ByteCode.NewsSubmit
Use quotes for an exact word.

Subscribe

RSSPodcast
Log in

Chat Logs

  1. * phaleth joined #primate
  2. blueHi phaleth.
  3. bluephaleth: https://github.com/primate-run/primate/issues/286
  4. nevetpema: `p.loose.bigint` / `i64` / `u64` / `i128` / `u128` coerce numeric strings via `Number()`, losing precision above 2^53 · Issue #286 · primate-run/primate
  5. bluePrimate is commited to delivering in record time :P
  6. blueI used that report as a reason to introduce changelogs per package, which is kinda cool
  7. phalethhi blue
  8. phaletha research group led by Professor Yu Feng
  9. phalethlooks like they don't care how is that implemented
  10. bluehow what is implemented?
  11. phaleththe number overflow
  12. bluewell it's good we got their attention
  13. bluealso, we're the fastest in delivering fixes, less than an hour from the original report to patch on npm
  14. phalethyeah, but it's a lot of mess in the code https://repopack.com/project/primate/tasks/414#file-70c75a59428a18c31f4c3a38
  15. phalethonly a student would be worried about 2^53 overflow, practically that's not going to happen
  16. blueI see if I can move some of the complexity into rcompat
  17. phalethwell, would be nice if a lot of the JS clanker recently added to the website's app js bundle would go away
  18. bluewhat JS clanker?
  19. bluecan you create task?
  20. phalethI mean the JS that clanker recently added
  21. phalethnah, prolly waste of time, and clanker has no chance anyway
  22. bluebut where? what task?
  23. phalethall of them I guess, like each release adds more and more kBs
  24. phalethit used to be at around 71 kBs
  25. bluecan you create task so I can review this? just write it used to be 71kBs and now it's whatever, and I'll look into it and figure it out why
  26. bluealso I didn't update repopack.com yet to primate .45, still need to do it
  27. phalethdon't worry, it does not matter
  28. blueI think it does matter if it bothers you :{
  29. blue:P*
  30. phalethyeah, would be nice if there was some work done for repopack
  31. phalethjust saying that bloat is being added
  32. bluein primate or repopack?
  33. phalethI mean you've propably promised those 64 bit ints in pema or whatever, so that bored student chuckled over it
  34. phalethbut in the end who cares if somebody runs into a bug there
  35. phalethboth
  36. bluewell please create tasks, I will look into it, memory consumption is kinda good in repopack, it's been running for a while and it's at 288MB
  37. phalethbut I've already created other tasks that are more important
  38. bluek
  39. bluemajor chat update is coming, phaleth
  40. blueand btw you can use with free models :)
  41. phalethoh, hmm
  42. phalethlast time I've tried running Mistral on my laptop the laptop was dying
  43. bluewe're gonna get hosted agents soon
  44. bluethis will be a paid feature
  45. blueand I know dreamreal is gonna love it
  46. bluephaleth: https://repopack.com/project/primate/tasks/423
  47. blueman I REALLY need chat. I keep confusing all those opencode tmux sessions :(
  48. blueI have three sessions per project. and I have the projects: repopack, primate, rcompat, flog, dog/cit (and sometimes socialcloud). also planning on future projects, superarch and many others
  49. blueimagine how much confusion this causes
  50. bluewith chat now, you always are in the context of a project (this is a local change I haven't deployed) yet. going to the homepage redirects you to your first, in the future default project
  51. bluebeing always in the context of a project is immensely helpful
  52. blueand btw: I think this is why people use herdr
  53. phalethyeah, I think so too
  54. bluebecause they realised tmux with a ton of agent sessions is gonna be totally confusing
  55. bluebut repopack.com solves this so nicely. you always have a project context
  56. bluealso with hosted agents, I can restart the computer as often as I want. the git checkouts survive because they're not on my computer
  57. blueit's perfect
  58. bluedreamreal: ^
  59. blueI might be wrong, but I think we've hit gold here. only time will tell
  60. bluethree workflows:
  61. blue1) you host your agent on your computer and connect it to repopack.com via rp; that's how I intend to work now; rp manages the worktrees in ~/.local/rp/agent-workspaces
  62. blue2) you host your agent on a vps or otherwise a server you own, and use rp to connect. same thing like 1), but not on your computer
  63. blue3) you buy a hosted agent (essentially a vps or another abstraction) from repopack.com, you don't need to setup anything
  64. blue3) will be the most elegant way to work, but also you might ask yourself - how do I get previews of the things I'm working on?
