00:54
<nikkibee>
hi annevk_, I just saw your email
00:55
<annevk_>
nikkibee: ah hey!
00:55
<annevk_>
nikkibee: email is still better, not the most reliable connection on the train here
00:55
<annevk_>
nikkibee: and I have to transfer soonish
00:55
<nikkibee>
vakila messaged me after they saw my message on issue 265 and we chatted about that
00:56
<nikkibee>
gotcha!
00:56
<annevk>
ah okay, vakila helped you out?
00:56
<nikkibee>
I was actually in the process of forking the repo when they saw my message on github saying I wanted to do it
00:56
<nikkibee>
so no problem there
00:57
<nikkibee>
so if I'm reading you right, working on most of the good first bugs means updating documentation as described, without changing the code. but I need to look at the relevant code to make sure the documentation matches?
00:57
<MikeSmith>
nikkibee: I'll be around and can help answer questions when needed (I won't be at the TPAC conference myself)
00:58
<nikkibee>
thanks MikeSmith!
00:58
<annevk>
nikkibee: nothing requires changes to html-build afaik
00:58
<annevk>
nikkibee: you only need to change "source", which is HTML
00:58
<nikkibee>
right
00:59
<annevk>
nikkibee: but that HTML does describe algorithms and such, which you sometimes will need to understand in order to make a change
00:59
<annevk>
nikkibee: but you can't really run it, since it's English
00:59
<nikkibee>
ohh. I get it. these are the html standards
00:59
<annevk>
nikkibee: so it's a slightly different exercise from writing code in general
01:00
<nikkibee>
implementations are up to vendors... so there's no code for me to look at to compare documentation to really
01:00
<annevk>
nikkibee: yeah "source" is basically the HTML standard, and html-build makes it a bit prettier
01:00
<annevk>
nikkibee: yeah, well, you can of course look at the source code of various browsers
01:01
<annevk>
nikkibee: or write a testcase/demo to find out what a browser does and whether that matches the standard
01:01
<nikkibee>
mhm. but I can't take how one browser does it as necessarily definitive
01:01
<annevk>
nikkibee: unfortunately not
01:02
<annevk>
nikkibee: you have to test all the browsers and then make a decision of sorts as to what is best, but we can help you with that once you get there
01:02
<nikkibee>
thanks
01:02
<annevk>
nikkibee: most of the good first bugs shouldn't require that much research though
01:02
<nikkibee>
just looking at the "source" file?
01:02
<annevk>
yeah, just looking at that and changing some things around
01:03
<nikkibee>
I get it a lot better now, thanks
01:03
<nikkibee>
I'll start with https://github.com/whatwg/html/issues/179 like you mentioned in the email
01:03
<annevk>
nikkibee: improving the documentation is very much welcome though, although not always the most rewarding
01:04
<annevk>
I've been meaning to do that, but got distracted
01:04
<annevk>
back later
01:04
<nikkibee>
that's alright, right now I'm looking to test the waters in general and get something in on time for my outreachy application
01:47
<nikkibee>
https://github.com/whatwg/html#pull-requests I don't understand the 100 characters convention
01:47
<nikkibee>
does this only apply to the text that will be shown? because a lot of lines go beyond 100 characters
02:21
<nikkibee>
https://github.com/whatwg/html-build#prerequisites is it possible to follow this on windows? like using cygwin or something
02:30
<Domenic>
nikkibee: it applies to the length of lines in the source file
02:30
<Domenic>
nikkibee: cygwin might work; I have not tried it. I use a Linux machine and most people seem to use Macs.
02:31
<nikkibee>
I thought it meant line length. but a lot of lines don't follow it
02:31
<nikkibee>
should I fix up lines in the same section that I'm working on rn?
02:31
<nikkibee>
I got a linux vm which I'm competent with, I just find it easier and faster to edit files on the host os
02:32
<Domenic>
nikkibee: generally we only fix up the smallest number of lines possible, to avoid generating diffs that are just moving whitespace around.
02:32
<nikkibee>
that makes sense
02:33
<nikkibee>
I'll try out cygwin later, right now I just want to get my first PR in :)
02:33
<Domenic>
:)
02:39
<nikkibee>
Domenic: for the line length, is it okay to split a <> block in half or so?
02:40
<nikkibee>
like I have a part where "<code data-x="dom-windowtimers-setTimeout">" is at the end of a line and half-way through it hits the 100 character limit
02:40
<nikkibee>
should I put a line-break there, or right before the whole block?
