00:14
<Kaiido>
Is { subtree: false, customElements } a valid input or is customElements relying on subtree being true?
02:50
<sideshowbarker>
I’m looking at https://github.com/mdn/content/pull/38254/files and wondering why MDN might have ever had NodeFilter.SHOW_ATTRIBUTE and NodeFilter.SHOW_CDATA_SECTION flagged as deprecated to begin with
02:50
<sideshowbarker>
Maybe just somebody was confused, or maybe just a plain mistake, I dunno
02:52
<sideshowbarker>
I’m going ahead and merging that patch, but… just curious about the history. git blame seems to shows it’s been that way since the old wiki days — before the sources were moved into a git repo (~5 years ago)
07:36
<annevk>
Kaiido: good question, that would be a valid input.
09:53
<annevk>

TabAtkins: minor thing, some kind of indication that the line numbers for

FATAL ERROR: IDL SYNTAX ERROR LINE: 55 - expected ";"

are based on the concatenated IDL and not the actual Bikeshed input would have saved me a bit of time.

09:54
<annevk>
sideshowbarker: we tried to remove Attr and CDATASection nodes from the DOM and failed.
09:56
<annevk>
keithamus: https://github.com/whatwg/html/pull/10869 and https://github.com/whatwg/dom/pull/1341 should be good now. I will update the PR template a bit. And I'll align WebKit and tests with the selfOnly member. I think that's the only thing we don't do yet.
09:57
<keithamus>
Thanks for updating these!
09:57
<annevk>
I also updated the HTML issue's OP so it has the latest IDL sketch. But I recommend working from the actual PRs.
09:57
<keithamus>
perfect, thank you
11:31
<Ms2ger>
I’m looking at https://github.com/mdn/content/pull/38254/files and wondering why MDN might have ever had NodeFilter.SHOW_ATTRIBUTE and NodeFilter.SHOW_CDATA_SECTION flagged as deprecated to begin with
Maybe because we wanted to remove cdata and attr nodes
11:31
<Ms2ger>
👴
11:32
<annevk>
TabAtkins: did Bikeshed change how it does references to terms? They all to be path-absolute URLs now? At least on https://whatpr.org/dom/1341/b64559c...71537d2.html I can no longer click on them...
13:20
<annevk>
keithamus: tests have landed. selfOnly test rename is up for review here: https://github.com/web-platform-tests/wpt/pull/50860
15:07
<TabAtkins>

TabAtkins: minor thing, some kind of indication that the line numbers for

FATAL ERROR: IDL SYNTAX ERROR LINE: 55 - expected ";"

are based on the concatenated IDL and not the actual Bikeshed input would have saved me a bit of time.

I just fixed those line numbers earlier this week, try updating Bikeshed.
15:08
<TabAtkins>
TabAtkins: did Bikeshed change how it does references to terms? They all to be path-absolute URLs now? At least on https://whatpr.org/dom/1341/b64559c...71537d2.html I can no longer click on them...
Uhhhh no, that's unexpected. I'll have to debug that. I shouldn't have made any changes around that.
16:42
<annevk>
TabAtkins: I only ever use CI Bikeshed. I got these results this morning.
17:03
<annevk>
TabAtkins: I suspect "Boilerplate: script-dom-helper" whatever that is. Perhaps it's time to audit all these scripts that execute on specifications again.
17:04
<annevk>
Except that I lack time. 🫠
17:10
<TabAtkins>
I'll check things out in a sec
17:47
<TabAtkins>
Which spec is this, btw?
17:48
<TabAtkins>
ah nm, I see the problem. didn't realize the error messages varied there.
17:49
<TabAtkins>
(I'm hacking around the lack of a line number offset by parsing the error message to look for line numbers.)
18:19
<TabAtkins>
So at a quick glance, this seems to be caused by the <base> element right at the top of the document. Is that a recent change? The links themselves are normal href="#foo" things.
18:26
<TabAtkins>
(I can't immediately find where/how whatpr documents are generated.)