GNU bug report logs - #35490
Creating a ./configure diff against gzip-1.9

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 35490 in the body.
You can then email your comments to 35490 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Ilya Leoshkevich <iii <at> linux.ibm.com>
To: bug-gzip <at> gnu.org
Subject: Creating a ./configure diff against gzip-1.9
Date: Mon, 29 Apr 2019 17:08:57 +0200

Hi,

I’m trying to backport IBM Z patches to gzip version 1.9.  The code
changes themselves apply without problems, however, I also need to
create a diff for configure and other generated build files.  If I just
use whatever distro I have at hand for this, there will be a lot of
noise from autoconf and automake version mismatch, so I'm trying to
figure out which exact versions were used for the 1.9 release.

I'm observing a few surprising things, and it would be great if someone
could help me resolve them and get to the point where I can reproduce
the 1.9 release from git sources:


- The released ./Makefile.in contains:

    distdir-am: $(DISTFILES)

which means that automake version used contained the commit ac47c2:
"make dist" did not depend on $(BUILT_SOURCES) from Nov 28 2017.

At the same time, the released ./build-aux/ar-lib contains:

    # Copyright (C) 2010-2015 Free Software Foundation, Inc.

which means that automake must have been older than commit d8add5:
maint: Update copyright years to 2017. from Mar 2 2017.

This is clearly impossible, and the only other version indication that I
could find is:

    Makefile.in generated by automake 1.15a

which does not really help, since this correspond to a lot of commits.
Can it be that a private branch was used?


- The released ./configure contains:

    Generated by GNU Autoconf 2.69.193-8974

However, 8974 is not an existing commit id.  Again, would it be correct
to assume that this is coming from a private branch?  The closest
version I could come up with is 2.69.193-dfb0.


- The am__DIST_COMMON list in the released ./Makefile.in contains
  build-aux/ar-lib.

However, there is no AM_PROG_AR or similar macros in the released
./configure.ac.  When I do `make dist`, it is not generated.  Where is
it supposed to come from?


- The released ./aclocal.m4 contains:

    # po.m4 serial 24 (gettext-0.19)

However, the released ./configure.ac does not contain
AM_GNU_GETTEXT_VERSION - where is po.m4 coming from?


Best regards,
Ilya

Message #8 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Paul Eggert <eggert <at> cs.ucla.edu>
To: Ilya Leoshkevich <iii <at> linux.ibm.com>
Cc: bug-gzip <at> gnu.org
Subject: Re: bug#35490: Creating a ./configure diff against gzip-1.9
Date: Thu, 3 Oct 2019 18:13:45 -0700

On 4/29/19 8:08 AM, Ilya Leoshkevich wrote:
> If I just
> use whatever distro I have at hand for this, there will be a lot of
> noise from autoconf and automake version mismatch, so I'm trying to
> figure out which exact versions were used for the 1.9 release.

This is not something I'd normally worry about, any more than I'd 
normally worry about exactly what x86-64 code gets generated by my C 
compiler.

Jim Meyering (who created the release) had his own platform with 
reasonably-bleeding-edge tools, not all of which correspond to any 
particular GNU release. He might be able to fill you in with the 
details, but this sort of thing is kinda low priority so I wouldn't 
expect a quick answer (and even he might not know the answer).

I skipped most of your questions, except for:

> where is po.m4 coming from?

I think it comes from Gnulib, and is filled in by './bootstrap'.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.