| 04:49 | <annevk> | tobie: yes, I can revive my old email thread with GitHub support and copy you all |
| 04:49 | <annevk> | tobie: maybe we can finally get it fixed then; will do once I'm behind my desktop |
| 05:57 | <hsivonen> | MikeSmith: hmm. maybe I've forgotten an hg add |
| 06:08 | <hsivonen> | MikeSmith: fixed: https://hg.mozilla.org/projects/htmlparser/rev/6a026284239e |
| 06:09 | <hsivonen> | MikeSmith: still no git version |
| 06:16 | <MikeSmith> | hsivonen: cool thanks |
| 06:33 | <MikeSmith> | hsivonen: so there’s a bug in template error-reporting that will need to fixed in the gecko sources too |
| 06:34 | <MikeSmith> | https://github.com/validator/htmlparser/commit/e6582f563751d2fd3f28d4e9c1b08c9f7494b372 |
| 06:34 | <MikeSmith> | I’ll raise a bmo bug for it for now |
| 06:35 | <MikeSmith> | and after that will make time later to write a gecko patch for it unless somebody else beats me to it |
| 07:23 | <annevk> | tobie: I copied you and Domenic on my GitHub support thread |
| 07:58 | <tobie> | annevk: ty |
| 07:58 | <tobie> | annevk: apparently zcorpan had similar issues yesterday |
| 07:58 | <annevk> | ah good |
| 07:58 | <annevk> | Maybe that means it'll finally get fixed |
| 07:59 | <annevk> | tobie: btw, when it works, do you have a better approach than cloning the whole repo? |
| 07:59 | <annevk> | tobie: jgraham once tried to teach me a trick of just adding that branch, but I never got that to work |
| 07:59 | <tobie> | annevk: sorry what? |
| 08:00 | <tobie> | annevk: oh, to "include" someone else's repo inside yours? |
| 08:00 | <tobie> | yes: |
| 08:01 | <annevk> | tobie: A forks repo R and creates a PR with branch B that the owner of R can push to because of this feature |
| 08:01 | <annevk> | tobie: does the owner of R need to clone the fork or can they do something with just B |
| 08:01 | <zcorpan> | annevk: I think whatwg/html checkout-pr script does this? |
| 08:05 | <annevk> | zcorpan: well, you also need to then push to B |
| 08:05 | <tobie> | https://www.irccloud.com/pastebin/BwBYvyNp/ |
| 08:05 | <annevk> | zcorpan: that's not possible with the checkout-pr thing afaik |
| 08:05 | <tobie> | annevk: ^ |
| 08:06 | <annevk> | tobie: interesting, I should maybe wrap that in a function and maybe document it in whatwg/meta |
| 08:06 | <zcorpan> | annevk: it is possible if the checkbox is checked |
| 08:07 | <tobie> | annevk: s/webidl/[REPO-NAME]/ obv. |
| 08:07 | <annevk> | zcorpan: oh yeah, https://github.com/whatwg/meta/blob/master/TEAM.md looks remarkably similar to what tobie pasted |
| 08:08 | <zcorpan> | annevk: what tobie wrote is essentially the same as checkout-pr... right :-) |
| 08:08 | <annevk> | I wish I had known this earlier |
| 08:08 | <annevk> | Although given how flaky the feature is maybe it's alright |
| 08:09 | <tobie> | annevk: what's wild is it seems to be the UI that's causing the flakiness. |
| 08:11 | <annevk> | tobie: yeah, would be nice if we could just outlaw that UI to begin with |
| 08:11 | <annevk> | tobie: "are you sure you want to PR this repo? they require to be able to push to your branch" |
| 08:17 | <tobie> | annevk: that's interesting. |
| 08:17 | <tobie> | annevk: sort of rubs me the wrong way, though. |
| 08:18 | <annevk> | tobie: collaborate or bust |
| 08:19 | <tobie> | annevk: the ownership model is inside out, though. |
| 08:19 | <annevk> | tobie: the other thing I think should be possible is for anyone to create a branch without a whole fork |
| 08:20 | <annevk> | tobie: might not layer well on top of Git, but for GitHub it'd make perfect sense |
| 08:21 | <annevk> | It'd be much easier to work on changes if you can just work on the main repo in your namespaced branches rather than keeping a fork up-to-date |
