1. @leftpaddotpy I ran into this bug on NixOS just yesterday so I'm glad you're on the case :D

  2. @alice_i_cecile I look forward to these every week! You do an incredible job of both being an great maintainer and an accessible communicator. I'm not a Bevy user (yet?) but I appreciate hearing about what's going on with the project and getting to pick up on some advanced Rust concepts along the way through your reviews :D

  3. @teapot_ben @neil I have been looking far and wide for an easier to deploy version of wallabag, thank you so much!

  4. Kinda hit a plateau on my weight loss journey and it's fucking with my motivation big time

  5. @arch @chirpbirb it's definitely a WAF configuration problem and Discord is the only responsible entity. They appear to have blocked the Chrome user agent which is just plain funny.

  6. @denmanrooke @hikari they did some configuration fuckery with their Cloudflare Firewall and managed to block the Chrome user agent, which in turn broke the Electron-powered app.

  7. @ybinnenweg that's an option I've been looking at, but realistically I might opt to have a second server with Gitea's CI runner on it since it'll significantly ease migration from GitHub Actions.

  8. I've been experimenting with running a #NixOS server to self host services that are on their enshittification path, and a Git forge was first.

    I ended up picking Gitea for no particular reason other than familiarity and ease of deployment, so I now have an active mirror of all my GitHub repositories and I will slowly try to make it the primary hub as I figure out CI and other things.

    git.msfjarvis.dev/explore/repos

  9. @GeePawHill It's unexpected that it didn't work but not really surprising. The new search interface is, to put it bluntly, quite garbage at being a search engine.

  10. @zsmb13 Can't imagine a better place for you, congratulations! Looking forward to the incredible work you'll do there :D

    Replying to source post

  11. #LazyWeb I had a #Firefox question that I was hoping I could get help with. Coming from Chrome I am used to address bar autocompletion work on full URLs, but Firefox seems to break them down to domains and path segments which I don't want. Can I change this?

    As a concrete example, my Tailscale admin panel is at login.tailscale.com/admin/machines and I visit this URL directly. When I type 'login' in the address bar, Chrome suggests the full URL but Firefox just says 'login.tailscale.com'

  12. The urge to port Claw to #Kotlin Multiplatform again is slowly rising

  13. @seb @cb hCaptcha is by far the worst. It might be my top 3 most hated things about the internet tbh.

  14. After over a month of consistent and daily exercise I'm sad to report that the gymbros were right, it does feel good.

    A mashup from Google Emoji Kitchen, with a blushing and smiling face in the middle and stars bursting out of them.
  15. @tkuenneth@snapp.social IntelliJ Plugin to launch Android apps on Windows Subsystem for Android?

  16. Gonna make a serious attempt at a blog post next week, it's been far too long.

  17. @samuteki Honestly, not much. I've maintained that the idea of having to hear everyone out just because you happened to scream into the void is stupid. If I see any post which rubs me the wrong way, I block the author.

  18. I was bored last night so decided to look into #Gradle sync performance for the $DAYJOB codebase, it averaged around 6 minutes when downloading all dependencies with over 1m30s being wasted in failed HTTP requests. After some massaging with `exclusiveContent` sync is now down to 3m17s with zero HTTP failures :blobDab:

  19. @seldo Living life on extra hard difficulty :blobHelp:

    Hope the interviews go well and you recover from COVID soon!

  20. @joschi @trisha_gee It does, but with a large enough set of modules it is faster for us to parse a Git diff than it is for Gradle to do configuration, compilation avoidance and up-to-date checks.

  21. @trisha_gee Determining affected modules by scanning the diff of a PR and only testing those modules and ones that depend on them.

    Edit: to clarify, modules here mean Gradle modules, not JPMS ones.

  22. @raptag yeah that's definitely an improvement but given the fact that GitHub is heavily leaning into being an "AI powered platform for developers" I wouldn't be surprised if those went away in a not so distant future to solely provide an algorithmic feed :/

  23. @raptag I made an effort to like it but it's simply too ugly and too buggy right now for me.

    Release notification cards somehow meld into each other, the horizontal paddings on the lists are completely inconsistent, the list goes on.

  24. How is it 5:40 in the morning and I have had zero seconds of sleep...

  25. @Gankra hope you escape SemVer hell soon 🤞

    GitHub timeline snippet showing Gankra making a GitHub release with the tag name ' v0.3.64-prerelease.999.oh-yeah+FIGHT-me.IRL123' in a test repository
  26. @skade If you have enough cores everything is fast :blobSmileSweating:

  27. @gryzor @zsmb13 I would love for the admins to come out and set the record straight in that case. Until then I will assume Nik's history of acting in good faith is being upheld in this incident as well.

    Replying to source post

  28. @SylvieLorxu At least this year there won't be another YouTuber from my country teaching decidedly unskilled college students how to spam PRs for free merch

  29. Android Studio has started a new canary track carrying the Iguana name but there are no release notes to be found anywhere, what is going on?

    #AndroidDev

  30. huh I thought it was Friday already, clearly I can't wait for the week to be over...

  31. First year in a while that I haven't been using the #Android betas and the FOMO is setting in lol

  32. @yosh just saw that and while I appreciate that your position is much more reasonable, there is definitely a significant number of people who do hold the position I'm talking about.

    The cURL maintainer has repeatedly shared emails they've received asking them to do free work so billionaire organizations can check off regulatory and compliance requirements. I do not believe they're alone.

  33. @bodil Would not be Twitter if there wasn't more disappointment up for grabs.

  34. @bodil Twitter discourse has gotten pretty middling.

    FWIW, the tweets aren't gone and neither is the media. The HTTP t.co links are no longer linkified for whatever reason but the actual stuff is still accessible via Twitter's object storage with more direct URLs.

  35. @sheevink @yosh yeah I'm not a fan of all the onus of all this supposed accountability always being forced to the bottom of the chain. If anybody using my library needs to cough up an SBOM and finds something amiss, I literally never want to hear about it. It's their problem alone.

    There are a lot of people calling in ethical responsibility in the Serde situation in the absence of legal ones which also seems pretty moot since caring about build times is a perfectly reasonable choice.

  36. @Vlad @py I remember it from Twitter, I don't think it was ever put up on Py's blog.

    The gist of it is that when investigating a bug, you start a Slack thread (or anywhere you can keep thread-like information) and as you research and try things, keep noting them down. By the time you find the fix, you have a whole log of your process that as Py said, easily becomes the body of a blog post as well as becomes permanent record for future use.

    Replying to source post

  37. One upside of using SQLDelight (github.com/cashapp/sqldelight) everywhere is that it periodically uncovers pretty embarrassing gaps in my #SQL knowledge that usually get papered over by more complex ORMs 😩

    #AndroidDev