18th of March, 2026
The quick summary of my current week has been explained in our PyScript Community Call that happened yesterday. There we discussed Saint Patrick’s Day at the very beginning, but then I’ve explained where I am, what I am doing, what I hope will work reliably for our first quarter challenge and idea.
On the other hand, while I try to focus as much as I can on my current task, working on OSS (Open Source Software) means I gotta help both the projects I am after and the projects I use to move forward … did I say just “projects”? Sometimes those “projects” are the Web platform we target daily.
This update is around OSS that must be, or should be, fixed:
- flatted is a library downloaded by too many people (340 million times per month right now) to keep eyes off when security concerns happen and lately 2 specific security things happened (you might see only one until the next one, already fixed and patched, will show up). We use that in PyScript, all things fixed the day after!
- Firefox Atomics.pause is different from others, I’ve been involved in that bug too. It’s a “won’t fix”; we had to change PyScript code (or related code) around it, we’re at least good now in a cross-browser fashion
- Firefox MessageChannel API is broken and I was redirected to that bug after filing my own one. You might think I am glitching in here but every bug I’ve recently filed against FireFox leads there … there is an issue with whoever edited the original issue topic, because that was supposed to be “MessagePort.postMessage for received MessagePorts will fail to send messages if blocking APIs (Sync XHR, Atomics) are used prior to the Entangling state machine stabilizing; workaround is to wait for receipt of a message” … and I’ve just realized I need to follow up …
- in this WHATWG issue I’ve interacted with standard bodies and provided a “polyfill” or usable PoC to solve the issue, resulting into a project that anyone can play around with … time passed, the polyfill might be outdated but it helped shape the new marker HTML5 standard. This will be used in PEP750 (which I’ve helped deliver in Python) because these markers are a “game changer” for template-literal-string-based parsers.
- for all discussions and ideas around the Web, whenever convenient I’ve written test cases, demos, created live examples and summarized some of these in this repo
I am guilty as charged: I am not tracking my time/effort too well around these topics, and that’s since about forever (I’ve helped ship CSS :has(...) selector with my previous company and gosh if that’s been time consuming) but here I am with latest things that concern my current project and I’ve tried to improve, help, besides the effort spent to indeed help land PEP750 (see my name in Acknowledgements) and helped shape PEP 818 (same as before) too.
The TL;DR is: do I encounter an issue or do I see a possibility to help the community? Count me in 😇