Re: bzr repository ready?

[Karl Fogel]

Jason Earl <address@hidden> writes:
> I will take a look at this over the weekend, but I think that you are
> over-estimating how much I know about CVS, Bzr, Git, tags, Emacs, and
> Free Software in general.  Really my only qualifications are that I was
> able to get both bzr's cvsps-import and bzr-fast-import to work on the
> Emacs repository :).  Andreas did the heavy lifting when it came to
> making a suitable repository.  From his replies to you on the problems
> with the bzr repository it is very clear to me that you and he both know
> more about source code repositories than I even *want* to know.
>
> I probably need another smiley after that last sentence.
>
> On the other hand, I do like learning new things, and it is quite
> possible that all that is really needed is for someone to chase down a
> few bugs (something I am more than capable of doing).  So I am happy to
> help, ecstatic even.  I just don't want to get anyone's hopes up about a
> fast turnaround.

No worries, Jason -- just do whatever you have time to.  We're really
close now.

I'd say just insert some debugging prints into your importers: figure
out where they're discovering the tag information, stick the debugging
statements there, and see if that prints out (or doesn't print out) any
of the tags I listed as missing.  You have the big advantage of having
actually done the bzr imports, which Andreas and I haven't.

It may well be that there is an innocent explanation: perhaps those tags
didn't cover all the files, so bzr had no way (?) to represent them, as
tags in bzr are basically symbolic names for whole-tree revisions --
whereas in CVS a tag is attached to a specific revision in each file, so
only user convention enforces that a given tag appears in all files.

(That's something I could look at in the CVS records, but it would be
great if we could get inside the guts of the bzr converter and see it
actually making that decision.)

-Karl