The purpose in creating and maintaining a client branch (used by TBird
and SeaMonkey) is to provide stability while also allowing increased
freedom on the LDAP C SDK trunk. The customers for the client branch
are the teams that use it in their products, so what they want is very
important. We don't want to repeat the bad past where no one was
motiviated to keep the client branch up to date though.
I am not exactly sure how things work with NSPR and NSS, but my
impression is that it is similar to what Mark Banner described (cut a
new branch or merge when a component is released and pick up critical
fixes whenever necessary).
-Mark
Anton Bobrov wrote:
Mark Banner wrote:
My general thoughts are that the client branches should be kept
stable, maybe being updated every time there is a c-sdk release. I
think that patches should only be applied to a client branch when they
are needed for bustage or major bug fixes, and as per Mozilla tree
rules, they will need the sr of a developer (typically Dan).
personally i would second this approach. right now we have bunch
of Sun changes to merge and some room to breathe in terms of srs
etc would be preferable to cut on time required. the drawback is
maintenance of course but if it becomes too much of a burden we
can always revisit it later. i would be also interested to know
how it works with other components like NSS/NSPR. maybe we can
follow the same, already established model to simplify things ?!
_______________________________________________
dev-tech-ldap mailing list
[email protected]
https://lists.mozilla.org/listinfo/dev-tech-ldap
