Hi. On Tue, 22 Apr 2025 18:33:43 +0200 Willy Tarreau <w...@1wt.eu> wrote:
> Hi, > > I've backported a few bits to 2.2 which had been accumulating low- > importance fixes over the last year without a release (and none of > them is really important as you can guess, most of the 2.4 ones are > not relevant there). This one is going to die in about a month when > 3.2 is out. > > I strongly doubt anyone on 2.2 continues to blindly jump on each new > version as they come. My guess is that either they've been picking up > fixes from the maintenance branch as they landed in it or they've just > being ignoring it. In any case, deploying updates for old versions > requires careful validation since bugs are features sometimes... > > Thus my question is simple: should I produce another 2.2 (2.2.34?) or > should I just leave it like this as a cumulative fixes branch ? I must > confess I'm seriously wondering if we shouldn't do that by default for > critical fixes branches, because tagged versions don't mean anything > when they're only produced when someone has enough time to go through > all patches to write an announce message. But if someone is really > relying on these tags, I'll go through that task. I just want to be > sure I'm not doing it only for the glory... I would even consider to stop at 2.8 for new versions/releases. From my point of view are 2 LTS releases more then enough to maintain. What's the reason to keep the older versions updated? > Thanks! > Willy Regards Alex