02:43
<Domenic>
nikkibee: it is OK to split that after "<code" but not anywhere else
02:44
<nikkibee>
gotcha! thanks
02:44
<nikkibee>
I'd rather split it at the start cause it's only a 5 character difference, and it's more readable that way imo. that's OK too, right?
02:46
<Domenic>
sure, np
02:46
<nikkibee>
good
02:49
<nikkibee>
https://github.com/whatwg/html/issues/179 I'm working on this issue, and I noticed near the start of the timers section, right above the big note block giving an overview of the functions, is the html comment <!-- in practice clearInterval() and clearTimeout() are synonyms, but we don't mention this here -->
02:49
<nikkibee>
since I'm now clarifying that this is the case, should I remove that comment, since it's contradictory to the section?
02:52
<Domenic>
yes, that seems like a good idea
02:53
<nikkibee>
done
03:16
<nikkibee>
https://github.com/whatwg/html/pull/282 PR made! @ Domenic
03:41
<nikkibee>
hey Domenic, I'm having a hard time with squashing
03:42
<nikkibee>
http://gitready.com/advanced/2009/02/10/squashing-commits-with-rebase.html I've been shown to follow this which I can do fine
03:42
<nikkibee>
but after squashing the commits I'm not sure what to do next
03:42
<nikkibee>
I did some history digging with somebody else and last time I did `git push origin master -f`, but they I was talking to about it said I should check with you about running that
03:48
<Domenic>
nikkibee: git push origin master -f is fine in this case
03:48
<nikkibee>
thanks
03:52
<nikkibee>
oof I must've done something wrong cause there's four commits now :S
03:53
<caitp>
uh oh
03:53
<nikkibee>
I tried to rebase again to change the syntaxt of the commit message to follow this better https://github.com/erlang/otp/wiki/Writing-good-commit-messages
03:55
<caitp>
if you do `git rebase -i HEAD~4` and replace `pick` with `s` in the text editor, and then force push, you should be okay
03:55
<Domenic>
Also when editing the commit message be sure to remove the previous ones instead of leaving them there
03:55
<caitp>
er, except for the first commit, leave that one as pick --- and uh, make sure they're all your commits
03:57
<nikkibee>
alright, I've done all of those now
03:58
<nikkibee>
argh that didn't turn out right
03:59
<caitp>
it has the changes from at least 3 of the commits, the diff looks pretty close to how it did before
03:59
<nikkibee>
but the commit title isn't what I left behind, and the full message has all the stuff from before for some reason
04:00
<caitp>
easy fix: git commit --amend
04:00
<caitp>
and force push again
04:00
<nikkibee>
thanks :)
04:00
<nikkibee>
how do I delete a line in VIM? I seem unable to
04:01
<nikkibee>
I've deleted like. all the text except the commit message I want
04:01
<nikkibee>
but I have a bunch of blank lines before and after I can't seem to remove
04:01
<caitp>
in the not-insertion-mode, press `d` twice to delete a single line that you're on
04:01
<nikkibee>
thanks
04:02
<nikkibee>
I just realized I can do it in insert mode, I just accidentally left it
04:02
<nikkibee>
there we go
04:04
<nikkibee>
Domenic: how does it look now? https://github.com/whatwg/html/pull/282
04:10
<Domenic>
nikkibee: almost right. Go with "Fixes #179: Add note about clearing timers" (titles don't consist of two sentences, and you want to use present tense consistently)
04:10
<nikkibee>
that makes more sense
04:11
<nikkibee>
done and pushed @ Domenic
04:12
<Domenic>
nikkibee: perfect! Now I just need to dig out my Linux machine to verify and merge... might take ~15 minutes, I'm in the middle of some packing for a trip.
04:14
<nikkibee>
alright, sounds good
04:49
<nikkibee>
urk I think I hit a weird key combo in vim by mistake
04:50
<nikkibee>
it took me back to the git prompt but I can't do anything, anything I type just makes "More?" appear on the input line
04:51
<Domenic>
nikkibee: there are ways to set the git editor to something less crazy than vim...
04:51
<Domenic>
nikkibee: probably to get out of your current state, Ctrl+C?
04:51
<nikkibee>
ctrl+c just makes the git prompt line appear again
04:51
<nikkibee>
I'm still stuck in this
04:51
<nikkibee>
can I make the git editor default to like, sublime text?