  65. bluethat's why we have task previews, that's exactly the reason they exist
  66. blueit's a new way to do development
  67. blueradically different, I think, but also very interesting
  68. blueand also crucially, it means your computer can burn down and you don't lose anything
  69. blueand you can do development on a weak(er) laptop
  70. blueyour computer becomes a terminal to your development environment, via your browser and repopack.com
  71. phalethlot of people will pay clanker and be wanting to get clanker hooked up to the project easiest way possible
  72. blueyes
  73. phalethif you make it so that the person has to write commands in terminal then you've lost them
  74. bluewell ideally, they're all gonna work through repopack.com chat
  75. phalethI mean for hooking up clanker
  76. phalethlike to me the chat says some nonsense cause I have no clanker
  77. phalethit should shove down a guide how to hook up clanker to the user's throat
  78. blueyes, this is going to change with hosted agents
  79. phalethalso once they set it up it should make them feel like a winner
  80. bluewell with hosted agents, there's no setup involved othre than connecting to your provider
  81. blueand if you buy a hosted agents, you would still get the free models immediately
  82. phaleththe gamification aspect is important, people are way too screwed nowadays, especially the youngsters
  83. bluelol
  84. bluedreamreal: you use claude code, correct?
  85. blueI currently only have rp wire opencode sessions. I do need to extend chat to claude code and codex
  86. phalethand to seepdeek
  87. phalethoops
  88. blueseepdeek?
  89. phalethyeah, try saying that
  90. blueseepdeek!
  91. phaleththis is difficult to understand without an example https://docs.repopack.com/start#connect-a-project
  92. bluecan you edit it or create task to improve?
  93. blueyou don't have to work on the task, just create it
  94. blueI want to make it perfect but I need to know about things that don't work well
  95. bluejust create tasks, it's cheap
  96. phalethyeah, but clanker is just gonna write more walls of texts
  97. blueignore it, just create tasks, tasks are the cheapest currency we have
  98. bluethe walls of texts are something we can work against in AGENTS.md
  99. bluethat doesn't make creating tasks invaluable
  100. phalethso does `rp bootstrap ...` create a project?
  101. blueor well, "not valuable"
  102. phaleththe bootstrap term is confusing
  103. blueit doesn't
  104. blueit sets up a project by cloning it
  105. bluetry this
  106. phalethah, now I get it
  107. bluemkdir ttt && cd ttt && rp bootstrap primate
  108. blueit will clone the project into ttt/primate, and print instructions. it's mostly for agents
  109. blueif you run rp bootstrap INSIDE a repopack.com project, it will just print instructinos
  110. blueit detects it by looking at the .git/config and finding repopack.com origins
  111. phaleththat information is missing there
  112. blueya
  113. phalethclanker would just turn it into a wall of text
  114. bluecreate task, I will fix
  115. blueor add manually :P
  116. blueeither way you need a task for it
  117. phalethyeah, I will try to read the docs, prolly impossible
  118. bluechat is working REALLY well
  119. blueonce I have this deployed, you gotta try it out phaleth
  120. phalethok, anyway, I don't have a clanker
  121. bluethere are open models
  122. phalethlast time I messed around with LLMs was in 2023 and since then I forgot everything
  123. phalethI remember I used some python environment thing that was painful to work with
  124. phalethevery single project worked only on specific version of python
  125. blueyou just need opencode
  126. phalethisn't opencode's source closed?
  127. phaleth:)
  128. phalethI think it was anthropic who published sourcemaps of their CLI to npm package registry and then somebody found two vulnerabs in their shitty CLI thanks to that
  129. * phaleth looks up if garbage is still running on the VPS
  130. phalethnope, it doesn't
  131. blueopencode is open source
  132. bluehttps://github.com/anomalyco/opencode
  133. blueMIT
  134. phalethwow, it can be installed via pacman
  135. phaleththen it must be important
  136. blueya
  137. blueanyway try it, rp can connect to it and then you can use the chat :) I'll soon push out this amazing update
  138. bluealmost done with it
  139. phalethmy computer does not have a GPU and it's CPU is too old
  140. bluedon't worry about it, this is a T480
  141. phalethyeah, but you use a remote clanker
  142. blueI don't
  143. blueopencode is running on my computer currently
  144. blueI don't have an extra vps atm
  145. phalethby clanker I mean the LLM
  146. phalethI found this knob https://i.imageupload.app/0ca74ba34e780ae573b2.png but in rp CLI there isn't anything like `rp task merge .` right?
