<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">On Wed, May 23, 2018 at 4:41 PM, Jordan Justen <span dir="ltr"><<a href="mailto:jordan.l.justen@intel.com" target="_blank">jordan.l.justen@intel.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class="">On 2018-05-23 16:00:53, Jason Ekstrand wrote:<br>
> On Wed, May 23, 2018 at 2:48 PM, Jordan Justen <<a href="mailto:jordan.l.justen@intel.com">jordan.l.justen@intel.com</a>><br>
> wrote:<br>
> <br>
> > On 2018-05-23 12:34:11, Jason Ekstrand wrote:<br>
> > > Mesa developers,<br>
> > ><br>
> > > tl;dr. Please go to <a href="http://gitlab.freedesktop.org" rel="noreferrer" target="_blank">gitlab.freedesktop.org</a>, create your account, and<br>
> > > upload your SSH keys. Instructions are the bottom of this e-mail.<br>
> > ><br>
> > > The <a href="http://freedesktop.org" rel="noreferrer" target="_blank">freedesktop.org</a> admins are trying to move as many projects and<br>
> > services<br>
> > > as possible over to gitlab and somehow I got hoodwinked into<br>
> > spear-heading<br>
> > > it for mesa.<br>
> ><br>
> > I thought we'd be able to decide if or when. It sounds like this is<br>
> > decided.<br>
> ><br>
> <br>
> You are free to state your objections. No one has actually pulled the<br>
> trigger yet. :-)<br>
> <br>
<br>
</span>Rather than "it's time to do this", how about "should we do this"?<br></blockquote><div><br></div><div>Fair enough. :-) How about, "We should do this some day and there are people who are almost ready with a fancy new website that's blocking on it so sooner would be better than later"?<br></div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
FWIW, if the fdo admins think this hosting is viable long term and at<br>
least as stable as the current servers, then I don't see a concern<br>
with changing *only* the repo location.<br>
<span class=""><br>
> <br>
> > > There are a number of reasons for this change. Some of those<br>
> > > reasons have to do with the maintenance cost of our sprawling and aging<br>
> > > infrastructure. Some of those reasons provide significant benefit to the<br>
> > > project being migrated:<br>
> > ><br>
> > > * Project-led user/rights management. If we're on gitlab, project<br>
> > > maintainers can give people access and we no longer have to wait for the<br>
> > > freedesktop admins to add a new person. And the freedesktop admins don't<br>
> > > have to take the time.<br>
> > ><br>
> > > * Better web UI for git.<br>
> ><br>
> > s/Better/Worse/<br>
> ><br>
> > > Ok, so some people will argue with me on this<br>
> > > one but it's at least how I feel. :-)<br>
> ><br>
> > /me bows<br>
> ><br>
> > > * [Optional] Integrated commit history and issue tracking. Bugzilla<br>
> > tags<br>
> > > are great but gitlab ties things together much better.<br>
> > ><br>
> > > * [Optional] Merge-request workflow. With the rise of github, there are<br>
> > > many developers out there who are used to the merge-request workflow and<br>
> > > switching to that may lower the barrier to entry for new contributors.<br>
> ><br>
> > On the down side, the contribution/review history is now tied up in<br>
> > the gitlab infrastructure rather than a simple email archive.<br>
> ><br>
> <br>
> Yup, and that is a down side. That's why I made it clear that this is<br>
> optional and not going to be a thing we do right away.<br>
<br>
</span>Ok, why don't you save it for a future refactor? :)<br></blockquote><div><br></div><div>That's the plan.<br></div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
We should say:<br>
<br>
"Gitlab has many other features, but they won't be used at this time.<br>
Perhaps in the future we can consider using some of them after<br>
further discussion."<br>
<br>
As opposed to a sales pitch for all of the things we won't be using.<br>
:)<span class="HOEnZb"><font color="#888888"><br></font></span></blockquote><div><br></div><div>Some of those features we will be using, just not all. Being able to add new contributors without the fd.o admins seems immediately useful. I tried very hard to talk gitlab up while still making it clear that we don't actually have to really change anything other than the push remote. Apparently, I failed. :-)<br></div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class="HOEnZb"><font color="#888888">
-Jordan<br>
</font></span><div class="HOEnZb"><div class="h5"><br>
> <br>
> > > * [Optional] Built-in wiki support<br>
> ><br>
> > The github (and gitlab) integrated wiki's look fairly busy/crappy with<br>
> > all the gibhub/gitlab UI around the contents.<br>
> ><br>
> <br>
> Yeah, I don't expect we'll actually use it for real things but it's there<br>
> if we want it. Might be useful for piglit to keep some notes or something.<br>
> <br>
> <br>
> > > * [Optional] Built-in CI. With gitlab, we can provide a docker image<br>
> > and<br>
