On 5/30/2015 1:47 PM, Daniel Ruggeri wrote:> Thinking about this more, what are the things preventing people from an > _easy_ upgrade path configuration-wise? A lot of this conversation > surrounded users and the impact of an upgrade to them. The interface for > the users' to the server is the configuration file. Maybe if we can > tackle that we can greatly reduce a barrier to upgrade (or maybe I'm too > optimistic)? > > For the majority of my configs, it was the changes to the authorization > directives - it takes brainpower to figure out what AdminXYZ was trying > to do years ago and reflect that with the new directives. However, this > is deterministic... a perl script could do this work for me if I'd just > write it. > > At $dayjob, this is the stuff I focus on. Tweaks to an existing script I > put together years ago to upgrade from 2.0 to 2.2 (or as Rich eloquently > put, "poop out" new configs) only required an hour's worth of work or so > to support upgrades from 2.2 to 2.4 minus the aforementioned authz > directives. > > -- Daniel Ruggeri
....aaaaaand on the original topic at hand, I'll share that none of us are really *forced* to focus on any particular branch or any particular area of the code base. As an army of volunteers, we scratch our own itch. I, too, have noticed that 2.2 hasn't had a whole lot going on lately. I don't know if that means we ought to get ready to drop the axe on it, but I also don't know if that means we shouldn't be thinking about it, either. I think Jim's email served it's purpose of at least vocalizing a thought that's probably in the back of all of our minds: 2.2 isn't getting a ton of love (or at least as much love as 2.4) lately. What that means to each of us is clearly different... for me, I noticed that the effort to backport some of the work I've done in 2.4 doesn't pay off so I don't do it. I haven't put much more thought into 2.2 than that. For others it might be a push to backport more stuff. At the same time, Bill points out a reality we're faced with. As the most widely deployed HTTP server, we have some sort of responsibility (whether real or imagined) to the users not to "cut them off" if we can avoid it. The point about distros taking their sweet time to update are well taken - it's one of the reasons I always build from latest source. Maybe there's a reason other than inertia for the slow adoption. Their own lack of volunteer cycles? Configuration differences? Performance differences? What we have works, so why change it? Do we know those reasons enough to try to keep ahead of them? P.S. I'm not a Member or PMC... do I have access to the report that spurred the conversation? P.P.S. Wow... I don't read my email for a few days... -- Daniel Ruggeri