On Sun, Apr 17, 2011 at 12:22:24AM +0200, Dominik Vogt wrote:
> 3. Import the cvs history with

Just clone from here:

https://github.com/ThomasAdam/fvwm

The "master" branch is tracking CVS.  I've not updated it with all the
recent changes, but I will be.

This will save you approx 10+ hours waiting for the import.  There's a lot
of history here.  :)

>   $ git cvs-import -d :pserver:[email protected]:/home/cvs/fvwm fvwm
> 
> (you can skip the -d option if CVSROOT is set).
> 
> 4. Wait until the import finishes.  That can take a long time.  I
> guess it will be much faster if the git repository is created on
> the same machine as the cvs repository.

See above.

Can we hold off a while though going gun-ho into shoving any Git repository
in place?  I've already started writing a lengthy email as it is about the
impact of this, and also how we go about defining development workflows,
etc., as well as all the helper scripts that will need updating, as well as
fvwm-web.

That's not to mention training and an ease of migration for people not used
to changeset-based revision control systems.

Give me till the end of tomorrow and I'll have everything laid out by then,
but I'd rather I propose the impact with key areas on migration, etc., for
developers before we just drop this in and expect people to suddenly think
it works like CVS [1].

Because it doesn't.

And I've had a lot of experience doing just this sort of thing in $DAY_JOB.

-- Thomas Adam

[1] I'm not trying to be a bottleneck here, or a killjoy, but I know how
these things pan out.  Without proper management and suggestions on common
idioms between Git/CVS, this won't work well.

-- 
"It was the cruelest game I've ever played and it's played inside my head."
-- "Hush The Warmth", Gorky's Zygotic Mynci.

Reply via email to