Re: [gentoo-scm] CVS -> git, list of where non-infra folk can contribute

[Brian Harring]

On Wed, Oct 03, 2012 at 06:22:11AM +0200, Michael Haggerty wrote:
> On 10/02/2012 06:15 AM, Brian Harring wrote:
> > [...]
> > 3) Robin afaik is putting together an email with the details; roughly,
> > the conversion process is conversion of cvs to svn, then svn2git
> > conversion; this is done since frankly it's the best/sanest conversion
> > pathway, and the fastest.  The validation of that conversion, and
> > getting it down to basically a set of known invocations is required.
>
> If you are using cvs2svn/cvs2git for the conversion, the I don't
> understand why you want to do the conversion via Subversion.  cvs2git
> (using the ExternalBlobGenerator) is much faster than cvs2svn.
>
> I look forward to hearing the details from Robin.

Well, it's 2012 now... optimization was from '10 (that 9x boost to 
cvs2svn came from us after all).

One thing to keep in mind with our tree; it's not exactly your 
standard source tree- for ebuilds, they sohuld up, rarely get 
modified, then get rename to -r{whatever}, or slightly tweaked and as 
a new version.

If memory serves- and it's 4am, so that trust is limited- the reasons 
for cvs2svn path were speed for our peculiar vcs history, and bugs in 
the cvs2git direct pathway.

This was long enough ago, it seriously wouldn't surprise me in the 
least if you've changed the playing field so that our original 
approach is no longer sane.

We're using your code one way or another, so advice is obviously 
welcome.


> > 3.a) Roughly, the plan will be snag the tree, start conversion.
> > Validate the results, repeat as necessary till we're happy with it.
> > This is the initial git core history,  This step should be <8h; mostly
> > cpu time, frankly, although re-validation of that pathway is required
> > (I did a fair amount of optimization to this, but I've not rechecked
> > the runtime in a while- nor if there is a better option in existence).
> > Basically, it's strongly preferable we're not sorting this at the time
> > we're trying to do the live conversion- the core issues need to be
> > sorted before.
>
> There is a program contrib/verify-cvs2svn.py that can be used for some
> verification of a conversion.  It tests that the contents of the tips of
> all branches (including master) and of all tags are the same in CVS as
> in the converted repository.  If the gentoo project tags the significant
> points in its history, then this is can give a lot of confidence in the
> conversion.  (Though, I must admit, I have little experience with this
> script and don't know how easy it is to use.)

We tag nothing, and branch nothing.  Just never fit our vcs usage 
patterns.


> > 3.b) Take all cvs activity that has occurred since the tree was
> > snapshotted and conversion started, and replay it into git via tailor;
> > this is minor- and avoidable if we just shut the tree down for however
> > long 3.a takes; that said, the tailor route is the intention, and
> > shouldn't be a problem.
>
> In my opinion, it is more prudent to shut down CVS during the
> conversion.  Otherwise, you effectively double the number of tools that
> you have to rely on and thus the number of possible points of failure.

Shutdown is my intent, and if the window is <4 hours, that's viable 
(frankly easier to just force it, regardless of whinging).

If we're talking more like 8-12 hours... well, that may be a different 
story.

Keep in mind that the the times quotes there, are just figured pulled 
out of the ass- if memory serves, we had it at something like 2 hours, 
which is a non issue.  If it's larger times, then the secondary tailor 
replay (which, note the validation of that fortunately isn't nearly as 
painful- can be done manually since it'll be <50 revs) isn't horrible.  
Best to keep in mind that a 4 hour maintenance window for something 
like this, on occasion gains a multiplier or two. :)

~harring