04:52
<nikkibee>
I don't want to close the terminal cause I'm in the middle of rebasing
04:52
<Domenic>
yeah you can, there's a slight trick to it, let me find that
04:53
<Domenic>
i would close the terminal then reopen it and do "git rebase --abort" so you can get into a fresh state
04:53
<nikkibee>
okay, that sounds fine then
04:53
<Domenic>
https://stackoverflow.com/questions/8951275/how-can-i-make-sublime-text-the-default-editor-for-git/9408117#9408117
04:54
<nikkibee>
that's rad, thanks
04:55
<nikkibee>
I miss the handy text colouring to let me know what I'm doing right, but this is much less stressful
04:57
<caitp>
theres probably a git syntax highlighter for sublime, isnt there?
04:57
<nikkibee>
not by default it doesn't look like
04:58
<nikkibee>
these files don't have an extension so I'm not sure how sublime would pick up on it
04:59
<caitp>
looks like there are lots of git plugins for it
04:59
<nikkibee>
that would be handy! I'm sure if there's plugins they know how to get sublime to recognize it
05:00
<nikkibee>
I've squashed and pushed the commits again @ Domenic
05:07
<Domenic>
nikkibee: great, re-building now!
05:09
<nikkibee>
:)
05:11
<Domenic>
Yay, merged!! Thank you so much!
05:11
<nikkibee>
:D thank you too for all the help!
05:11
<nikkibee>
I'm excited to be part of this
05:14
<Domenic>
\o/
14:26
<Ms2ger>
TabAtkins, where does bikeshed get its [[REFS]]?
14:58
<TabAtkins>
Ms2ger: Mostly SpecRef, but also Bikeshed's linking database (if [[ref]] is the shortname of a spec indexed by Shepherd/Bikeshed, it'll link).
15:07
<Ms2ger>
Pointer? It's missing OMGIDL
15:52
<caitp>
are iterable<> interfaces with an integer length property necessarily supposed to use %ArrayProto_values% as their iterator, or does that only apply to interfaces with the integer-typed length attr, indexed element getter, and no other form of iterable<> declaration
18:00
<hatmaker>
Hello!
18:00
<nikkibee>
hello
18:00
<botie>
privet, nikkibee
18:01
<hatmaker>
anyone here know Alystair?
18:31
<nikkibee>
http://testthewebforward.org/docs/configuration.html#branch I'm following these steps for https://github.com/w3c/web-platform-tests/issues/2277
18:31
<nikkibee>
I'm trying to think of a good name for the branch
18:31
<nikkibee>
"timers-cleartimeout-clearinterval"?
18:39
<nox>
https://reviewable.io/reviews/servo/servo/8114#-K1BFtVuDw8ymsMAeG5Q Apparently, dataset is badly specified, could someone clarify whether this is intended or not?
18:39
<nox>
https://html.spec.whatwg.org/multipage/dom.html#custom-data-attribute 'name starts with the string "data-", has at least one character after the hyphen'
18:40
<nox>
https://html.spec.whatwg.org/multipage/dom.html#dom-dataset 'For each content attribute on the element whose first five characters are the string "data-" and whose remaining characters (if any)'
18:40
<nox>
That 'if any' contradicts the first HTML spec link I pasted.
18:54
<jgraham>
nikkibee: The branch name doesn't matter at all, but that sounds fine
18:55
<nikkibee>
okay, thanks
18:55
<jgraham>
I mean having something somewhat descriptive helps, but it's shortlived anyway
18:55
<jgraham>
nox: Without reading the spec that sounds like UA requirements vs author requirements
18:57
<nikkibee>
https://github.com/nikkisquared/web-platform-tests#running-the-tests I'm trying to follow this and I get an error from a python file after running ./serve
18:57
<nikkibee>
note that I'm on windows
18:57
<nikkibee>
I followed the windows notes
18:58
<nikkibee>
https://gist.github.com/nikkisquared/5217fb6847348e149107 this is the error
18:59
<nikkibee>
I tried following the note at the end about `openssl` not working on the commandline but I don't think I'm doing it right
18:59
<nikkibee>
it doesn't work on the command line for me, even though I put it in the path
19:00
<nikkibee>
oh what the! the path didn't save my changes
19:01
<nikkibee>
I can run `openssl` now but I still get that error
19:03
<nikkibee>
I'd like knowing I can run the tests myself
19:27
<nox>
jgraham: Oh I see.