| 08:21 | <tobie> | annevk: totally agree. I can see branches emerging as some form of lighter weight thing people coalesce around. |
| 08:22 | <tobie> | annevk: well, namespaced-branches would require a different permission model for that to work |
| 09:36 | <hsivonen> | should never reboot a server except early in the morning on a day that can be used for dealing with it not actually booting cleanly... |
| 13:29 | wanderview | turns to IRC for entertainment while twitter is down... |
| 17:40 | <annevk> | rbyers: it would be good to know who in Blink have bz-level knowledge of navigation / document objects / browsing contexts/ etc. |
| 17:40 | <annevk> | rbyers: I want to create some kind of @whatwg/navigate team on GitHub to ping on these kind of issues |
| 17:41 | <annevk> | rbyers: going on vacation soon, but I plan to continue tackling it when I get back |
| 17:43 | <rbyers> | annevk: That's an easy question. Nobody has "bz-level knowledge" of this stuff ;-) |
| 17:44 | <rbyers> | .. but perhaps the question is, who are the people responsible for owning this space on blink (and continuing to build up towards bz-level knowledge). |
| 17:44 | <jyasskin> | rbyers: Navigation would maybe be one of the site-isolation folks? |
| 17:44 | <annevk> | No easy wins in navigation |
| 17:45 | <annevk> | jyasskin: yeah maybe, though it was somewhat suspect that the answer to what needs changing in the HTML Standard was "nothing" |
| 17:45 | <rbyers> | jyasskin: Yeah we certainly have a number of experts in various pieces |
| 17:45 | <annevk> | (I think I asked on Twitter, but it's down...) |
| 17:46 | <jyasskin> | annevk: They did a lot of work to make their changes invisible. But it's true that they're not spec folks. |
| 17:48 | <rbyers> | annevk: It would help me if you could collect a list of related spec issues which don't seem to have clear ownership from the Chrome team. We're rapidly filling in a lot of the gaps in our team structure, but still have a lot of work to do to ensure every API / algorithm is officially owned by some team. |
| 17:49 | <annevk> | There's quite a large series of interop issues undermining this quest of refactoring navigation (not even thinking about all the stuff folks are doing on top with service workers and such); such as: what is the scope of browsing context names, when do load events dispatch, how does about:blank work, when do beforeunload / unload dispatch, how do Document |
| 17:49 | <annevk> | objects inherit state, which WindowProxy objects can access each other, etc. |
| 17:49 | <annevk> | rbyers: ^^ |
| 17:50 | <annevk> | rbyers: https://github.com/whatwg/html/issues?q=is%3Aissue+is%3Aopen+label%3A%22topic%3A+navigation%22 has 50+ issues relating to this |
| 17:50 | <annevk> | rbyers: also things like what can you do and not do during beforeunload and unload and how is that governed |
| 17:51 | <wanderview> | basically all the parts of the spec/web people try to avoid |
| 17:51 | <annevk> | Gotta go relax a bit, hope that helps looking around |
| 17:51 | <annevk> | wanderview: hah |
| 17:52 | <rbyers> | annevk: That's really helpful, thank you! |
| 17:52 | wanderview | thinks annevk is asking who on the blink team rbyers doesn't like and would like to punish... |
| 17:53 | <rbyers> | FWIW some of our "site isolation" people (like lucasgadani@ who I see filed at least one of those bugs) are on my team and are currently moving to other projects. Perhaps there's something I can do within my own team here... |
| 17:54 | <rbyers> | wanderview: Hah, what do you mean? Working on interop on my team is one of the most glamorous, well-recognized, enjoyable things you can do ;-) |
| 17:54 | <wanderview> | thats good |
| 17:54 | <rbyers> | I'm half joking. But only half... ;-) |