04:31
<sideshowbarker>
No idea, either. But I tell you, the more you look at the build stuff for that spec, the more it's gonna make just scratch your head in puzzlement
04:32
<sideshowbarker>
Kinda like with the content of the spec itself
04:34
<sideshowbarker>
That thing has such a long way to go that it makes me wonder if we might be facing a kinda “You can't get there from here” situation
04:35
<sideshowbarker>
I mean, we might want to be asking if we'd be better off tossing it all and rewriting it from scratch.
04:36
<sideshowbarker>
…With the caveat for the “we” part meaning, somebody other than me specifically doing the actual work
04:37
<sideshowbarker>
“Belling the cat” and all that
06:13
<annevk>
TabAtkins: does <hx algorithm> not result in variable highlighting? Example: https://encoding.spec.whatwg.org/#big5-decoder
06:55
<sideshowbarker>
https://wadakatu.github.io/w3c-hackathon/
06:56
<sideshowbarker>
https://github.com/wadakatu/w3c-hackathon
09:52
<zcorpan>
sideshowbarker: oh there's a build.py that creates .include files. Some custom syntax pre-processing...
09:56
<sideshowbarker>
Lovely
10:27
<zcorpan>
TabAtkins: is there a way to have bikeshed include the filename in the error messages? uievents has many build errors in include files and there are only line numbers in the messages
12:27
<zcorpan>
TabAtkins: for HTML parse errors; I see some messages have source file info
13:22
<zcorpan>
TabAtkins: fyi, uievents has its own syntax for markdown tables
15:07
<TabAtkins>
TabAtkins: is there a way to have bikeshed include the filename in the error messages? uievents has many build errors in include files and there are only line numbers in the messages
That should now be happening in Bikeshed 7 already. If it's not, lmk and I'll look.
15:08
<zcorpan>
TabAtkins: ah yes, I had an outdated bikeshed when I started. It works now
15:30
<annevk>
TabAtkins: did you see my question above? Not particularly pressing, though it would have been useful today if that had worked and I think I expected it to work when I made the changes to Encoding. But evidently I didn't test or it regressed somehow.
15:31
<TabAtkins>
I did see, but it was late and I didn't want to check my code. And now it's early and I don't want to check my code. I'll check in a few hours 😄
22:13
<TabAtkins>
Hm. I was under the impression that HTML would sometimes split a section "halfway" with an <hr>, and was trying to come up with a consistent behavior to implement for Bikeshed to handle that. But looking over it now, it seems that's not true, and instead it conforms to the restrictions I first wanted to apply (all pages are started by a heading, and can't contain headings of a lower level). Am I missing some examples of more complicated page-splitting, or is that indeed it?