| 12:16 | <peterv> | annevk: I'm a bit confused, HTMLAnchorElement doesn't add unscopables, does it? |
| 12:47 | <annevk> | peterv: I was confused! |
| 12:48 | <peterv> | ok 🙂 |
| 12:49 | <annevk> | I cannot think of an example like the one OP mentions, but maybe it's still theoretical at this point |
| 13:01 | <peterv> | annevk: we certainly don't have any WebIDL in Gecko that does it currently, and from what I remember we'd just have the added unscopables in the derived interface |
| 13:01 | <peterv> | with the current implementation I mean |
| 14:53 | <Andreu Botella> | sideshowbarker: https://w3c.github.io/csswg-drafts/timestamps.json |
| 14:53 | <Andreu Botella> | it's showing the timestamp of the latest commit in the repo for all specs |
| 14:58 | <Andreu Botella> | oh, that might be because the git checkout in Github Actions is sparse |
| 15:05 | <sideshowbarker> | ah — yeah in general since there’s not a good way to test GitHub Actions locally, I hadn’t anticipated that the workflow might be set up that way. When I tested it locally, I just did a full clone and I found that the timestamps were what I’d expected them to be |
| 15:07 | <sideshowbarker> | but still I would expect that even with a sparse checkout, git would give accurate timestamps. It seems counter-intuitive to me at least that git would do otherwise. I guess I can test it locally with a sparse checkout to see |
| 15:09 | <Andreu Botella> | I think git is trying to find any commits that modify the file, and here there aren't any in the section of the history it has |
| 15:10 | <Andreu Botella> | I'm not sure why it would show up the earliest commit rather than erroring or something, though |
| 15:13 | <sideshowbarker> | well, looking at https://github.com/w3c/csswg-drafts/blob/main/.github/workflows/build-specs.yml now, nothing jumps out to me that seems to be explicitly doing a sparse checkout. And I’m now also realizing that I don’t really know what the use case for a sparse checkout is — when I mentioned it, I had initially conflated it in my mind with a shallow clone. But I see it’s something completely different actually |
| 15:13 | <Andreu Botella> | oh |
| 15:13 | <Andreu Botella> | I'm also conflating those, I guess |
| 15:13 | <Andreu Botella> | I meant a shallow clone |
| 15:13 | <sideshowbarker> | ah OK |
| 15:14 | <Andreu Botella> | the actions/checkout@v3 action by default does a shallow clone with -depth=1 |
| 15:14 | <sideshowbarker> | aha |
| 15:14 | <Andreu Botella> | you can do to make it a full clone |
| 15:14 | <Andreu Botella> | https://github.com/actions/checkout |
| 15:15 | <sideshowbarker> | OK yeah let’s please do that |