Community
Chat Logs
Monday, November 2, 2026 (America/New_York)
- * skillbot quit (Remote host closed the connection)
- * skillnews quit (Remote host closed the connection)
- * skillbot joined #primate
- * skillnews joined #primate
- * jreicher quit (Quit: In transit)
- * jreicher joined #primate
- liolocmorning !
- liolocIs there any way to undo migrations?
- bluelioloc: morning, hello
- bluelioloc: no, migrations are forwarding breaking by design
- blueforward breaking*
- bluethe way to deal with it is create another migration that fixes the problem
- bluesome orms/db systems do implement reversible migrations, but it's a can of worms: destructible migrations cannot be reserved, and non-destructible are easy to forward-fix. so reverting migrations doesn't gain you much
- bluereversed*
- liolocah i see
- blueso the question is, what is your use case here?
- blueit's better to accept migrations as historic evidence and forward-fix than messing with them
- liolocif we had a database that operated kind of like git does, with "commits" then it would probably be an easy fix, but there's probably good reasons why that doesn't exist
- liolocMy use case was a personal project where I added a migration but decided to revert those changes, nothing crazy
- liolocI don't mind adding another migration to revert
- blueyes, if you had a database that operated like git, it would be easy to 'revert' migrations. however, you have to remember that databases do not operate in a way; their write operations aren't all gated by acceptance, like commits here, and any user performing an action post-migration already has the expectation that that action is persisted, rather than lost when you 'revert' to the old stage
- bluedatabases do not operate in the same way*
- liolocah ok
- bluegit's graph is visible all the way, databases are generally not ledgers... and even so, the complexity of reconciliating a back-migration with already-persisted changes is too high, I'd say even impossible
- bluelioloc: primate 0.44 released, btw
- blueit's not a particularly big release like 0.45, but I wanted to `primate update` out already
- bluelike 0.43*
- blueI'm also adding more features to flog. `flog format` and `flog lint` will hopefully soon arrive, as a fast replacement to eslint. then we'll have `flog test` and `flog test --cover`
- bluefinally, `flog typecheck` is planned. it will ts-check your project, but only against modern ts syntax. so erasable syntax only, no namespaces, etc.
- bluegenerally speaking this will be replacing eslint + tsc + test runners (like proby)
- blueso flog will do everything except run your code
- bluehttps://primate-master.repopack.app/docs/testing
- blueexciting
- blue(upcoming in 0.45)
- bluedreamreal: ^^^ that might be interesting for your tests, as it doesn't require starting a primate app
- blueprimate is getting a visual website refresh in 0.45, https://pewtxngij-task-295.repopack.app/
- nevetPrimate — The Universal Web Framework
- nevet[ByteCode.News] CheerpJ 5 launches: Java 21 and Java 25 in the browser - https://bytecode.news/posts/2026/10/cheerpj-5-launches-java-21-and-java-25-in-the-browser
- liolochey blue, 0.44 is broken with a simple react app, created task https://repopack.com/project/primate/tasks/302 for it
- lioloci couldn't access the visual website refresh, sasy 404 page not found
- bluehi lioloc, looking at it
- bluelioloc: @primate/core@0.13.1 is published
- liolocnice thanks
- liolocconfirmed it works
- liolocthanks
- blueno worries, fixing on master + adding tests now
- bluefeed list
- nevetjustine.lol - https://justine.lol/rss.xml Node.js Blog - https://nodejs.org/en/feed/blog.xml Obsidian Blog - https://obsidian.md/feed.xml W3C - Blog - https://www.w3.org/blog/feed Jan van den Berg - https://j11g.com/feed WebAssembly News - https://webassembly.org/feed.xml James Shore - https://www.jamesshore.com/v2/feed Distant Province - https://distantprovince.substack.com/feed Ember.js Blog - https://blog.emberjs.com/rss.xml The Astro Blog - https://astro.build/rss.xml Entropic Thoughts - https://entropicthoughts.com/feed Loris Cro's Personal Website - https://kristoff.it/index.xml Flog releases - https://flog.run/changelog.rss zer0ver - https://0ver.org/atom.xml Simon Willison's Weblog - https://simonwillison.net/atom/everything/ Andrew Kelley - https://andrewkelley.me/rss.xml danluu.com - https://danluu.com/atom.xml Primate Blog - https://primate.run/blog.rss Diátaxis - https://diataxis.fr/atom.xml The JetBrains Blog - https://blog.jetbrains.com/feed/ purplesyringa's blog - https://purplesyringa.moe/blog/feed.rss Rust Blog - https://blog.rust-lang.org/feed.xml Read the Tea Leaves - https://nolanlawson.com/feed/ Adoptium Blog - https://adoptium.net/rss.xml FREEZE64.COM - https://freeze64.com/feed/ Terence Eden’s Blog - https://shkspr.mobi/blog/feed/atom Docker - https://www.docker.com/feed/ Feed - https://hollycummins.com/rss.xml bun.com - https://bun.com/rss.xml Alan Williamson - https://alan.is/feed/ Ziglang.org News - https://ziglang.org/news/index.xml Svelte blog - https://svelte.dev/blog/rss.xml CheerpJ - https://cheerpj.com/feed.xml Chicory Blog - https://chicory.dev/blog/rss.xml yosh's blog - https://www.unix.dog/~yosh/blog/feed.atom tRPC Blog - https://trpc.io/blog/rss.xml Eignex - https://eignex.com/feed.xml Typotheque - https://www.typotheque.com/rss Improbable Research - https://improbable.com/feed/ Alberto Denia's Writing - https://www.albertodenia.com/writing/feed Deciphering Glyph - https://blog.glyph.im/feeds/all.atom.xml cekrem.github.io - https://cekrem.github.io/index.xml Apache NetBeans - https://netbeans.apache.org/blogs/atom The PHP Foundation - https://thephp.foundation/atom.xml That Privacy Guy! Blog - https://www.thatprivacyguy.com/blog/feed.xml Help Net Security - https://www.helpnetsecurity.com/feed/ maevyn.codes - https://maevyn.codes/rss.xml David M. Lloyd - https://word-bits.flurg.com/rss.xml TypeScript - https://devblogs.microsoft.com/typescript/feed/ thoughts from the red planet - http://feeds.feedburner.com/ThoughtsFromTheRedPlanet Piotr's TechBlog - https://piotrminkowski.com/feed/ Vlad Mihalcea - https://vladmihalcea.com/feed/ Deno - https://deno.com/feed WebKit - https://webkit.org/feed Eclipse IDE and platform on eclipseide.org - https://eclipseide.org/index.xml Fernando Borretti - https://borretti.me/feed.xml Enigmastation.com - https://enigmastation.com/feed.xml JoshLong.com - https://api.joshlong.com/feed.xml OpenTelemetry - https://opentelemetry.io/index.xml MDN Blog - https://developer.mozilla.org/en-US/blog/rss.xml Random Notes and Cheat Sheets - https://www.kodewerk.com/feed.xml The Nuxt Blog - https://nuxt.com/blog/rss.xml Spring - https://spring.io/blog.atom Vercel News - https://vercel.com/atom Christian Rackerseder | Blog - https://www.echooff.dev/rss.xml Nushell - https://www.nushell.sh/atom.xml Angular Blog - Medium - https://blog.angular.dev/feed Paged Out! Magazine - https://pagedout.institute/rss.xml Vanish is... - https://vanish.is/feed.xml Open Liberty - https://openliberty.io/feed.xml Apache Fory™ Blog - https://fory.apache.org/blog/rss.xml Poimandres - https://pmnd.rs/feed.xml The Vue Point - https://blog.vuejs.org/feed.rss Next.js Blog - https://nextjs.org/feed.xml Biome | Blog - https://biomejs.dev/blog/rss.xml Mystic’s Substack - https://www.wecodefire.com/feed Nate Meyvis - https://www.natemeyvis.com/feed/ マリウス - https://xn--gckvb8fzb.com/index.xml Nature Reviews Computing - https://www.nature.com/nrcomput.rss Blog posts | RSS Feed - https://www.cockroachlabs.com/rss.xml OpenAPI Initiative - https://www.openapis.org/feed The Go Blog - https://go.dev/blog/feed.atom React Blog - https://react.dev/feed.xml ByteCode.News - https://bytecode.news/feed.xml Codeberg News - https://blog.codeberg.org/feeds/all.atom.xml DuckDB - https://duckdb.org/feed.xml Martin Fowler - https://martinfowler.com/feed.atom Terminal Trove - https://terminaltrove.com/blog.xml GitHub hosting -> Codeberg - https://discuss.cachyos.org/t/github-hosting-codeberg/33963.rss
- bluedreamreal: ^ this really should show the feeds for this channel, showing all feeds is... not useful
- bluearticles
- bluearticles blue
- bluenothing is working :(
- blue`!articles` without [user], should show my articles, I think
- bluewith user, show should user's articles
- blueshould show*
- bluedreamreal: and coolest feature of all: link my irc user to repopack user, link channel to project, then be able to do `+task [title] [text]`
- bluethis would allow quick prototyping of ideas
- * phaleth joined #primate
- bluehi phaleth
- bluehttps://primate-master.repopack.app/
- nevetPrimate — The Universal Web Framework
- bluenew website coming in 0.45
- phalethhi blue
- phalethnew primate website?
