:56 PM, Jim Cromie <[email protected]> wrote: > On Fri, Jan 1, 2010 at 1:58 PM, karl williamson <[email protected]> wrote: >> Jim Cromie wrote: >>> >>> doc recent advice about including changes >>> wrought by 'make regen' in patches. >>> >> >> This helped me understand things. And either I misunderstand, or it could >> be made clearer that a series of patches may be an exception to the rule. >> Something like: >> >> +submit your patch without them. An exception to this might be if +you're >> submitting a series of >> +patches: it might be best to submit the regenerated changes >> +immediately after the source-changes that caused them, so as to have >> +as little effect as possible on the bisectability of your patchset. >> > > hi Karl, > > Im glad it helped. > I was rather hoping that the process might be further discussed and > perhaps refined. > > The preference was (originally?) expressed here: > http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2009-11/msg00841.html > > I invented the separated "derived-products-only" patch approach as > satisfying the preference, but also further highlighting the "make > regen needed". > > Pumpkings clearly do the needed 'make regen' before committing such > patches to blead, but separating the derived files would allow a > double-check of sorts. > > So - what happens when a make-regen-needed commit is 'make regen'd, > followed by applying the 'derived-files' patch ? > Hopefully, the apply would succeed, and also say 'no files changed' > This would be a nice easy-to-use double-check. > re-raising the issue, with a patch to undo the advice.
0001-undo-make-regen-advice-bulk-of-94e892a68016a417b1b32.patch
(application/octet-stream, 1.5 KB)
From 4362a8bba9a964e65dcb66f90c50776ec0755264 Mon Sep 17 00:00:00 2001 From: Jim Cromie <[email protected]> Date: Wed, 16 Jun 2010 20:38:12 -0600 Subject: [PATCH] undo 'make regen' advice, bulk of 94e892a68016a417b1b324213d39309b8b744d0e revert 'make regen' special case advice, leave the minor whitespace cleanup. --- pod/perlrepository.pod | 11 ----------- 1 files changed, 0 insertions(+), 11 deletions(-) diff --git a/pod/perlrepository.pod b/pod/perlrepository.pod index 7cf75e9..05ae621 100644 --- a/pod/perlrepository.pod +++ b/pod/perlrepository.pod @@ -417,17 +417,6 @@ $install_root/lib. If you are unsure about the proper location of a file that may have gotten copied while building the source distribution, consult the C<MANIFEST>. -As a special case, several files are regenerated by 'make regen' if -your patch alters C<embed.fnc>. These are needed for compilation, but -are included in the distribution so that you can build perl without -needing another perl to generate the files. You must test with these -regenerated files, but it is preferred that you instead note that -'make regen is needed' in both the email and the commit message, and -submit your patch without them. If you're submitting a series of -patches, it might be best to submit the regenerated changes -immediately after the source-changes that caused them, so as to have -as little effect as possible on the bisectability of your patchset. - =for XXX What should we recommend about binary files now? Do we need anything? -- 1.6.6.1