#microformats 2026-05-02

2026-05-02 UTC
[morgan], plantroon and [jeremycherfas] joined the channel
#
Zegnat Ah yes. The JS parser is under the microformats org on GitHub, right? But I guess the package on npm is not, so only Aimee can do it?
#
jkingweb I'm finally satisfied with the state of my revived mf-tester, and it's helped me to improve the test suite a bit. If anyone's curious about results, I've posted them here: https://jkingweb.ca/mftester/
#
Zegnat jkingweb++ that is really cool!
#
Loqi jkingweb has 9 karma over the last year
#
jkingweb And if anyone would like to review my patches against the test suite which are pending, I'd appreciate it.
#
Zegnat jkingweb did you ever consider doing something like https://babelmark.github.io/ ? I have been going back on forth on that, wondering how I would do it. Might have to check out your source to see how you run the different languages side-by-side.
#
jkingweb I'd never heard of this, so no. Intriguing idea, though.
#
Zegnat It was a great help during the early commonmark days to figure out spec discussions. And might be able to assist those for mf too.
#
Zegnat I though about maybe doing one that queries the public websites listed on https://microformats.io/#parsers . That is how Babelmark works, it queries different servers over HTTP for different parsers.
#
jkingweb Yeah, this tester is purely local. Which gets rather involved, as you can imagine given all the different execution environemtns for the different languages.
#
jkingweb It'd be nice to have it in a Docker container or something, but I have no experience with such things.
#
Zegnat I am pretty comfortable doing some containerising for javascipt/php/python. But no experience with the other languages. Additionally my problem is that I need to put something in there that receives payloads from the outside to parse. A webserver or something else. That usually stumped me for the languages I am not good at :(
#
Loqi aww, cheer up
#
Zegnat Lookling at your test repo, that one I could probably get every language running in a container for. Simply because it does not need to take input. It just needs to run and outputs its results to files, which can be a directory mounted from outside the container.
#
Zegnat (Happy to work on a PR if you think it helps your tests to not require all the language pipelines to be installed outside.)
#
jkingweb Yes, please!
#
jkingweb Right now, it works—on my Linux desktop. I wasn't able to get all the tooling set up on my Mac laptop, though, so it'd be nice to have something less fragile.
#
Zegnat I will do a bit of testing. In my mind it is a Docker Compose file that spins up a container per parser. But I need to read a little bit more into how you configure each parser right now.
#
Zegnat I will come back with ... something :P
#
jkingweb Eager to see what you come up with.
#
jkingweb And I'm sure together we can figure out a reasonable way to make a Web interface for single-file input.
barnaby joined the channel
#
Zegnat jkingweb are your test results lying about the test suite commit ID? Your results show more tests being run than mine ...
#
jkingweb Oh. Yes. It's that commit plus the three unmerged PRs.
#
Zegnat Was just checking, made comparing output a tad harder. I think I got almost everything running in Docker now. Though probably not at all correctly packaged ;)
#
Zegnat There you go jkingweb! I do not have Elixir/Perl/Rust pipelines on my machine, but was able to run all the tests. https://github.com/mensbeam/mf2-tester/pull/1
#
Zegnat That should get all of them to run on your Mac laptop, if you are in for a test. As long as you have Docker installed. Hopefully :D
#
jkingweb Wow, that was fast. I'll have a look.
#
Zegnat Fast because I left a lot up to future us to figure out. It does as few changes as possible to the repo.
pfych, gRegor and saccharin joined the channel