> > > CI tasks to run in it which can do things such as build the website, run<br>
> > > build-tests, etc. I'm not sure if build-testing Android is feasible but<br>
> > we<br>
> > > could at least build-test autotools, meson, scons, and maybe even run<br>
> > some<br>
> > > LLVMpipe tests.<br>
> ><br>
> > Why can this only be done with gitlab?<br>
> ><br>
> <br>
> It *can* be done other ways but the fd.o admins are tired of everyone<br>
> wanting some hand-rolled solution for this that or the other thing. With<br>
> the gitlab CI, we provide a docker image and a script and it "just works".<br>
> There is literally zero overhead for the fd.o admins to maintain the thing.<br>
> <br>
> <br>
> > > Before anyone freaks out about the possible changes that may be<br>
> > incoming, I<br>
> > > would like to make it crystal clear that many of the above things are<br>
> > > optional.<br>
> ><br>
> > Yeah, this doesn't feel like a steamroller at all.<br>
> ><br>
> <br>
> I'm really not trying to steamroller. :-( Like I said below, we don't need<br>
> to use any of those optional features and we'll turn them off initially so<br>
> that it's not even possible to make a PR or file an issue and the wiki<br>
> won't even exist. Once you add your key and change your push URL, you<br>
> won't even notice that it's hosted on gitlab.<br>
> <br>
> <br>
> > -Jordan<br>
> ><br>
> > > We can continue to use Bugzilla for issue tracking and the<br>
> > > mailing list for patch review. Both cgit and annongit will continue to<br>
> > > work for the foreseeable future. The new fancy features such as merge<br>
> > > requests will all be disabled initially and we can consider enabling and<br>
> > > using those features on a case-by-case basis. The only immediate change<br>
> > > will be that pushes will have to happen to gitlab instead of git.fd.o.<br>
> > No<br>
> > > one is trying to change your workflow, they're just trying to move our<br>
> > git<br>
> > > hosting to a different platform.<br>
> > ><br>
> > > One of the motivations for doing this now is that there has been some<br>
> > > desire to move the mesa website away from raw HTML and over to a platform<br>
> > > such as sphinx. If we're going to do that, we need a system for building<br>
> > > the website whenever someone pushes to mesa. The solution that the fd.o<br>
> > > admins would like us to use for that is the docker-based gitlab CI.<br>
> > Laura<br>
> > > has been working on this the last couple of weeks and the results are<br>
> > > pretty nice looking so far. You can check out a preview here:<br>
> > > <a href="https://mesa-test.freedesktop.org/intro.html" rel="noreferrer" target="_blank">https://mesa-test.freedesktop.<wbr>org/intro.html</a> Using sphinx gives us all<br>
> > > sorts of neat things like nice text formatting, syntax highlighting, and<br>
> > > autogenerated searchable navigation. Right now, it's still using one of<br>
> > the<br>
> > > standard sphinx themes so it looks a bit "default" but that's something<br>
> > we<br>
> > > can change.<br>
> > ><br>
> > > Making this transition happen will, obviously, require a small amount of<br>
> > > involvement from the mesa development community. In particular, you'll<br>
> > all<br>
> > > need to get your SSH keys set up through gitlab. Here's what you need to<br>
> > > do; it should take less than 5 minutes:<br>
> > ><br>
> > > 1. Go to <a href="http://gitlab.freedesktop.org" rel="noreferrer" target="_blank">gitlab.freedesktop.org</a><br>
> > > 2. Click "Sign In / Register" in the upper left-hand corner<br>
> > > 3. You already have an account. Click "Forgot your password?", type in<br>
> > > your fd.o-associated e-mail, and click "Reset Password". Follow the<br>
> > > directions in the e-mail.<br>
> > > 4. Once you've successfully signed in, click on the little circle in the<br>
> > > upper right-hand corner and select "Settings"<br>
> > > 5. Click "SSH Keys" in the bar on the left and add your keys<br>
> > ><br>
> > > Assuming no one explodes too badly, we'll do the actual migration soon.<br>
> > > Ideally, I'd like to not drag this out for more than a couple of weeks.<br>
> > > When the actual migration happens, the only change mesa devs will have to<br>
> > > make when this happens is to change the git remote they use for pushing<br>
> > to<br>
> > > point to gitlab.<br>
> > ><br>
> > > Thanks for your cooperation (was that premature?),<br>
> > ><br>
> > > --Jason<br>
> > ><br>
> > ><br>
> > > ______________________________<wbr>_________________<br>
> > > mesa-dev mailing list<br>
> > > <a href="mailto:mesa-dev@lists.freedesktop.org">mesa-dev@lists.freedesktop.org</a><br>
> > > <a href="https://lists.freedesktop.org/mailman/listinfo/mesa-dev" rel="noreferrer" target="_blank">https://lists.freedesktop.org/<wbr>mailman/listinfo/mesa-dev</a><br>
> ><br>
</div></div></blockquote></div><br></div></div>