ByteCode.NewsSubmit
Use quotes for an exact word.

Subscribe

RSSPodcast
Log in

Community

Chat Logs

Thursday, November 5, 2026 (America/New_York)

  1. blueWB jreicher.
  2. jreicherHowdy
  3. bluejreicher: Kotlin WASM support is coming in Primate 0.45. With a bit of luck, we're gonna get hotreloading baked in as well.
  4. blueIf this pilot proves successful, I'm gonna target Java WASM for 0.46. I haven't added new backends in a while time
  5. blueJust to be clear what I mean with hot-reloading. When you change routes in Primate, during development, it rebuilds the running server and then replaces it. So you don't need to restart the server to see any changes you make to routes or backend logic.
  6. blueWith the wasm boundary, that's a bit more complicated: you need to rebuild .wasm files, too, but ideally only those that are affected. This is supported in Go, Ruby and Python for Primate. The Kotlin backend will likely create one big .wasm file, which might make hotreloading a bit slower at first. I might possibly have to switch to compiling many small files for dev, and one monolith for production.
  7. blueBut the effects are pretty cool: you edit a .kt file, primate refreshes the browser and you can use the changed backend
  8. bluedreamreal: I *LOVE* that I can write something like this: "Q: now that repopack is migrated completely to `key`, we can inform flog#225 to take action appropriately, correct?", and clanker GETS it.
  9. dreamrealyeah, they're getting better about context
  10. bluedreamreal: also, it REALLY loves saying "failed closed". So much so that I started adopting it when telling it to review stuff, as in, "the state of the task must be left in a conclusive state"
  11. blueI'm totally misusing the term there, but so does it, and I so don't care
  12. blue"Whilst you're reviewing: a few things. FINDINGS.md and other results from the GNG study must disappear. Also, the task must stack upon the docs reorg at 356, to be its tip. At its current state, it must fail closed at REVISION."
  13. bluedreamreal: ^ is this parsable to you as a human?
  14. dreamrealNot to me without enough context, on first read, no, but i don't think that matters
  15. bluek, good to know
  16. bluethe context is: the kotlin task started as a GO-NO-GO study, before transitioning into full impl. And '356' is just another docs reorganisation task that moved around a few docs that the Kotlin task would also be changing, hence the restacking. But clanker GETS it, because context!
  17. dreamreal*nod*
  18. nevet[Node.js Blog] Notable V8 changes from Node.js 26 to 27 (Alpha) - https://nodejs.org/en/blog/announcements/notable-v8-changes-from-node-js-26-to-27-alpha
  19. systemfaultOh, Iterator#zip, what a surprise...
  20. systemfaultTold Voldemort 3+ years ago that it would be a worthwhile addition...
  21. bluesystemfault: and did Voldemort listen? of course not
  22. systemfaultblue: I actually took 2 minutes to check the TC39 proposals to see if he was the one behind it.
  23. systemfaultOf course not.
  24. bluethat guy lost it when he started banning people from #js for not being part of his crazy progressive agendas
  25. blueI remember talking to him in private about `master` vs `main`. He acknowledge the move was stupid, but said "if one black guy feels bad about it, it's good enough for me"
  26. bluewhat an idiot, by God
  27. bluegood riddance that I don't have to interact with him any longer
  28. systemfaultI'm "generally" more against performative progressivism than progressivism itself...
  29. systemfaultI haven't gone back to his channel since the day he kicked me... from both the node and js channels
  30. blueyeah, terrible :/
  31. blueI found it a certain irony that the guy who advocates for keeping backcompat for nearly forever (in his packages, he supports node 0.4!), has no qualms about a git change that broke everyone's flows...
  32. bluebut of course, blue tribe, no brain, what I can say
  33. nevet[Node.js Blog] Notable V8 changes from Node.js 26 to 27 (Alpha) - https://nodejs.org/en/blog/announcements/notable-v8-changes-from-node-js-26-to-27-alpha
  34. bluedreamreal: why did nevet post it again? or did the node people mess up something in their feed?
  35. dreamrealYeah, it showed up as a new entry somehow
  36. dreamrealYou would be surprised - or maybe not - at how bad people are at RSS
  37. dreamreali had to do a whole update to help sanitize feeds that don't do base urls properly, or use *internal urls* - one site has localhost:3000 as the host/port for their PUBLIC FEED
  38. dreamrealAnd yes, I've tried to tell them
  39. blueouch
  40. bluethis is kinda bad though, RSS is a wonderful technologuy
  41. bluetechnology*
  42. dreamrealit's a good idea with a set of atrocious specs but eh
  43. bluedreamreal, clanker is really struggling with me demanidng that kotlin hot reload should be <1s
  44. blueit's currently apparently at 5-7s. am I asking for too much?
  45. blueidk it just doesn't appear a very nice experience to wait 5 seconds for your app to reload
  46. dreamrealDepends on the codebase size, the compiler is not super-fast
  47. bluewith Go/Python/Ruby it's immediate
  48. blueI don't know that it should take that long with Kotlin
  49. blue"It is not WASM initialization alone: standalone initialization measured roughly 7–21 milliseconds. A warmed, unchanged Gradle invocation previously took 1.84 seconds."
  50. bluehm
  51. blue1.84 seconds??????????????????????
  52. bluethis is like, 20 lines of code
  53. blueit should take 1.84 milliseconds :(
  54. dreamrealyes. Are you using build.gradle.kts or build.gradle?
  55. bluebuild.gradle.kts
  56. * dreamreal is not esp a fan of gradle, but it may be the right solution here
