WHATWG
2025-11-29
plaintext logs
prev next
source on github
07:33
<sideshowbarker>

annevk: TabAtkins Sorry about the checker change breaking builds; going forward, I’ll try to make sure I don’t end doing something like that again. For now:

  • I’ve changed the checker messages about role=directory to being warnings rather than errors.
  • In https://github.com/validator/validator/commit/4cd3a20ef1 I added some automation that will now cause every single WHATWG spec to be checked under CI on every push to the checker sources
07:34
<sideshowbarker>
TabAtkins: If there are other specs you’d recommend I also add automated checking like that for, lemme know — I’d be happy to add more.
07:36
<sideshowbarker>
And for the record, I’ll confess that if I had been following the ARIA spec requirements strictly/properly, the messages about role=directory should have warnings to begin with — not errors. That’s because the ARIA spec defines role=directory as just being deprecated — not strictly non-conforming.
07:39
<sideshowbarker>
But that said, for document/author-conformance requirements that I have to maintain code for in the checker, in general I really don’t like the (anti)pattern of specs redefining some feature as deprecated but then over time never ever getting around to eventually graduating those legacy features to being fully obsolete (=strictly non-conforming).
07:41
<sideshowbarker>
Because over the long term — even after documents/authors have mostly long ago stopped using whatever particular legacy/deprecated feature — I still have to keep that code in there, despite it having reached the state of no longer serving any good purpose.
07:43
<sideshowbarker>
The CSS specs unfortunately also seem to kind of follow that “deprecate things but never actually get around to going back and fully-obsoleting anything” (anti)pattern.
07:49
<sideshowbarker>

annevk: By the way, there are about 18 WHATWG specs that are still using role=directory. See https://gist.github.com/sideshowbarker/2f300510b0ec2af9f4832b1396b20244

So I guess those all probably should be re-published. But that said, because I’ve now changed the message about that to being a warning rather than error, at least it won’t break the build for any of those specs.

16:32
<krishna bodigiri>
Hi all
16:36
<krishna bodigiri>
display:flex means default behaviour is all item will be arrange in horizontal flexible iam thinking like above is that my statement correct