RFC: RPM Changelog thoughts
Paul W. Frields
stickster at gmail.com
Mon Oct 31 17:23:54 UTC 2005
On Mon, 2005-10-31 at 07:40 -0600, Tommy Reynolds wrote:
> I'm looking at the RPM packaging steps that Paul has done. AIUI, he
> uses a small python script to pick the changelog information out of
> the document's XML. Cool idea.
It is, if only I were indeed doing that! The only thing I do is a grab
of the title information out of <title> within the <book> or <article>.
I am just learning Python baby steps, so even that was a stretch.
Didn't want to take credit where it wasn't due. I had considered this,
though, as an eventual goal for Python learning... if I could figure out
enough SAX stuff to make this work, that would be a real coup as far as
I'm concerned. I was under the impression that there were ways to even
parse and read back entities, but I have no idea if that's a correct
impression.
> My question is this: should we build the RPM changelog from the XML
> content or from the CVS check-in log? Using the XML info is way
> cool, but we can get more file-specific information from the CVS log.
> I think developers, er, RPM users, would like the additional details.
>
> Opinions?
I think the CVS log would be nice, but my understanding is that it's
difficult to parse for this content. A couple difficulties that
occurred to me were (1) tying the CVS login name to an email address,
which is normally used in the spec file's %changelog; (2) mitigating the
situation with the CVS changelog having entries that do not correspond
to RPM package versions -- in other words, CVS gets revisions for which
a new package is not rolled; and (3) tying the CVS revision number into
a version number used for the RPM package. I don't know if any of these
are showstoppers, but they are definitely gobsmackers, if you get my
meaning.
On the other hand, the DocBook XML <revisionhistory> at least has a
"rational" version number and date which should roughly correspond with
packaging. We still have the email address problem, however, and other
problems likely exist with this way of doing it. *sigh* This is so much
easier dealing with regular code like in Extras!
--
Paul W. Frields, RHCE http://paul.frields.org/
gpg fingerprint: 3DA6 A0AC 6D58 FEC4 0233 5906 ACDB C937 BD11 3717
Fedora Documentation Project: http://fedora.redhat.com/projects/docs/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : http://lists.fedoraproject.org/pipermail/docs/attachments/20051031/eae07e56/attachment.bin
More information about the docs
mailing list