[gentoo-scm] Re: Git Validation - Update

[Rich Freeman]

On Tue, Oct 16, 2012 at 1:59 PM, Rich Freeman <rich0@g.o> wrote:
> Once I have both I can start working on validation rules and perhaps
> get feedback to the conversion team.  We'll need to work out what does
> and doesn't count as OK.

Hmm, it didn't take long after working through the cvs side of things
to realize that my design will fail to detect file deletions, at least
in part.  I'm storing a history of when files change, starting by
looking at commits in complete isolation.  In cvs the deletion of a
file shows up as the existance of an entry that changes the file
state.  In git the deletion of a file is only marked by the absence of
the file in the tree - looking at the tree in isolation you'd never
know that there used to be a file there.

Right now I can't tell when a file was deleted, or even be sure it
ever was deleted.  Any commit info associated with deletion cannot be
checked.  If one of those commits happened to include modifications
(but not deletions) to other files then those would be checked.

Off the top of my head I can't think of a simple fix (other than
pairwise comparison the way that git show  $commit works - which would
be much more difficult to run in parallel).  One way that comes to
mind is to do a second pass that just looks for deletions, using the
cvs data to cheat - each deletion could be checked in parallel doing a
pairwise compare on a single file.  If I only wanted to check that the
file got deleted at about the right time I could probably tweak the
reduce function to keep around the hash for the last time a file was
present.  The author and message info would be worthless, but the
timestamp would be fairly close since we have so many commits.

Well, something to come back to once I get the cvs side done.  If
anybody has thoughts let me know...

Rich