  147. blueYes. I think I removed `task merge` from the cli at some point. It should return, but I think it should always guarded
  148. blueClankers should not be able to autonomously push code onto master. So that will become an always-guarded action.
  149. phalethyeah, with a permission on the project
  150. bluehere, https://repopack.com/project/repopack/tasks/868
  151. bluewe don't need a permissoin on the project
  152. bluetask merge will always be guarded
  153. phalethbut people also use CLI so it's impossible to distinguish people from clankers
  154. blueyes. but the CLI isn't assumed to be 100% safe
  155. phalethand people should also be denied to take that action
  156. blueyes
  157. blueanyway can script your CLI
  158. bluethe last authority is the website
  159. blueguarded actions always require you manually approving on the ui
  160. bluethere are not a lot of them, but important things are guarded
  161. blues/anyway/anyone/
  162. phalethok, that task says: "equires a deliberate human confirmation in Repopack's browser for every invocation"
  163. phalethrequires*
  164. phalethbut how is that confirmation gonna look like? will it be enough that I'm logged in only via rp CLI?
  165. phalethcause now I have to be logged in on repopack site itself and hunt for the merge button everytime
  166. bluethe human confirmation is a popup that appears on the top right in the browser
  167. blueand yes, you wouldn;t need to hunt for the merge button
  168. phalethyeah, but in the browser after logged in to repopack website right?
  169. blueyes
  170. phalethwell, that's still the same then
  171. bluesame as what?
  172. phalethmakes no sense to implement that task at all
  173. bluewait, are we talking past it other? `rp task merge` would allow your clanker to request a merge
  174. phalethlike currently, in CLI you get link to task, open it and hit the merge button there after logging into the website
  175. bluethat would show up as a popup on the UI
  176. phalethyeah, but I'm in role of the clanker too :D
  177. bluethat spares you from looking for the task and scrolling down to the merge button
  178. blueif you're in the role of the clanker, that's useless for you, yes
  179. phaleththere is no looking up the task, the URL to the task is already in the terminal after `rp task submit . --push`
  180. blueyou wouldn't type `rp task merge`, then go to the website and confirm it
  181. blueyou'd just go directly to the website, I agree
  182. bluebut human-driving rp isn't the most common use case
  183. phalethok, then lets move on, no need to implement
  184. phaleththat means I should enable that knob I shown on the picture above again
  185. bluewell arguably it can move things a bit faster... let's imagine I'm in chat
  186. bluethen I tell the review clanker, review this task, and if approved, request a merge
  187. bluethen it can work in the backgronud and at some point I get a merge popup request
  188. blueI think there is some rationale for that task in the task description
  189. phalethyeah, that makes sense, but I was thinking it should just be available to anybody, but clankers will not have permission by default, or better yet that merge availability is off by default, which is already done by that knob I posted on the picture above
  190. blueso the probem is, you cannot distinguish who's driving rp, right?
  191. phalethI dunno, prolly not
  192. blueit's literally impossible to distinguish it. maybe in chat because we run in a sandbox. but even then we cheat with RP_CHAT_SANDBOX=1
  193. bluean env variable should not become an access gate
  194. blueI believe in repopack, we should have two levels of access. rp and browser. browser can do less actions, but has higher permissions
  195. bluerp can do anything, but important stuff is guarded, because it's scriptable and agent-runnable
  196. phalethok, how about doing something like `rp login` does
  197. phalethit opens up a web browser to confirm and only a human can confirm that
  198. bluerp login just gets you the token
  199. phalethyeah, I mean that behaviour
  200. blueessentially we have that behaviour already for guarded actions. except it doesn't open a new browser, it sends a push SSE to your existing session
  201. phalethso just show a merge button there, no need to logged into the website itself and click some notification or whatever
  202. bluebut remember that with rp login, you already need to be logged into the website anyway
  203. bluethis is more or less the exact same flow as push popups, just in a new tab
  204. phalethyeah, that's true
  205. blueI honestly don't even know why we're doing rp login in a new tab
  206. blueshould just do a push notif
  207. phalethso the client would just have to generate a one time token that the server can recognize as something belonging to the user
  208. blueI guess, yeah
  209. phalethsomething super obscure
  210. bluephaleth, there's a new feature on repopack.com if you're interested
  211. bluewhen you're in a project, click on the project switcher
  212. bluethe last line allows you to subscribe
  213. blueyou can choose which events you're interested in getting informed about
  214. phalethI see, looks nice
  215. bluethe transport mechansim is confirmed in your account settings: any combination of email, bells and push notifications
  216. phalethI mostly care about tagged releases
  217. bluecan you create repopack task for that?