- blueyes
- bluealso testing, ORM translactions, changelog for primate, and API docs
- bluetesting is already done: https://primate-master.repopack.app/docs/testing
- blueprimate already uses it internally for its frontend tests, this bypasses starting a real server and occupying a port, pretty cool stuff
- phalethcolors are a bit different, margins are a bit off in some places but many that's fine
- blueya
- phalethmaybe that's fine*
- bluealso flog is gaining many new commands, `flog check types`, `flog check imports`, `flog checks deps`
- blueso it'll replace many tools, eslint/prettier/biome, tsc, npm, knip, and many more. anything aside from the runtime itself, essentially
- bluethis is the bun thing I think systemfault was complaining about. flog is basically bun but without the runtime, so it's only a package manager
- bluethis is the one thing flog will never be, a runtime. because it's in go. but that's fine, I don't want it to be a runtime
- bluedog will carry that torch
- phalethshould be separate yeah, make the container too big
- bluewell flog right now, on master, is 4.5mb
- bluethat's not very big at all. bun is like 90mb
- * skillbot quit (Remote host closed the connection)
- * skillnews quit (Read error: Connection reset by peer)
- * skillbot joined #primate
- * skillnews joined #primate
- phaleth9.8M on my disk
- blueflog-0.0.0-dev.1540bd0af077-linux-amd64.tar.gz 4.5MB
- blueI guess there's a lot of text inside? binaries are supposed to be compressed already I thought...
- blueanyway I'll see if I can make it smaller
- phalethunarchived
- blueya
- bluewell 9.8M is still ok. but I'll optimise once we get all the features
- blueI'm just trying to get rid of all this unreliable dumpster fire tooling that never works for me properly
- bluelike I install knip, then it shows a thousand errors where about 99.99% of them are false positive because it doesn't understand node private imports (#)
- bluedoes anyone actually test/use those tools?
- phalethno, people still use webpack
- blueomg webpack
- phalethit's always like yeah, dog shit slow, but it's battle tested and it works
- bluebattle tested my butt
- bluebut tbh, I delimit flog to *my* use cases. like for TS type checks, it's only modern syntax because I don't care about enums or interfaces or all that old garbage
- blueso it's just erasable syntax and no interface. it's anyway what the TS more or less recommends today
- bluesorry, no namespaces*
- blueinterfaces will be fully supported
- bluenamespaces was a module-like attempt of TS that predated proper ESM modules, I believe. so anyway you shouldn't be doing it anymore
- bluealso no type decorators and so on
- phalethI like TS enums
- bluethey don't degrade well at all
- phaleththey allow string values
- phalethlike Java enums, but Java enums look weird
- blueyou should just do `const ENUM = { 1: "HI", 2: "BYE"};`
- bluethat's as good an enum as you'll get
- phalethnamespaces in TS are a failure
- blueyes
- bluethat's why `flog check types` won't support htem
- blueso obviously if you need to check the world, you'd be still using tsc. but for modern projects, flog will be just as good
- bluethese are primate's devdeps right?
- blueright now:
- blue "@rcompat/test": "^0.22.0",
- blue "@types/node": "^26.6.4",
- blue "@typescript/native": "npm:typescript@^7.0.2",
- blue "apekit": "0.0.58",
- blue "eslint": "^10.12.0",
- blue "proby": "^0.27.0",
- blue "typescript": "npm:@typescript/typescript6@^6.0.2"
- blueafter flog is done, they'll all be removed
- blueit's all tooling upon tooling
- blueand I'm sick and tired of it
- bluejust need one good tool, (tm)
- phalethflog will replace tsc?
- blueyes
- phalethok
- phalethalso isn't proby your thing?
- blueyes
- blueproby is a test runner. flog will replace it
- blue@rcompat/test, that might stay there until primate can switch to std:test and dog
- phalethok, well concurrent test runner is better
- bluebut all those devdeps are heading to the dumpster yard
- bluethey're not really devdeps, they're tooling
- bluethe future is flog
- blueand not all those dumpster fires
- phalethso I've tried to use encore as just a JS runtime and it just wanted me to generate their project from their template and shit I was like hell no
- bluelol
- bluewhat is encore?
