On Fri, Feb 6, 2009 at 16:04, Basile STARYNKEVITCH <bas...@starynkevitch.net> wrote:
> But given that most people agree that (since the plugin idea is not too > ambitious) the plugin effort is not a big one (e.g. much much smaller than > your heroic old-gimple => tuple work) why did'nt you consider accepting > plugin related stuff into trunk? Because we are not in stage 1. Unless a patch is very small, carrying it around is a hassle, particularly when the group of folks interested in testing it out is large as in this case. > Or is the current stage of the trunk still in a phase where plugin related > stuff are not accepted? Right. > My point is that by creating a plugin branch, we have to duplicate the > review effort, since every thing has to be reviewed twice (once to be > admittted into the plugin branch; once to be transported from plugin branch > to trunk). And review time (& reviewers effort) is the most scare resource > in GCC. There will be additional review when the merge is proposed, but given the review during the lifetime of the branch, it should not take too long. I will propose merging the branch into mainline during the next stage 1, if the branch is ready (we can decide what 'ready' means later). Diego.