03:11
<Domenic>
<3 cvrebert
04:16
<MikeSmith>
The Prolific Mr. Rebert
04:17
<MikeSmith>
I'm sad in part because I first came to know cvrebert from PRs and issues he raised against the HTML Checker
04:17
<MikeSmith>
from his work on his extremely-clever https://github.com/cvrebert/lmvtfy tool
04:18
<MikeSmith>
but he raiseth no HTML Checker PRs or issues lately
04:18
<MikeSmith>
so I feel a bit neglected 😢
04:19
<MikeSmith>
s/raiseth no/raiseth nigh a/
04:57
<Domenic>
one of the best s//s I've seen in recent memory
05:00
<annevk>
So my thinking with https://html5.org/tools/web-apps-tracker is that we redirect if the URL is exactly that to https://github.com/whatwg/html/commits
05:00
<annevk>
Coupled with the Twitter account also pointing there I think that's sufficient
05:00
<annevk>
If anyone wants to write a script that redirects links such as https://html5.org/r/8891 as well be my guest
05:01
<annevk>
Domenic: so one concern with our commit conventions is that the tweets end up rather big
05:01
<Domenic>
annevk: they seem fine to me?
05:01
<annevk>
Domenic: you don't mind the double newline?
05:01
<annevk>
Takes up so much space
05:01
<Domenic>
Not at all... great to separate "this is definitely worth reading" from "the rest of this is details that will probably be cut off anyway"
05:02
<Domenic>
We could try to (get someone with knowledge of Ruby to) add an option to the twitter thingy to cut off anything after the double-newline if you want
05:03
<annevk>
That might be of interest, usually the links placed there are wrong anyway
05:03
<annevk>
I would prefer the tweets to be a bit less obtrusive
05:04
<annevk>
And that would certainly make them so
05:05
<Domenic>
The double newlines are actually because of a recent instance of me getting someone to fix a bug that omitted them
05:05
<Domenic>
https://github.com/github/github-services/issues/959 https://github.com/github/github-services/pull/1075
05:05
<Domenic>
I wanted the separation for streams where I do the same thing
05:06
<Domenic>
Through the latter link you can also learn about a bizarre behavior of Ruby's string.split(" ")
05:24
<MikeSmith>
about the idea of redirecting links such as https://html5.org/r/8891 it seems not worth the hassle
05:24
<MikeSmith>
because "ain't broke don't fix it"
05:27
<MikeSmith>
in other news by my count, 50 of the open W3C bugzilla bugs were closed out yesterday (out of the 428 open bugs the day started out with)
05:27
<MikeSmith>
but unfortunately it seems pretty much all the low-hanging fruit has now been picked there
05:28
<MikeSmith>
and the remaining open bugs are all mostly things that will actually require some work
05:31
<MikeSmith>
I guess I could try to start a TODO list of stuff that needs to be done, if that would be worthwhile
05:32
<MikeSmith>
my own personal TODO list is to actually finish getting my build environment set up and working
05:33
<MikeSmith>
but first, time to take a walk to Shinjuku station and eat a hamburger
05:36
<annevk>
Domenic: hah, it sounds like you're my Twitter hook enemy :-P
05:37
<annevk>
MikeSmith: jaleous
05:38
<annevk>
MikeSmith: I think a high-level TODO thing might be worthwhile since a lot of bugs are related
05:38
<annevk>
MikeSmith: also, there's a bunch of easy bugs left but they're all filed by philipj so I don't want to start fixing them for him
05:53
<annevk>
TabAtkins: is https://drafts.csswg.org/mediaqueries/#scripting being implemented?
05:54
<TabAtkins>
annevk: It will be, yes. Approved as part of Houdini.
05:56
<annevk>
TabAtkins: so https://github.com/whatwg/html/pull/25/files makes sense?
05:56
<TabAtkins>
Oh, wrong feature, that's not a Houdini thing. Still yeah, should be implemented, and yes, makes sense at first glance.
05:57
<TabAtkins>
You can simplify that to just (scripting) btw
05:59
<annevk>
Ah shit
06:03
<annevk>
Okay, addressed in a follow up commit
06:03
<annevk>
Thank you
06:03
<annevk>
MikeSmith: Domenic: btw, I'm not exactly locally building the specification
06:03
<annevk>
MikeSmith: Domenic: currently relying on making simple changes, but I guess I should change that...
06:08
<annevk>
Loading https://github.com/whatwg/html/graphs/contributors seems impossible
06:27
<cvrebert>
MikeSmith: Perhaps all the issues+PRs on https://github.com/w3c/css-validator-fork should be closed since it's dead?
06:42
<MikeSmith>
cvrebert: Yeah, will do so later
06:46
<MikeSmith>
actually I think I may just transfer ownership of it to Alex. Because it's his fork really (never actually been a w3c thing), and nobody other than him actually cares at all about it
06:46
<MikeSmith>
fuglied up as it is with scala/sbt junk
07:22
<cvrebert>
Am I missing any straightforward reason why these two "tables" shouldn't render the same?: http://jsfiddle.net/cvrebert/2xk62L2x/
07:25
<TabAtkins>
cvrebert: <fieldset> has magical rendering behavior.
