RE: Is there really any interest in a patch to allow cvs 1.11.6 to run

[Kelly F. Hickel]

> -----Original Message-----
> From: Derek Robert Price [mailto:derek@ximbiot.com]
> Sent: Friday, October 03, 2003 10:48 AM
> To: Kelly F. Hickel
> Cc: bug-cvs@gnu.org
> Subject: Re: Is there really any interest in a patch to allow cvs
1.11.6
> to run
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Kelly F. Hickel wrote:
> 
> |FAIL: basica-6.2
> |make[2]: *** [localcheck] Error 1
> |make[2]: Leaving directory `/usr/local/cvs/cvs-1.11.6_nsk/src'
> |make[1]: *** [check-am] Error 2
> |make[1]: Leaving directory `/usr/local/cvs/cvs-1.11.6_nsk/src'
> |make: *** [check-recursive] Error 1
> |/usr/local/cvs/cvs-1.11.6_nsk:
> |
> |How much does this matter?
> 
> 
> If there are any sanity.sh tests failing, your fix is wrong because it
> breaks another part of CVS and the fix will not be accepted.
[Kelly Hickel] Well, The fix isn't wrong, because the unchanged code has
the same problem.  All the fix does is get around the platform
strangeness about size limited read/writes.  So, I haven't broken
anything new, but cvs is actually usable, instead of un-usable.
> 
> |   I still haven't gotten any good feeling
> |from someone who actually has the power to put my patches into the
> |source that they're interested in doing so....
> |
> |I'm about to upgrade the OS to the latest version, so I want to wait
> |any make sure that the unmodified source still has problems on the
> |latest version, then I'll be ready to send out the (relatively small)
> |patch.
> 
> 
> Per the HACKING in the top level of the CVS source distribution:
> 
> |The general rule for portability is that it is only worth including
> |portability cruft for systems on which people are actually testing
and
> |using new CVS releases.  Without testing, CVS will fail to be
portable
> |for any number of unanticipated reasons.
> 
> 
> This means that unless we see that Tandem gets regular testing, the
CVS
> developers are unlikely to feel that your patch is useful even if it
> begins to work since it will likely not stay useful for long.
> 
> i.e. If somebody is not running nightly testing on the system, then it
> is much easier for another developer to check in a fix for something
> else that breaks your portability cruft but which is not spotted until
> you tell it's broken after the next CVS release.
[Kelly Hickel] Well, it seems a little unlikely that anyone (except
maybe me) is going to loan you a $250,000 machine (smallest
configuration available) to run nightly tests on.  If you have a scheme
that lets remote people do tests for you and mails back the results we
can maybe work something out.
> 
> The exception to this rule can be when you can find a fix already
> incorporated in some GNULIB module we use (or get your fix
incorporated
> into GNULIB).  We hope to pick up portability cruft from GNULIB that
we
> might not actually be testing but which gets testing elsewhere.
[Kelly Hickel] Well, I guess I may try to find time to take a look at
GNULIB, I doubt it's fixed, but maybe it is.  This is the reason that
I'm still using cvs 1.10.7 (which doesn't pass it's tests either), it's
just too difficult to upgrade, because I have to re-port the world every
time.  Given the above comments, I'll probably just send the latest to
the ITUG folks, so that they can update their ported GNU tools
repository.

By the way, I didn't end up changing any source files other than subr.c,
if that helps (I added two methods).  Again, (if you didn't see the
earlier posts) I basically just added a safe_read and safe_write, which
some of the comments in client.c seem to be asking for.

Still, I suppose it's better to have a "no" (which is how I interpret
your response), to having no answer at all.

Thanks,
   Kelly
> 
> Derek
> - --
> ~                *8^)
> 
> Email: derek@ximbiot.com
> 
> Get CVS support at <http://ximbiot.com>!
> - --
> I will not buy a presidential pardon.
> I will not buy a presidential pardon.
> I will not buy a presidential pardon...
> 
>     - Bart Simpson on chalkboard, _The Simpsons_
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.0.7 (GNU/Linux)
> Comment: Using GnuPG with Netscape - http://enigmail.mozdev.org
> 
> iD8DBQE/fZoULD1OTBfyMaQRAp6CAKDPO2BR0Q7bJX/w+siMqPeyu0ZZ0gCeKX4+
> OAOXilw3oKR90T8enCyo834=
> =+yK4
> -----END PGP SIGNATURE-----
>