- phalethit's like you buy one product from apple and you have to dump everything else into trash to replace it with other apple crap
- phalethnot too sure, it's somebody having similar ideas to yours and just kinda replacing everything
- phalethbut they wanna lock you in you know, and maybe you will be accused of the same
- phalethyou will get comments like you do support TS, but it's not a TS, what it is
- blueI don't give a rat's butt
- phalethanyway, impossible to keep up with the standards
- blueI develop all my tools for myself, others are welcome to partake, if they don't want to, it's their prerogative to use their broken old tools
- blueI don't have to carry every backcompat nonsense the TS team feels obliged to. they have semver, they could decide to not do it
- bluepeople have apparently forgotten semver was designed so you can freely break stuff, like a bird
- blueinstead, people put themselves into self-inflicted prisons, not my problem
- phalethnah TS doesn't follow semver
- blueif TS doesn't follow semever, then why does it publish new majors? it should just switch to the firefox/chrome model
- blueso they kinda follow semver but also they don't
- bluealso your argument might hold for TS, but it certainly doesn't hold for node
- blueit does follow semver, but also it almost never breaks anything
- blueit could break stuff, it just doesn't want to
- blueif it broke stuff, people would be forced to update their old rotten software still using fs.write(callback) or whatever it is
- blueor stay on older, unsupported versions
- bluewhy do open source projects shift the burden of keeping your software uptodate, to themselves? and then complain they have no resource to maintain stuff?
- bluedreamreal: ^ honest question
- bluejust break forward, like linux does, and you don't need to maintain compat forever
- phalethpeople still use express and cjs, and still actually write more cjs than esm server side
- phalethand clanker keep supporting that crap too
- blueyeah but they do it because they can
- phalethso that means node cannot break, period
- bluewhen what's his name.. the esm guy, started publishing esm-only packages, people complained, but then they stopped complaining and continued to use his useful tools
- bluesindre*
- phalethso the thing is node is everywhere where non-browser JS is, I think even electron might use node, but dunno
- phalethand so if you wanna invade that space you just have to have everything well integrated, including editor support if you think people should write code
- phalethand in that integration sense also well automated, cause when people have to make effort they will not change their tools
- phalethit's always the same answer: PHP has already solved those problems
- phalethor whatever you put instead of PHP there
- phalethalso the biggest problem is mobile
- phalethmobile will be the biggest problem also for repopack, cause people like to sit on toilet when writing down ideas for whoever to work on, be it clankers
- phalethor lay down in bed or on a couch
- phalethlike who's having a desktop PC at home, only total nerds
- bluecan use repopack on normiephone too
- dreamrealblue: laziness
- bluedreamreal: ya but they complain about that one guy in nebraska maintaining it?
- blueidk, it just doesn't fit together
- dreamrealeveryone wants everything for free
- blueopen source should be something you do as a passion, you don't ned to maintain compat beyond what you have time to
- bluethat's what I do with primate. it's take it or leave it
- dreamreal"you gave it away, why are you not STILL giving it away"
- bluebecause I'm not obliged to :P
- bluebut no law of man or god
- blues/but/by/
- bluethe answer is simple to me. the question is why do so many open source maintainers put themselves in this bind
- dreamrealsure but people still want everyone ELSE to hold the burden
- blueI would never put myself in this bind
- dreamrealand since a lot of open source maintainers feel the same they bind themselves
- blueunless someone pays me to do primate, I'm not gonna support something I don't have the time, or more importantly, the passion, to support
- bluethis is something I will never compromise on
- dreamreal+1
- blueif I stop enjoying it, I'll stop doing it, I don't care
- dreamrealbut definitely fix all of MY issues
- blueyour issues are ALL fixed and more!
- phalethheh
- dreamrealI wasn't being serious anyway :D
- blueya, I know
- bluephaleth: did I tell you cit is compiling directly to assembly at&t, no libc?
- blueI think I also figured out ownership
- phalethyeah, so simple
- blueso clanker asked, how to deal with a scenario where you have a helper which needs to modify a resource, but also you want to give it back
- blueapparently rust has something like &mut or whatever craziness, like temporarily ownership or wahtever
- bluecit isn't having any of it, you transfer the ownership with !, and you give it back with ! again if you need to
- blueif you need to give it back AND provide more data, you give back a 2-tuple
- blue[!resource, bytes_read]
- blueit's all very simple, I don't know why rust complicates it so much
- blueit's a bit of go, a bit of rust, only the good parts
- blueit's very impotant to be able to read code in a linear way
- bluea bit like repopack tasks
- blueall those pointers are doing nothing good
- blueit's like all hidden sideeffects and whatnot. I think this is what zig is trying to avoid, but of course as we all know zig has become a dumpster fire
- blueso now we get zig simplicity, with rust-style ownership, and go returns
- blueall good and proper
- bluedreamreal: see how I actively engage in tikkun olam here?
- dreamrealI do, yes
- bluegood
- phalethblue: you should watch `Async Rust` from tsoding then you will remove rust completely :D
- phalethi can actually upload it for you if you want, it's a fun stream
- * phaleth quit (Quit: phaleth)