Chat Logs

  1. phalethsolidjs is doing some breaking changes https://github.com/solidjs/solid/blob/next/documentation/solid-2.0/MIGRATION.md
  2. bluephaleth: wow
  3. bluephaleth: https://repopack.com/project/primate/tasks/80#message-3046
  4. nevetRepopack 路 Sign in
  5. bluedid you configure the app to have another bundle for docs?
  6. bluealso I'll look look at the size increase
  7. phalethnope, another entrypoint has to be configured for each route?
  8. phalethlooks the size of @primate/core has been incresing, but it's not too bad given the trouble with SSE
  9. blueyes, phaleth; see https://repopack.com/project/primate/tasks/80#apps%2Fwebsite%2Fviews%2Fdocs%2Fdocs%2Fconfiguration.md
  10. nevetRepopack 路 Sign in
  11. phalethdoes that mean for primate website we have to define `groups: { docs: "/docs", guides: "/guides", blog: "/blog" }` and only then those routes will be split?
  12. blueyes; note that the splitting as it is currently will create full (not shared) bundles for all three. so browsing between those fences will load those full bundles instead of reusing share code. If we need a shared code functionality, then we need to extend this feature
  13. phalethah, ok, sounds good
  14. bluesonuds good as in, we need shared bundles, or you're ok with how it is? probably for primate website shared common code is better for perf
  15. phalethit depends on how big is the "core" of the primate website app
  16. blueI'll have clanker analyse.
  17. * blue clankering
  18. phalethor in other words if the chunk being split out is 40 kB of JS or more then it's worth splitting the route
  19. bluehm, that's hard to know in advance
  20. phalethyes, has to be tested on primate website, I think this scenario is ideal `groups: { app: "/", docs: "/docs", guides: "/guides"}`
  21. phalethif the homepage app JS bundle size drops significantly then the implementation is good
  22. bluegood goal definition
  23. bluephaleth: please check out primate#80 (you should get a bell)
  24. phalethalright
  25. bluethe last note is a bit clankered I feel "This would be a separate follow-up capability rather than an implicit change to #80's isolation contract.". I would implement everything as part of #80
  26. blueI wonder this feature done right in one go, not different tasks
  27. blueI want*
  28. phalethyeah, I can see what you are trying to do there
  29. phalethbut it's getting a bit hairy
  30. blueyou have a better direction?
  31. blueclanker came up with some overengineered conf object, I told him to simplify it
  32. phalethyeah, checking the route prefixes is fine, no need for other operations
  33. phalethI will just try it on primate website
  34. bluewell, the graphs thing is good to avoid having full bundles
  35. blueotherwise /, /docs and /guides will all be very big, you gain nothing
  36. bluegoal is for / to be small I think
  37. phalethyeah, I'm reading
  38. phalethtypescript is getting in the way, but that's because primate still uses pnpm which kinda ignores TS errors it seems
  39. blueyes
  40. phalethmind if I just fixed all those errors so we can get replace pnpm to make the development faster?
  41. bluesure
  42. phalethok, let me comment, I think it's not going well
  43. bluek
  44. phalethdone :)
  45. phalethlooks like for primate website itself current changes are not a big deal as the core of the JS bundle is just too big
  46. phalethbut if those changes are good for repopack after the bug I found is fixed then I think it's all good
  47. phalethshould I also respond to the `graphs` idea?
  48. blueyes
  49. phalethalright
  50. phalethclanker is always so confident :)
  51. phalethanyway, this change prolly won't do much for primate website even with the graphs feature, but worth a shot I guess
  52. blueya
  53. phalethalso it seems like the graphs feature won't be flawless, may as well abandon the idea
  54. bluewhy not? also dev clanker finished, review clanker is now checking
  55. blue(finished fix, I mean)
  56. bluebtw if you have issues with graphs feature, please post in tank
  57. bluetask*
  58. bluephaleth: pinged you
  59. phalethit just looks to me the rabbit hole got too deep, but I can test
  60. bluewe're not doing graphs yet, the fix was for the error you encountered
  61. blueok phaleth, looks now like it's working for you, yes?
  62. blueby the way, what do you mean with "full load"? it's supposed to load the additional assets
  63. phalethI will comment what I mean by full load
  64. bluek
  65. phalethok, commented
  66. bluegreat, I see you also asked for changes, so it's now in REVISION
  67. bluegreat
  68. phalethheh