[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Fishing for a General Permissions API (inspired by "importers"

From: Derek Robert Price
Subject: Re: Fishing for a General Permissions API (inspired by "importers"
Date: Tue, 16 Mar 2004 12:19:44 -0500
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030624 Netscape/7.1

Hash: SHA1

Mark D. Baushke wrote:

> Would the commitinfo script be called once for the entire
> tree as loginfo is called at present or once for each
> directory in the import tree? This has implications for how
> the contrib/commit_prep.in and contrib/log_accum examples
> presently operate.

Good question.  I would have said that it should be called for each
directory if you hadn't mentioned that loginfo wasn't being called
separately.  Perhaps it should be kept in sync?  As long as commit_prep
is called only once, I think log_accum will still work properly.  I'm
not sure I understand all the issues involved, but you are probably
right that a once-per-directory/once-per-commit combination would break
the examples.

> It might also be useful for the 'cvs add' and 'cvs import'
> commands to share logic with 'cvs commit' as to validation
> that the creation of a new directory in the repository.

Hrm.  Yes, probably, but beyond the scope of my voting for acceptance of
this patch, I think.

> >ALL $CVSROOT/CVSROOT/myscript %r %p $T %{t} -- %{sV}
> I do not understand the $T in the above line. Where is it
> set?

Er, it should have been %T, and I meant it to be the destination
branch/vendor branch.  Maybe %B would be better?  Anyhow, taginfo is
using %t for the name of the tag being created and %b as a
branch/non-branch flag, and I was trying to remain somewhat consistent.
I thought %t/%T was somewhat consistent with the %{Vv} being used for
old revision number/new revision number model being used for loginfo &

And %T isn't being set yet.  I was just suggesting it.

> For a 'cvs import' I would be under the impression that
> there are two active tags in the operation: 1) the vendor
> tag, 2) the version tag.

Actually, there can be more than one release tag (what you are calling a
vesion tag):

Usage: cvs import [-d] [-k subst] [-I ign] [-m msg] [-b branch]
    [-W spec] repository vendor-tag release-tags...

I don't know how often people use that functionality, but as long as it
is enabled, it would need to be accounted for.

> It is still not clear to me if you
> intend to integrate commitinfo with the taginfo operation by
> calling the taginfo script during import or not...

Ouch.  I hadn't thought of that.  It's a thought, but I was still
thinking of leaving that to the commitinfo script...  if a user really
wanted, it shouldn't be too hard to call their own taginfo script from
their commitinfo script on an import.  I think I would vote for leaving
taginfo out of the import until it looks like it is really necessary, or
at least useful, to separate the functionality.

[Issue #157 kept in sync.]

- --

Email: derek@ximbiot.com

Get CVS support at <http://ximbiot.com>!
Version: GnuPG v1.2.1 (GNU/Linux)
Comment: Using GnuPG with Netscape - http://enigmail.mozdev.org


reply via email to

[Prev in Thread] Current Thread [Next in Thread]