On 04/23/2010 11:27 AM, Evan Schoenberg, M.D. wrote: >> I must confess to being a complete novice when it comes to integration of a >> new libpurple version into our code base, but seeing the relatively large >> numbers of patches which are necessary to fix security and compatibility >> issues in our existing version, I wonder if it wouldn't be better to pull >> the new version of the library and do both a 1.3.11 and a 1.4B18 release >> with it. (Of course, I defer to the experts on the wisdom of that path). > > Doing a major libpurple version upgrade on the 1.3 branch would be a large > undertaking; I definitely don't think it's worth doing. A minimal release > with the plucked changes to fix some protocol breakages seems reasonable, and > it should easy enough to do.
"Should" and "are" are two different things in this case. The Gadu-Gadu and Yahoo changes are much more in-depth than I expected and will take at least a couple hours for me, and I'm the idiot that put them in im.pidgin.pidgin and im.pidgin.adium.1-4 in the first place ;) These revisions don't pluck cleanly and lead to merge hell in the gg prpl (conflicts in every file) and will result in a broken build for Yahoo without additional post-merge changes. It might actually be easier in terms of actual work expended to just ship the same libpurple for 1.3 and 1.4. If it's still desired to backport these changes, I will have to defer this until the weekend, when I'll actually have time to take a look at it in-depth. I may be able to dissect the Gadu-Gadu changes into smaller changes to ease that burden, but it will almost certainly be easier (for me, anyway) to ignore monotone's pluck facility and just backport the Yahoo stuff the old-fashioned way. (This is because all the changes necessary are written for libpurple 2.7.0 and need backported to libpurple 2.5.9 for Adium 1.3.) John
signature.asc
Description: OpenPGP digital signature