  57. bluehttps://repopack.com/project/primate/tasks/350
  58. dreamrealso you're invoking the compiler chain *twice*
  59. dreamrealonce to verify the build, once to RUN it
  60. bluedamn
  61. blue"There’s a real concern, but “compiling twice” is misleading.
  62. bluebuild.gradle.kts is Kotlin code, so Gradle must compile the build script—but normally caches that compilation. It does not mean every edit compiles your routes twice."
  63. dreamrealyes, but it has to *check*
  64. dreamrealit will cache if it can
  65. blueBefore switching to Groovy or replacing Gradle, #350 should measure:
  66. blue1. Where that no-op invocation spends its time.
  67. blue2. Whether Gradle’s configuration cache materially reduces it.
  68. blue3. Whether development is unnecessarily using expensive production compilation/link settings.
  69. blueGroovy?????
  70. * blue is losing it
  71. dreamrealpatience, young padawan
  72. dreamrealproblems are things to be solved, not endured
  73. blueI have no idea out of which butt it pulled that one, but it seems concerning
  74. dreamrealbe a programmer
  75. * blue patiences
  76. dreamrealbuild.gradle = groovy, build.gradle.kts = kotlin
  77. bluegroovy as in the apache groovy we talked about?
  78. dreamrealthe groovy and kotlin interfaces are *materially* the same, but kotlin gives you typesafety in your build, etc
  79. blueoh
  80. * blue doesn't know a lot, very small
  81. dreamrealyes, gralde is the PRIMARY place people still write groovy
  82. bluewhy not use makefiles?
  83. dreamrealas far as I know, at least
  84. dreamrealbecause invoking the entire build chain via make is an utter drag
  85. blueI hear makefiles solve everything, if you only allow yourself to have a psychosis every time you look at them
  86. dreamrealtheer you go
  87. dreamrealif you're okay with psychosis, make or cmake or just win
  88. * blue is def not okay with psychosis
  89. * blue leaves psychosis to the clankers
  90. dreamrealI actually use just+[maven|gradle|make] quite a bit myself: just gives me broad profiles for invoking the build tools the way I want
  91. dreamrealjust I have stuff like `just mr-watch 1117` to watch the gitlab CI/CD run, or `just coverage-view` to run the coverage profile and then open the summary pages, etc
  92. bluewell, sol 5.6 is busy profiling kotlin wasm builds
  93. dreamrealmy team has those files (they're in teh repos) but don't use them, so I tend to run circles around them even in basic tasks
  94. blueI see. just seems cool, I never had need for it though, I think
  95. dreamrealthey'll reload the gitlab CI/CD pages on their own, and I'm like "... why"
  96. dreamrealit's a lot like make with better syntax
  97. dreamrealyou could do everything just does WITH make
  98. blueI feel like reinventing make isn't a really good pattern, given uh, the aforementioned psychosis angle
  99. dreamrealwell, okay: but look at how make handles propagation. just isn't trying to BE make, it's trying to be make for the *absurdly* common cases
  100. dreamrealif you need advanced substitution patterns, etc., make is still better and everyone knows it and expects how it works
  101. bluemy point is, that this is a broader pattern. scripting languages often feel like a bandaid for building stuff. that's something I liked about zig: you use the language itself to build stuff
  102. dreamrealsure
  103. blueI do wish the zig people had thought this through and made zig equally capable of building anything, but they mostly focus on zig/c
  104. blueWASM executable generation/linking 2.31–2.63s
  105. nevet[Flog releases] Flog 0.1.14 - https://repopack.com/api/v0/projects/flog/releases/products/flog/versions/0.1.14
  106. nevet[Flog releases] Flog 0.1.13 - https://repopack.com/api/v0/projects/flog/releases/products/flog/versions/0.1.13
  107. nevet[Flog releases] Flog 0.1.12 - https://repopack.com/api/v0/projects/flog/releases/products/flog/versions/0.1.12
  108. bluesystemfault: did gurrr or someone complain about that meme?
  109. systemfaultThe one on the TS channel? No. I did.
  110. blueya
  111. systemfaultI would have said the same has it been a pro-democrat one.
  112. blueya, I get the point, you're trying to keep it professional
  113. blueit's a pity our leftist friends don't give us the same courtsey...
  114. systemfaultI'm fine with political "discussions" TBH as long as people remain civil.
  115. systemfaultBut I'm not sure about just posting memes
  116. blueyeah I totally get, this is no criticism
  117. blueI noticed the schwifty guy kind of got quiet after you don't him you don't think the US are particular great either :P
  118. blueyou told him*
  119. blue(which is true, btw)
  120. bluein other news... I'm 42 today
  121. bluenot getting any better I fear, the rest :P
  122. systemfaultOh, happy B day
  123. bluethank you :)
  124. systemfaultYeah, I'm now 44... we're getting old
  125. blueyeah :/ I don't *feel* old, it's a weird thing. In my mind I'm still in my 20s...
  126. jreicherOld people give being old a bad name
  127. blueheh
  128. bluebut my wife got me a lego set of the Space Shuttle
  129. blueI'll probably spend most of tomorrow assembling it, I'm a sucker for these things
  130. jreicher12 year old me is really jealous right now. Can I come round to your house?
  131. bluesure jreicher
  132. bluesystemfault: import defer * as X is ugly, but it is exciting because it solves the circular dependency problem...
  133. systemfaultYeah, looks useful
  134. systemfaultI didn't know it was a proposal until now
  135. blueme neither...
  136. bluebut it solves the User imports Post (hasMany) imports User (hasMany - author)
  137. bluehasOne - author*