autoconf-archive-maintainers | |
[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Serial lines for Automake
From: |
Francesco Salvestrini |
Subject: |
Re: Serial lines for Automake |
Date: |
Sat, 18 Jul 2009 00:06:41 +0200 |
User-agent: |
KMail/1.9.10 |
Hi Peter,
On Friday 17 July 2009, Peter Simons wrote:
> Hi Francesco,
>
> > A git-pourcelain/git-hook approach could avoid running git-log + grep
> > over more than 500 macros on each dist target run (the count get
> > worse if the dist target fails ...).
>
> how would that procedure work? I am aware of the fact that there are
> hooks in Git, but I don't know what exactly they do or when they are
> triggered.
Briefly:
A hook is a script placed in $GIT_DIR/hooks directory and it is triggered
into the local repository or into the remote one, at certain points (commits,
receive-packs etc.).
See http://book.git-scm.com/5_git_hooks.html (some examples) or
http://www.kernel.org/pub/software/scm/git/docs/githooks.html
The pre-commit hook should be the good one. It should be activated *only* into
the release repository (your repository, Peter).
The hook use doesn't solve the problem obviously, it (could) avoid running
long operations while building the distribution tarball.
But ... since we don't have a solution yet I would like to postpone this
discussion (keeping doors open for further Q&A as well ;-) ).
> > We will need to keep the numbering scheme coherent, once it is
> > stable: we shouldn't decrease a macro serial once it has been
> > released ... we could lead into user related problems otherwise.
>
> I see is one particular problem with those serial lines. Say the Archive
> distributes a macro AX_FOO at version X. Now Joe Doe downloads the file,
> patches AX_FOO, and re-distributes it on his homepage. How can Joe
> compute a serial "X + 1" that won't conflict with a version that the
> Archive might distribute some day?
We could partition the number-space by using a slight different scheme than
the initial one. The previous one was:
# NUMBER
The newer could be something like:
# NUMBER1.NUMBER2 COMMENT
Where NUMBER1 is:
*) 0 for the archive macros
*) >= 0 for third-party tweaks/changes/redistribution
eg:
# 0.XXXX Please change 0.XXXX to 1.YYYY if you are going to change this macro
NOTE: The format used for the serial line is the extended one '# NUMBER
GARBAGE' where NUMBER could be a dotted\.
I don't see a full-fledged-fool-proof solution at the moment. The solution
could solve problems generated by some "misuse" of the archive itself (eg:
the archive installed system-wise and some user copies scattered in the
aclocal include path).
> The Archive has tried to provide accurate version information for the
> last 6 years or so, but to be honest, it never worked well because of
> that problem.
I see. We should keep into account all the problems, prioritizing them
accordingly. I likely wouldn't apply the macro versioning (or other
versioning approach as well) until all pros/cons have emerged.
Ciao,
Francesco
> Take care,
> Peter
--
You'll feel devilish tonight. Toss dynamite caps under a flamenco dancer's
heel.
- [autoconf-archive-maintainers] Serial lines for Automake, Peter Simons, 2009/07/17
- Re: [autoconf-archive-maintainers] Serial lines for Automake, Francesco Salvestrini, 2009/07/17
- Re: Serial lines for Automake, Peter Simons, 2009/07/17
- Re: Serial lines for Automake,
Francesco Salvestrini <=
- Re: Serial lines for Automake, Peter Simons, 2009/07/18
- Re: Serial lines for Automake, Francesco Salvestrini, 2009/07/18
- Re: Serial lines for Automake, Filippo Giunchedi, 2009/07/18
- Re: Serial lines for Automake, Dustin J. Mitchell, 2009/07/18
- Re: Serial lines for Automake, Francesco Salvestrini, 2009/07/19
- Re: Serial lines for Automake, Francesco Salvestrini, 2009/07/19