  218. blue`add tagged releases as subscription option`
  219. phalethok
  220. bluethanks
  221. bluethanks, I got a push notif for that :)
  222. phalethheh
  223. * phaleth continues to read those walls of texts
  224. bluehttps://repopack.com/project/repopack/tasks/918
  225. blueis ok phaleth? it's not a wall of text..
  226. phalethyeah, that means clicking on that notif will open the release page I guess
  227. blueyes
  228. phalethok
  229. bluerp is kinda nice, don't you think
  230. bluecan really do cool things with it
  231. phalethyeah, would be nice if I could merge without logging into the website
  232. phaleththat's the only piece missing I guess
  233. bluethat would give rp a lot of power to merge things into master autonomously
  234. phalethalso I dunno if it's possible to list task with review state, but that prolly already is possible
  235. bluemaybe we should allow this per rp token, phaleth?
  236. blueso when you do `rp login`, we can could offer an option "Allow rp to merge tasks without confirmation, subject to project settings."
  237. phalethper project
  238. phalethoh, that login token, yeah
  239. blueok, per project is also fine by me
  240. bluemy thought about the login token was that you could have your personal rp login on your computer, to merge stuff
  241. blueand then your clanker login on your vps, with no merge capabilities
  242. bluewe could even define that that `rp login --headless` never gets merge unguarded merge capabilities
  243. phalethno clue how that works currently, but I can see on the repopack website there are like a lot of already expired tokens from the CLI
  244. phalethnot sure if they get autodeleted at some point from the db
  245. phalethyeah, that's a good idea
  246. phalethI noticed that `rp login --headless` in the docs, but the docs are not telling much I can understand
  247. blueit's mostly self explanatory but should still document
  248. bluerp login --headless will print a url, I think repopack.com/auth/cli/dervice or so
  249. blue[blue@vmd170320 ~]$ rp login --headless
  250. blueRepopack server: https://repopack.com
  251. blueOpen: https://repopack.com/auth/cli/device
  252. blueEnter code: REDACTED
  253. blueThis code expires in 10 minutes. Waiting for authorization...
  254. phalethand clankers just curls that url?
  255. blueit's a server-server token exchange, yes
  256. blueyou need to authorise the token
  257. phalethcool
  258. bluethis is for environments that have no browser
  259. bluelike a vps.
  260. phalethyeah, that's understandable form the docs, it's just no example there
  261. phalethfrom*
  262. blueok, could you create task to improve that? I will do. no walls of text, promised
  263. phalethnah, don't worry about it, I just need to read all repopack docs
  264. bluek
  265. phalethall of those functionalities are going to change several times anyway
  266. phalethat this stage docs are just a pure waste of time
  267. blueyes but remember we will add to AGENTS.md, to update docs on every task
  268. blueand remove as much internal docs as possible
  269. blueanything that's not 'secret' (whatever that is), should be public-facing
  270. phalethI think then in AGENTS.md there should be something about docs in admin area, if there is going to be anything internal ever documented
  271. blueYES
  272. bluecreate task?
  273. phalethat this point, no
  274. phaleththere is nothing internal, except in #392 which should just be closed already
  275. blue392 is a dumpster fire :)
  276. phalethwall of text, yeah
  277. phaleth`rp task list --status review` takes a few seconds to respond, is it not like one db query behind the scenes?
