bug-cvs
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Log bug in CVS 1.12.10?


From: Thomas Singer
Subject: Re: Log bug in CVS 1.12.10?
Date: Tue, 23 Nov 2004 17:43:21 +0100
User-agent: Mozilla Thunderbird 0.8 (Windows/20040913)

Hi Mark,

I would rather you didn't send me bug reports directly to me and instead
posted such to the bug-cvs@gnu.org list.

Well, I had the fear, that it might get lost in all the spam.

Hmmm... cvs 1.12.10 is using zip 1.2.1 up from zip 1.1.4, and I suppose
there could be some kind of a problem with compression levels.

If possible, could the user try to do the same 'cvs log' command on the
server with the cvs 1.12.10 command? If it really is some kind of
corruption in the protocol, it would be nice if it were reproduceable
with the 1.12.10 client/server combination too...

You are right. Using no compression works. If I recall correctly, CVSNT had a similar bug at the beginning of 2004 and Tony Hoyle fixed it somehow. It looks like zlib is broken. At our side, we cannot change it, because SmartCVS uses Java's native compression methods, which have proven to work without problems.

--
Cheers,
Tom


Mark D. Baushke wrote:
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Thomas,

I would rather you didn't send me bug reports directly to me and instead
posted such to the bug-cvs@gnu.org list. (Yes, I know that you gave me a
SmartCVS Professional License, but to be honest, I have not had time to
try to download it and use it myself.)

Looking over the report...

So, you are saying that the final 'ok' is never received?

Hmmm... cvs 1.12.10 is using zip 1.2.1 up from zip 1.1.4, and I suppose
there could be some kind of a problem with compression levels.

If possible, could the user try to do the same 'cvs log' command on the
server with the cvs 1.12.10 command? If it really is some kind of
corruption in the protocol, it would be nice if it were reproduceable
with the 1.12.10 client/server combination too...

Suggest that the user may be able to work around the problem with a
different compression level like -z3 (or even dropping compression all
together) instead of -z6 and if that fixes the problem, it would be good
to know the OS of the server.
If that does work, then it is also possible that using an external
- --with-external-zlib may also fix the problem (it presumes that their OS
vendor has a better zlib than what comes with cvs 1.12.10).

It would also be good to know if there may be a problem with smartcvs
decrypting the gzipped data stream.

If you can reproduce the problem using a ,v file from an opensource
project, that would be great.

Suggested action:

  a) get the Operating System of the cvs 1.12.10 system
  b) submit the bug report to bug-cvs@gnu.org
  c) If possible, include enough data to allow us to reproduce
     the bug with a ,v file that demonstrates the problem.

        Thanks,
        -- Mark
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (FreeBSD)

iD8DBQFBo1cQ3x41pRYZE/gRAh0dAKCaSVmTSry/o6XV63kAkhYAWHY2mwCdEEH6
KVCxGlatZZkNCppQJChjc2s=
=8AJ/
-----END PGP SIGNATURE-----






reply via email to

[Prev in Thread] Current Thread [Next in Thread]