Re: [Gnu-arch-users] cscvs *huge* memory usage [was: Re: ViewARCH-0.0.2]

[Robert Anderson]

On Fri, 2003-09-12 at 08:09, address@hidden wrote:
> Robert Anderson <address@hidden> writes:
> 
> > Why do you want to do that?  The history is still there, and accessible
> > through CVS like it always was.  You've not "lost" anything to just
> > import your latest sources and go from there.
> 
> Because I don't want to have to use 2 tools to access my project's
> history.

You already know CVS and that's the way you would have done it anyway. 
I don't see any "loss" there.

> I also want "annotate" support and I want it to be able to give me useful 
> info for history that predates migration to arch. etc.

One possible way to do this is the other way around - through CVS, until
the migration boundary is sufficiently removed from where you are
working.

Commit to CVS with each arch commit, preferably with a trigger.  Use CVS
for annotate when the boundary is an issue.

> If history is not integrated, then, for all practical purposes, it
> might as well be lost.

I went through exactly the same process you're going through, and I
expended a lot of effort with the same preconceived idea that my history
was "lost" if I didn't redundantly copy it into arch somehow.  It turned
out to be a huge time sink, and for basically nothing.  My  humble
advice: just import and start working.  If and when you have a real
show-stopper due to the format boundary, work on it then.

Bob