  278. blueyes
  279. blueshould be faster
  280. phaleth2 seconds https://upaste.de/raw/bER
  281. blueI'll be looking into it
  282. phalethok, great
  283. blueworking on perf now
  284. bluephaleth, for the record, at some point we're gonna need to develop our own agent to make things even more streamlined
  285. bluesupporting opencode and codex and claude code is nice, but having our integrated agent makes things a whole lot smoother
  286. phalethI thought the chat replaces opencode
  287. blueno. the chat is not an agent. the chat is a multiplexer over agent sessions
  288. blueit connects to your local opencode/codex/claude code
  289. bluean agent is a complex thing, needing instructions, context compaction, tools, and so on
  290. blueit's doable, but it's not what the chat or rp currently does
  291. phalethah, ok, so it could be doing that in the future, but looking at the size of the opencode project that is a huge undertaking
  292. blueyes and no. because our chat agent will be 100% headless, just a server, not a client like opencode/codex/claude code
  293. bluethat saves you a lot of complexity
  294. blueall those opencode/codex/claude code are both. they are a server and a client. and I think a lot of the complexity is in the client
  295. phalethso I don't know how a web browser can control a CLI tool, sounds like there is a security problem
  296. blueour chat replaces the client essentially
  297. blueit doesn't
  298. bluerp starts an outgoing connection
  299. bluethe web browser doesn't really control it, it's more like rp connects to repopack.com and they exchange info
  300. blueyou could say in essence it does control it, but that's not a real security problem since connection can only be initiated from rp
  301. blueinstalling or running rp does not allow any repopack.com user to control rp
  302. phalethyeah, the docs do not explain any of that
  303. blueyou need to explicitly run `rp agent start`, and enter the code shown to link it
  304. blueyeah, the docs are also unfinished :P
  305. phalethhttps://docs.repopack.com/guides/agent-chat
  306. phalethwall of text
  307. blueyes, we're gonna need to deal with that
  308. bluebelieve me, I'll handpick all the docs and make sure they're reduced and understandable
  309. blueand then I'll have dreamreal go over them and scream at me when he doesn't get something
  310. phalethno, don't worry about the docs, they are a pure waste of time
  311. bluegive it some time, all good things take a while
  312. phalethno no no, no time given to that
  313. bluethe docs are actually important because people have been asking for them
  314. bluedreamreal did, as did stagas
  315. blueI can't ignore that forever
  316. phalethyeah, but those current docs are useless
  317. bluenot entirely. they already have structure, and some useful information
  318. bluethe problem is they're gonna outdate fast
  319. phalethso it's like waste of token, and people can type in `rp agent --help`
  320. blueand for that we need to update AGENTS.md
  321. phalethtokens*
  322. blueya
  323. phalethyeah, but if you update rp CLI's help then clanker can figure things out
  324. phalethand so can people
  325. phalethno person is gonna read AGENTS.md
  326. blueya
  327. bluephaleth: you can now subscribe to new releases, too
  328. blueit's deployed :)
  329. phalethcool, thanks
  330. blueoh man, it's not marked on by default
  331. bluestupid clanker :(
  332. blueneed to fix that
  333. bluephaleth, are you on rp unstable?
  334. blueshould switch to the unstable channel for now, I'm not cutting releases for everything
  335. blue`rp update --channel unstable`
  336. blueI'm gonna deploy the perf improvement in a few mins
  337. phalethI think stable but it's working well so far
  338. blueok
  339. phalethok, I could test that then
  340. blueyou will need the latest version for the chat. I might cut a new rp version. but for now you can switch to unstable
  341. phalethI have no local LLM
  342. phalethbut maybe later I can try finding some
  343. phalethever since hugging face requires a log in I just don't bother with LLMs anymore
  344. phalethyeah, clanker is stupid cause of coding this way https://repopack.com/project/repopack/tasks/919
  345. phalethI mean, by making the problem in the first place
  346. phalethwell, whatever it wrote there looks like a bloat
  347. phalethah, so clanker added a bunch of flags to the CLI it seems
  348. phalethor query options
  349. phalethon task stacks page in docs there is this: "the stack merges from parent to child" and later "An unmerged parent cannot be closed while open children remain"
  350. phaleththat sounds contradictory
  351. phaleththe first statements is most likely wrong, or at least I'd think so
  352. blueya
  353. bluethe problem is phaleth, before it was filtering in the cli, now on the server, that's faster
  354. blueanyway, I merged 919, gonna redeploy
  355. phalethyeah, figured
  356. bluephaleth: try now
  357. bluerp task list --status review
  358. blueis much, much faster
  359. blueyou need `rp update` first and to be on the unstable channel
  360. phalethyeah, trying to update
  361. blue`rp update --channel unstable`
  362. phalethmuch better, yeah https://upaste.de/raw/eA5
  363. blueyup
  364. * phaleth quit (Quit: phaleth)