Create a global git commit view

Note that I've seen some ??s around this - do we really want to retain this functionality? Does it really tell us all that much?

Yes, the proper views integration is an absolute necessity, since there are other views that need to be generated of commit data.

Especially given that it's been a performance bottleneck in the past, I'm inclined to say we skip this one.

I defaulted to creating issues in the infra queue. There were a lot of issues to create, and given that the project can easily be changed later, I didn't feel like sweating it at the time :)

If it's just narshty queries in cvs module to fear, then that's not so bad. I was focused on the fact that it's basically worthless to cache this data. But if the query itself is cheap enough, I suppose that doesn't matter. I don't really know where the d.o infra comes down on caching/not caching commit log queries...though it seems like they're mostly uncached, now that I think about it.

Anyway, putting performance questions aside, is it really hugely useful?

Or is it more appropriate to just focus on replicating functionality 1:1 right now; maybe knocking this thing out is more a question for redesign folks...?

I don't know about you, but I find exposing data to views really a) easy and b) fun. I think we're already approaching more meta time talking about this issue than it would actually take to fix, unless the schema for vcapi is completely insane and twisted.

Tablewizard could spit out the field definitions pretty readily, I suspect, but the thing that's gonna make it a not-so-standard job is defining all the table relationships. For someone who isn't already pretty familiar with how views' base tables and joins (implicit and explicit) work, it could be a headache. Though you're probably right, I should stop spinning meta-wheels and just do it.

That said: for phase 2, I think we should replicate functionality 1:1 unless there are compelling reasons not to. I don't want to see phase 2 delayed while we try to figure out what functionality people are making the most use of, etc. We should just get back to where we are, and then we can plan all the other cool stuff we want to build on top of it in phase 3.

Cool, I was leaning in that direction as I wrote that. /me is on board with that.

And, there are no mythical "redesign folks". In general, there's no such group, and especially when it comes to stuff like this, there's definitely no "team" eagerly waiting to implement our stuff. The redesign didn't even touch the issue queues, barely scratched the project nodes (and the "party line" from Mark + Leisa was "you guys know this stuff much better than we do, we trust you to do what's best") -- they certainly didn't consider the details of our version control integration and development process...

I only brought that up because I know there's a lot of brouhaha in the DA right now about hiring someone to actually implement the redesign. But if the line was "you know your issue queues better than we do," then it's all moot, anyway!

User story: "Create listings of all commit activity in a site's repositories that can be viewed by site visitors."

The base view exists. Needs a lot of work, including some custom display handlers and such, but it exists. Good enough for this sprint :) Will open subsequent, more specific issues later.