This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: The pain of committing ChangeLogs...
- To: Neil Booth <neil at daikokuya dot demon dot co dot uk>
- Subject: Re: The pain of committing ChangeLogs...
- From: "Joseph S. Myers" <jsm28 at cam dot ac dot uk>
- Date: Thu, 11 Jan 2001 00:21:31 +0000 (GMT)
- cc: <gcc at gcc dot gnu dot org>, Gerald Pfeifer <pfeifer at dbai dot tuwien dot ac dot at>
On Wed, 10 Jan 2001, Neil Booth wrote:
> I suspect a major cause of this is the ChangeLog file, and its
> thousands of revisions causing considerable server load.
For normal update/commit operations on the mainline, this shouldn't be an
issue - since the RCS file format stores the most recent revision and
backwards diffs to go to older revisions.
However, when committing, at least one whole copy of the ChangeLog needs
to go across the link, since CVS doesn't store the unmodified files
locally in a checked out tree for comparison with the modified files.
Try reading the CVS protocol specification (cvsclient.texi) and running
strace on a CVS connection (without -z so the uncompressed protocol data
can be seen).
However, ChangeLog.3 covers a 6-month period from January to June 2000; it
is time to rotate at least gcc/ChangeLog again (to ChangeLog.4 covering
the rest of 2000).
--
Joseph S. Myers
jsm28@cam.ac.uk