We should make sure all top-level directories have at least one OWNER:
pnin:wpt2 gsnedders$ find . ./css -mindepth 1 -maxdepth 1 -type d '!' -exec test -e "{}/OWNERS" ';' -print
Round 2:
wpt-pr-bot now adds a comments whenever there's a pr that has no OWNERS files.
Problem is the process we have to handle this at this point is to send the OP on irc. Seems subpar. Maybe suggest something else here and modify the comment wpt-pr-bot makes accordingly?
IMO, we should try and seek out people for all of these directories. I think most should be relatively easy?
@annevk http?
@halindrome annotation-protocol?
@miketaylr @annevk compat?
@fred-wang mathml?
@fred-wang mathml?
Yes, I can take ownership for that one. Thanks.
I can take annotation-protocol
On Aug 1, 2017 7:41 AM, "Geoffrey Sneddon" notifications@github.com wrote:
@fred-wang https://github.com/fred-wang done in b0c089d
https://github.com/w3c/web-platform-tests/commit/b0c089d479a61bc3d486dc880438782af2665917—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/w3c/web-platform-tests/issues/5915#issuecomment-319390945,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AAfx8JhH12J5P4Drr3ityaGgw27YY_G6ks5sTzkWgaJpZM4NaPsf
.
Maybe @mnot wants to own http/? I'm okay with looking at PRs though.
Sure.
@gsnedders, which are still missing? I feel like I've added rather many in the past months.
@foolip updated
Thanks! Lots of those have OWNERS files in Chromium, so probably >50% of them could be fixed by copying those, converting to GitHub user names and sending out lots of reviews for people to find additional co-owners.
I updated the first comment with the current state. I've created PRs for a few CSS directories, getting names from the spec's editors.
I can take acid and js
annevk on IRC
<annevk> zcorpan: seems innerText should mimic html
<annevk> zcorpan: x-frame-options should mimic csp
<annevk> zcorpan: another way of solving that might be to move some of those directories
I think Tab and I would be happy to hand off ownership to anyone who wants it, but in the meantime it's probably fine to have us added to unowned modules we edit. (I can try to be more responsive wrt reviews going forward, but tbh I've been shit at responding to review requests in general.)
@grorg @junov thoughts on how to deal with all the canvas tests? Should we try to move those into html/ as they are part of that specification or is keeping them in separate top-level directories preferable for some reason? Any of you willing to review incoming PRs?
@annevk I agree these should not be top level. I volunteer to review PRs. There's a bunch of canvas tests in html/semantics/embedded-content/the-canvas-element, which IMHO should be closer to the other canvas tests in the directory structure.
There are now PRs for everything. If somebody feels they're missing or don't want to be listed, please comment in the PR. If you are added and approve, please mark as approved using GitHub's review feature.
@zcorpan Way to go! Do you think we can lint for this once all those patches land?
I intend to go over the PRs tomorrow and remove non-responders and merge.
These are still not merged:
Running the command in the initial issue gives:
./keyboard-map
./x-frame-options
./.well-known
./css
./async-local-storage
./css/css-filter
./css/vendor-imports
I think css and probably css/vendor-imports are the only things that shouldn't have OWNERS files.
.well-known should have OWNERS in subdirs.
I think we have everything now. @gsnedders can you confirm?