Re: [Gnu-arch-users] Re: cscvs--experimental--1.1 nearing doneness; call for testers

[Charles Duffy]

On Mon, 2003-09-29 at 14:54, John Goerzen wrote:
> Now the bad news.  cscvs starts by generating at least one -- and
> perhaps two -- changesets for the initial add of each file.  That is,
> it starts out wanting to commit -- get this -- 30,000 changesets to
> tla just to add the files from the initial cvs import.  That's not
> acceptable at all.

Okay, I've got that one beat. Creation of the kernel repository still
takes 4 changesets (and almost everything is in #4) -- but that's
clearly bkcvs's fault, and I don't see any particular reason to work
around it.

There are also some cases where bkcvs intentionally gives multiple
revisions within one changeset with different names (ie. deleting files;
"ChangeSet" revisions) and does other weird stuff. If I had
bkcvs-specific support I could handle these situations nicely --
something for the TODO list, maybe, to be bumped up to a higher priority
upon request.

BTW, I fixed a few general-purpose high-severity bugs that bkcvs's
oddities exposed in addition to the big obvious one. Many thanks for
bringing this to my attention.

(FYI, the 2.4 kernel now comes out to 12254 changesets).