Re: One more little bug with "file X was added on branch Y" revisions

On Mon, Sep 14, 2009 at 3:55 PM, Michael Haggerty <[email protected]> wrote:
> Hmmm, the cvs2svn source code suggests that files added on a branch
> should have a log message that matches the following regexp:
>
>    r'file .* was initially added on branch .*\.\n$'
>
> The current CVS source code agrees, as does the CVS source code back to
> version 1.1 of the file commit.c.
>
> Your example files' log messages are missing the word "initially" and
> the trailing ".".

Indeed they are.

I really was the one who added the file in question back in 2005, and
I can honestly swear on a stack of Unix man pages [1] that I have
never built CVS, patched CVS, or even looked at the CVS source code.
We have been loyal users of Fedora/Red Hat Linux for many years here,
and I for one have only ever used Red Hat's build of CVS on my various
workstations.  I suppose it's always possible that some mischievous
hacker at Red Hat patched CVS to introduce this change, but I haven't
been keeping track.

Regardless: I am 99.999% positive that the "file X was added on branch
Y" was generated by a clean, kosher, non-weird CVS version by a
non-malicious, clued-in developer (me).

> If not, I'd rather leave things the way that they are.  (I want to be
> very conservative about discarding revisions.)  In this case, I suggest
> that you hand-edit the three log messages to match the expected pattern;
> then they will be omitted by cvs2svn.

...or patch the regex in cvs2svn to conform to my version of reality.
;-)  I'll do that and let you know how it works.

(Incidentally, another pretty sure sign that you're looking at a
discardable revision is that the "text" part of the RCS file is empty.
 Is there any feasible way to check that in
remove_unneeded_initial_trunk_delete() and/or
remove_initial_branch_deletes()?)

Greg

[1] really, who uses bibles anymore?

------------------------------------------------------
http://cvs2svn.tigris.org/ds/viewMessage.do?dsForumId=1667&dsMessageId=2395209

To unsubscribe from this discussion, e-mail: [[email protected]].