Sorry - false alarm - I've been experimenting with nix for generating identical home and utility environments at home, on VPS, and at work, and I neglected to clean up a few parts of the nix environment that were interfering with the build. I'll investigate why the built files couldn't find the nix libc later, but for now I can build both liquid-rust and cobalt-bin
I'm assuming I should start on liquid, as it's a bit behind?
cobalt is going through a re-architecture, so liquid is a good place.
Things liquid needs:
- A re-sync of the conformance tests with the ruby implementation's tests (creating issues and ignoring tests that fail), see cobalt-org/liquid-rust#419
- Going through the backlog of conformance issues, see https://github.com/cobalt-org/liquid-rust/issues?q=is%3Aissue+is%3Aopen+label%3Astd-compatibility
- Random fixes
I've repeatedly noticed now that some of the (newer) features are undocumented… Thought about opening a PR to add a line to
CONTRIBUTING.md that requires to change the documentation accordingly when a feature is added or changed in a significant way. Thoughts? I mean, whoever adds an entire feature probably knows how to use the software enough to quickly describe how to use it, so it's not like this should be impossible to do… And the problem is there isn't really a point in adding features if most users don't know how to use them or even that they exist. I've opened a few issues on that sort of thing already, the latest one that would have to be added (which I only found coincidentally by looking at release notes) is file minification…
(haven't opened an issue on that one so far)
I see… That's of course an issue. Should I open an issue on the minify thing? If you tell me how this is configured, I could also just PR to the docs directly…
I'll have to look to see if there is a way to merge the repos and keep github pages working
Alright. If there isn't any way, I could also offer my webserver, you'd just have to point some domain or subdomain to it, or I could give you a subdomain like cobalt-org.artemislena.eu 🤷🏻♀️
I don't think so. It uses the excerpt, afaik… Perhaps you can adjust the excerpt to default to the entire content somehow? Adjusting
default.excerpt_separator in _cobalt.yml might work.
Out of curiosity, what value did you set it to?
self-documenting values are even better than comments.
Ah, alright. I'm not too aware of how Cobalt works internally, so didn't know whether not finding the excerpt separator might lead to it being empty.
https://cobalt-org.github.io/docs/front/A marker for what leading content should be extracted from the content as an excerpt. "" will cause no excerpt to be generated.
When there is no excerpt, we use the full content
Now cobalt seems to try to interpret that. At least "cobalt serve --drafts" errors like shown here: https://paste.sr.ht/~tsdh/e8d899adc04844c4b558f9db21895e31b68cb72c
I'd expect that listings aren't processed but taken verbatim. Is there something I can do about it?
Or
language\n...\n
epage: That's the post's content: https://paste.sr.ht/~tsdh/fd02358adddbfca3f4f7e0e1c9ee085f0ce72ab3
tsdh @tsdh gotta go to bed now but read answers tomorrow
You'll need
{{ raw }} for that
Is it possible to insert equations into Cobalt posts? Preferably with LaTeX syntax.
Natively? Not yet. We have an issue for cobalt-org/cobalt.rs#280 though I'm not particular about what style we support.
Is there any documentation about how to use the scss support in Cobalt?
Little direct documentation: https://cobalt-org.github.io/docs/assets/
You can choose our CSS is rendered: https://cobalt-org.github.io/docs/config/
You have the _sass directory for files you can import: https://cobalt-org.github.io/docs/directory/
Other than that, it should just be a matter of having the files on disk and they should get processed
3 replies
although I believe our DateTime implementation will need some tweaking in terms of what it accepts - our implementation should be able to handle PrimitiveDateTime (no offset), but right now it only accepts OffsetDateTime. Ideally, it should handle any single date, time, and datetime with or without offset. So I'm going to fiddle with our DateTime to get it to that point, and then fix with_time. Sound good?
Note that when I'm talking about PrimitiveDateTime and OffsetDateTime, I really meant the string representations of those types, because everything internally should be OffsetDateTime