[Gnu-arch-users] Re: cscvs totla data and Arch revisions

[Robert Collins]

On Mon, 2004-08-30 at 17:06 +1200, Martin Langhoff wrote:
> I am reviewing the "cscvs totla" data in the sqlite database. I was
> hoping to get enough data from the database tables to match Arch
> revisions to totla csets. Curiously, the data isn't there.

I'm working quite heavily on cscvs at the moment, so may be able to add
this in the near future... if its needed.

> So my questions go out to Robert, Charles and the list: 
> 
>  - isn't there a mapping between cscvs changesets to the comitted
> revisions? If it is there, where should I find it?

Yes, in the tla commit messages, there is a CSCVS-ID: header.

>  - is it possible to get that info from `tla commit ` and put it in
> the sqlite database? 

Why?

> Reading through runTla() in totla.py, it seems
> stdout is being discarded. I am tempted to try and capture it and do
> some nasty perl-ism there, like parse the output for a match to
> m/\-patch\-(\d+)/ and then put $1 in a new table in the database.
> 
> The goals are to automatically:
> 
>  - be able to create a cached revision on every important symbolic name/tag.

The info for this is there already - cscvs cache -T should help.

>  - be able to create a new "tla branch" on every symbolic name/tag
> that marks the beginning of a branch, and then use totla to track
> those branches too.

That is on my Todo already :}. But there are some major issues - notably
the different syntax for tla branches & cvs branches.

> I am trying to put an arch gateway for a cvs-ridden project, which is
> very tidy in its use of branch names for symbolic and branch tags.
> While I can open each branch manually, I really want to automate it.

Do you mind me asking, which project?

Rob

signature.asc
Description: This is a digitally signed message part