On Tue, Mar 22, 2016 at 04:11:15PM +0000, tsbockman via Digitalmars-d wrote: > On Tuesday, 22 March 2016 at 15:33:23 UTC, PmLk wrote: > >On Tuesday, 22 March 2016 at 00:20:56 UTC, tsbockman wrote: > >>It's to the point where I feel kind of guilty about opening new pull > >>requests. > > > >You shouldn't. It's not your fault if the authors let their PR > >sleeping during 6 months without maintaining them. If you look at the > >tail of the queue, 60% of the PR don't even pass anymore. > > Yes, but it is also quite common for pull requests that *are* being > maintained to get stalled for weeks or months, simply because no one > with the right skills/merge rights can find the time to respond. > > It's not uncommon to find pull requests whose last few messages are > mostly just "pings" from the author, separated by weeks or even > months, trying to get someone to finally review/merge the thing.
I think the problem is that we're horribly short of manpower here. The thing is, Phobos is large -- *very* large -- and encompasses a pretty wide range of functionalities, and I, for one thing, don't feel qualified to review a lot of them. I'm reasonably confident to review range algorithms and the like, because I use them on a regular basis and so am reasonably familiar with how they ought to work. But for something like a new numerical algorithm, I have no idea where to even begin. Or things like std.xml, or std.json, that I never used, so I simply don't have the confidence that my review would do the PR justice. Or things like GC changes, std.regex engine hacks, that I don't feel confident to review because I simply don't have the time to dig into all the gory implementation details to know how to do a good job reviewing. And this is on top of the fact that I don't always have the time to sit down and go through a large changeset in detail, so when faced with a growing Phobos queue with many changes to unfamiliar modules, stuff I don't use (and thus don't really know how they *ought* to be used), and limited free time, I simply balk and just stick to relatively small PRs that can be reviewed quickly, that affect familiar modules or only involves relatively simple changes like doc improvements, etc.. This generally shouldn't be a problem if there are enough Phobos committers so that most of the areas are covered by somebody with expertise in that area, but the fact is that Phobos is too big for the current number of Phobos committers, and the committers we do have don't quite cover all of the Phobos modules, or even if we do, too many of us have only limited free time and may not get around to reviewing what needs to be reviewed within a reasonable timeframe. tl;dr: We need more Phobos reviewers, and more importantly, more committers. Many more, IMO. I've noticed recently that there has been an increase in the number of reviewers, which is a good sign, but not enough of them have been given commit rights. Which is understandable, since we don't want to just hand out commit rights to anybody who shows up -- but at the current rate we simply don't have the manpower to keep the PR queue to a reasonable size. T -- Error: Keyboard not attached. Press F1 to continue. -- Yoon Ha Lee, CONLANG
