[jacky]Yeah, I meant self-hostable! I'm wondering if there's a space/need for something like Gatsby but for the IndieWeb (in the sense that it'll ask for endpoints to Micropub, Webmention and IndieAuth and from there, you just do everything else)
[jacky]very late night thought but I've been thinking again about [tantek]'s nudging towards seeing Microsub as a handling for people and less on the feeds (in a way) and I think that could be a basis for leveraging things like `only to people I follow` or the like in a Microsub reader
[jacky]namely if Microsub was able to give a list of the h-cards that are the representative ones of a feed in a timeline, that could be used to determine the 'audience' of said feed
[jacky]the only thing there would be needed to give a human-friendly name to it (like "people Jacky follow", "Jacky's followers interested in tech" <- this could be via following a specific tag feed on one's site but needs to be done by the site's feed generation system, not the Microsub server!)
samwilson, hendursa1, [KevinMarks], pepperoni_, barnaby and [Murray] joined the channel
[Murray]jacky: if it were more like Next/Nuxt/latest version of 11ty, and allowed for dynamic page load options (SSR/SSG/hybrid) then I'd jump on something like that in a shot 😄 Then again, I think it would possibly make more sense to just create webmention/indieauth/etc. plugins for something like 11ty? And then just have an indieweb starter kit type thing
[Murray]oh also, on Drupal, I've been really interested in how much it seems to be re-entering tech conversations this year. I feel like even a year ago it was seen as "legacy tech" (eurgh), something bloated and undesirable. Suddenly, I see Gatsby have specifically targeted Drupal for version-release optimisations, Next are mentioning it in blog posts, several conferences are adding talks centred on it. Definitely feels like it's being
[grantcodes]I've always seen drupal as something for monolithic sites like charities and universities. I don't know if it improved but the ui was really daunting to me
barnabyRE ^^^: I’d love to do some developer usability testing for these libraries, i.e. we have a screen-sharing video call where I watch you try to get them working with your app. Anyone interested?
barnabyI only just learned how to use GH actions, and set up actions for my new indieauth and micropub libraries which run tests with coverage and static analysis on PHP 7.3, 7.4 and 8.0
barnabycoveralls.io’s design is really nice. I decided to just GH actions for simplicity, but integrated phpunit’s HTML coverage report into the project’s GH pages, so I get automatic static coverage reports for free
barnabyGWG: it’s a piece of software from vimeo which does static analysis of PHP software, and picks up on all sorts of type-related problems which can otherwise be hard to find
barnabyGWG: my GH actions/GH pages coverage reports is a bit of a hack and isn’t quite as integrated as I want. When developing, I run ./run_coverage.sh locally, which generates an HTML coverage report in /docs/coverage, which I open locally and check. This is also what gets uploaded to the GH pages along with the docs, e.g. https://taproot.github.io/indieauth/coverage/
barnabyadditionally, the GH action runs coverage tests and reports a basic text output in the action logs. this is intended for me to quickly be able to review PRs without having to pull them locally and run the tests myself
barnabyI was thinking of setting up a similar system for php-mf2, if the current maintainters are up for it. I noticed that the service currently used for CI there will shut down soon
barnabyI’d love to find a static analysis tool for python which works similarly to psalm. all the ones I found so far were either for niche use cases, or were more for policing coding style
[snarfed]granary (~1100) and webutil (~100) are just 1-2s each, total. bridgy (~600) is longer, ~25s total, mostly because it reads/writes to a database. isolated, in memory, but still.
barnabyI get especially annoyed at pylint for telling me that my lines are long, when python makes breaking call chains over multiple lines awkward and ugly. at least you can turn those warnings off
barnabyooh interesting, is that similar to intelligent fuzzing? seems like there are some great fuzzing tools available, I’ve never really had a project where it makes sense to use them
LaBcasse[m]I have changed the doctype of the exporter (https://orgmode.org/org.html#HTML-doctypes) and now everything works ! So, xhtml seems to be interpreted as xml ;)
jamietannaso the `accept` header from XRay is `Accept: application/mf2+json, application/activity+json, text/html, application/json, application/xml, text/xml`
[jacky]Murray: that's another option for sure re: site rendering options! I don't actively work on that level of the Web much nowadays so I've been a bit out of touch with it (outside of Eleventy)