19:31
<Domenic>
Wasn't there some attempt at making inputEl.files assignable? sicking might remember?
19:32
<sicking>
Domenic: yeah. Mainly we need the .item() problem solved
19:33
<Domenic>
sicking: but couldn't we allow inputEl.files = any sequence and then convert it to a FileList? Separately from fixing FileList vs. array
19:33
<sicking>
Domenic: i think that's the only remaining blocker
19:33
<sicking>
Domenic: true, we could. You couldn't express that in WebIDL, though that's a relatively small hurdle
19:33
<Domenic>
Hmm I see.
21:53
<TabAtkins>
sicking: Use a getter/setter pair and some prose.
21:53
<sicking>
TabAtkins: that doesn't let us use the .files syntax, does it?
21:54
<TabAtkins>
ms2ger: https://github.com/tobie/specref
21:54
<TabAtkins>
sicking: Of course it does.
21:54
<TabAtkins>
getter/setter pairs looks like an attributde
21:54
<sicking>
that's webidl syntax i don't know if on that case
21:56
<sicking>
oh, i don't think we'd want the overhead of turning HTMLInputElement into a proxy, if that's what you're referring to?
22:06
<sicking>
Domenic: i would be a very happy camper if we got sequence_with_item<> in WebIDL. Or if ES Arrays grew a .item() function. Though the latter is probably not web compatible given that even adding .contains() wasn't possible.
22:07
<sicking>
Domenic: hmm... we probably need FrozenArrrayWithItem actually
22:25
<TabAtkins>
sicking: It's just getter/setter pairs in JS. WebIDL supports it. All "attributes" in WebIDL are actually getter/setter pairs when translated to JS, too.
22:25
<sicking>
TabAtkins: ok, this is some WebIDL feature I don't know about then
22:26
<TabAtkins>
Again, this isn't WebIDL. It's JS. WebIDL just supports it.
22:26
<sicking>
i don't know what you mean by "WebIDL just supports it"
22:26
<sicking>
i know JS has getters and setters. That doesn't mean that webidl can do what Dominic is asking for
22:26
<TabAtkins>
WebIDL has syntax to express it.
22:27
<sicking>
if that's the case, then that's some WebIDL feature that I don't know about
22:28
<TabAtkins>
I'm still confused. There's nothing to "know about". If you know about JS getters and setters, then you know about WebIDL ones. The only detail is the precise way to write one in an interface.
22:28
<sicking>
i just know about "[readonly] attribute <type> <name>"
22:28
<sicking>
which is not enough here
22:28
<sicking>
since Dominic was asking for different types for the getter and setter
22:29
<TabAtkins>
I'm not convinced you know about JS getters and setters here. ^_^
22:29
<sicking>
ok
22:29
<TabAtkins>
^_^
22:29
<TabAtkins>
You just have a getter that takes a sequence, converts it to a FileList, and assigns that to the internal [[files]] slot. Then a setter that returns the [[files]] slot.
22:29
<sicking>
show me the WebIDL syntax
22:30
<sicking>
i know perfectly well that JS suppots this
22:30
<sicking>
i haven't claimed otherwise
22:31
<TabAtkins>
Okay, I'm the stupid one here. I'm recalling a WebIDL bug that heycam hasn't fixed yet. IDL getter/setter is for proxies. Ugh.
22:32
<sicking>
cool
22:32
<TabAtkins>
Sorry about that.
22:32
<sicking>
no worries, glad we came to clarity :)
23:48
<jsbell>
tuning in late, but sicking: stats show it'll be difficult to remove .items() on the DSL from Location.ancestorOrigins and .contains() on the DSLs from IndexedDB :(
23:49
<jsbell>
(i.e. yeah, sequence_with_XXX<> looks necessary)
23:49
<jsbell>
I'll mail webapps when I have cycles to care
23:49
<sicking>
jsbell: yup. That was my expectation
23:49
<sicking>
jsbell: heh
23:50
<sicking>
jsbell: Gecko is blocked by still not supporting subclassing Array as I understand it
23:50
<sicking>
once that is possible I'd like to simply switch a few classes over rather than waiting for WebIDL to catch up
23:51
<jsbell>
We welcome your implementation feedback. ;-)
23:51
<sicking>
cool :)