[quagga-dev 8373] Re: [PATCH 7/7] ospf6d: Remove obsolete code

Greg Troxel gdt at ir.bbn.com
Tue Nov 30 13:11:37 GMT 2010


@balaji: Thanks for explaining this and keeping patches organized.

@quagga-dev:

I have indeed been trying to apply some patches.  The basic difficulty
is that I feel the existing codebase is fragile and underdocumented in
terms of calling conventions, invariants, etc.  Reviewing each patch to
be sure that it won't cause regressions is very tricky, and I've found
that I'm not comfortable yet with a significant number of them.

I did put one on a branch and asked for regression testing, but haven't
yet heard any reports.  It's on feature.bgp-vty-view-af.  If a few
people running BGP for real told me that they updated to that and it was
fine, I'd merge it master.

I will be trying to examine patches and for each either apply it,
feature branch it and call for testing, or explain on quagga-dev why I
didn't do one of those.

It's easiest for me if proposed changes are in a publically-accessible
git repository that's essentially a clone of the official repository,
with each logically-related group of changes on a feature branch off
master.  That means changes can be merged independently.

Balaji has been taking patches mailed to the list and storing them in
his git repository (see HACKING and HACKING.pending), which I
appreciate.  Patches sent to the list often don't have commit messages,
though, and those are very important.  If the only issue is not being
able to publish a git repository, then emailed patches with git
format-patch are fine, because those end up with commit messages after
git am.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 194 bytes
Desc: not available
URL: <http://lists.quagga.net/pipermail/quagga-dev/attachments/20101130/c00dbbc4/attachment-0001.sig>


More information about the Quagga-dev mailing list