07:27
<cvrebert>
TabAtkins: Aside from <legend>-related stuff, https://html.spec.whatwg.org/multipage/rendering.html#the-fieldset-and-legend-elements does not seem to mention anything magical.
07:33
<annevk>
cvrebert: regarding https://www.w3.org/Bugs/Public/show_bug.cgi?id=28954 I'm not sure it's worth clarifying since browsers haven't implemented their current set of rendering rules based on that section
07:34
<annevk>
cvrebert: the Rendering section is mostly reflecting what is implemented, sometimes tightening up things that were not defined
07:34
<annevk>
cvrebert: so it's kind of reasonable that there's a couple of mismatches
07:36
<annevk>
mkwst: woohoo, I think you added the first feature since GitHub
07:50
<zcorpan_>
is https://html5.org/tools/web-apps-tracker supposed to work? (broken because of the rename?)
07:51
<annevk>
zcorpan_: no, I want to redirect it, see logs
07:52
<zcorpan_>
oh, ok
07:52
<annevk>
zcorpan_: Twitter account no longer uses it
07:52
<annevk>
zcorpan_: if you're willing to provide patches :-)
07:54
<zcorpan_>
is the green button OK if there's just one commit in a PR?
07:56
<Ms2ger>
No
07:59
<MikeSmith>
The green button still causes multiple commits
08:01
<zcorpan_>
ok
08:01
<cvrebert>
See https://github.com/isaacs/github/issues/2 , https://github.com/isaacs/github/issues/5
08:08
<cvrebert>
TabAtkins: Is the "is expected to establish a new BFC" sentence sufficient to explain that, or is there further magic?
08:16
<zcorpan_>
should i do what github's "command line instructions" says or something else, to merge a pr?
08:17
<Ms2ger>
rebase the branch, confirm with git log, and push
08:17
<cvrebert>
drop the "--no-ff". my particular PR is already freshly rebased.
08:24
<MikeSmith>
zcorpan_: unfortunately the github "command line instructions" also cause multiple commits, I think
08:24
<zcorpan_>
MikeSmith: yeah i noticed
08:26
<MikeSmith>
an alternative to what Ms2ger suggested is, just cherrypick the commit from that branch
08:26
<MikeSmith>
to me, that's the easiest
08:27
<Ms2ger>
Yeah, I guess that would work
08:28
<MikeSmith>
it doesn't have any downside that I can think of
08:31
<cvrebert>
well, the PR won't show as merged even though it's already been rebased atop master, but whatever
08:31
<MikeSmith>
cvrebert: that's true I guess, yeah
08:33
<MikeSmith>
But I think for this repo/project, that's not a problem
08:33
<MikeSmith>
Anyway it's idiotic that we have to repeatedly have these discussions
08:33
<MikeSmith>
git is the complete opposite of elegance
08:33
<MikeSmith>
it gets in your way constantly
08:35
<zcorpan_>
i did what cvrebert said
08:38
<TabAtkins>
MikeSmith: It's not git's fault, it's github's. Rebase is the better option most of the time, but GH strongly favors merge, and makes it hard to do anything else.
08:39
<MikeSmith>
true yeah
08:39
<MikeSmith>
github is very slow about fixing broken stuff like this, too
11:28
<roc>
every single project I work on rebases commits to master
11:28
<roc>
so github PRs basically never show as merged
11:29
<roc>
yet github wins
11:29
<roc>
another triumph of network effects over common sense
11:33
<Ms2ger>
See also: git
11:46
<TabAtkins>
roc: One downside doesn't necessarily wreck a solution. ^_^
11:47
<TabAtkins>
Also, I'm lazy, and just merge things into Bikeshed.
11:47
<Ms2ger>
Exactly
11:51
<Ms2ger>
annevk, fixed https://github.com/whatwg/html/pull/21
11:53
<roc>
TabAtkins: sure, github's not completely useless, or network effects wouldn't be enough to overcome its defects
11:55
<Ms2ger>
Sadly
12:13
<nox>
Ms2ger: Made me laugh.
15:50
<annevk>
Ms2ger: you should check that parsedURL is not failure, no?
15:51
<annevk>
Ms2ger: "is not successful" seems wrong
15:54
<Ms2ger>
resolve-a-url says "abort these steps with an error"
15:54
<Ms2ger>
Might be better to make that "return failure" or something
15:55
<annevk>
oh okay, it's the wrapper
15:55
<annevk>
hopefully we can remove that at some point
15:55
<annevk>
I guess for now this is okayish then
15:55
<Ms2ger>
Yeah, wasn't too happy about that part either
15:57
<annevk>
I like how the HTML commit log has lots of different icons
15:57
<annevk>
Hopefully that'll go on like that
22:57
<Domenic>
annevk: you might be interested in https://github.com/git-up/GitUp/blob/master/README.md