Re: a few Q's

[Christos Zoulas]

On Oct 22,  5:36pm, mdb@cvshome.org ("Mark D. Baushke") wrote:
-- Subject: Re: a few Q's

| > 1. When is the next 1.12.x version scheduled for? I would like to import a
| >    1.12.x flavor in the NetBSD tree and 1.12.1 is quite old.
| 
| Derek needs to answer this question.
| 
| > 2. I am maintaining cvs for the NetBSD project and I have quite a few pat=
| ches
| >    accumulated. Now the patches have gotten quite large, so it is a pain =
| for
| >    me to maintain them. Last time I submitted one of them, I was told that
| >    CVS was not actively developed, and there were too many bugs in handli=
| ng
| >    symlinks, so the patch was not accepted. Is this still the case?
| 
| Patches may be sent to bug-cvs@gnu.org

I will do that as soon as I merge things. Do you want one giant patch? Or
split by functionality?

| > Here's the list of patches:
| >     1. make client server cvs work when the remote cvs server path
| >        contains a symlink.
| 
| This is fixed in the cvs 1.12.1.1 (top-of-tree).

I will take a look.

| >     2. add -u [unlocked] flag to do a checkout without locking.
| 
| I think using -R might do this for you...

I'll check into that too.

| >     3. make CVS work as a setuid [not setgid] process when SETXID_SUPPORT
| >        is defined.
| 
| I'd love to see this patch.
| 
| >     4. make CVS work with IPV6
| 
| This would also be good.
| 
| >     5. make CVS work with krb/gssapi auth
| 
| Okay... I am someone biased against auth stuff in the current CVS, but I
| understand the need. I probably would not be helping to commit this one.

It is all ifdef'ed properly, so it is not a big deal.

| >     6. add -D dirname flag, to specify an alternate directory name for
| >        cvs to store its administration files.
| 
| I am not sure I understand this one.

Imagine you are working on a case insensive system and you have a directory
called cvs. You can change -D to make the default directory .CVS for example
to avoid the conflict.

| 
| >     7. use asprintf() instead of malloc()/sprintf() pairs.
| 
| Yes, I would like to see this one.

Ok.

| > Please let me know if you are interested, and I'll send the patches.
| 
| If you want to work with top-of-tree and give us patches, it might even
| be possible to get some of these changes into the tree before 1.12.2 is
| created.

That would be wonderful, thanks!

christos