phaleththe idea of "Inter-procedural (whole program) escape analysis" sounds great
phalethanother good idea "The architecture separates CPU-bound work from I/O-bound work into different thread pools"
phalethactually that continues with "without requiring async/await syntax" which sounds great
phaleth"The compiler catches the dangerous patterns (use-after-free, null pointers, data races) without requiring you to annotate lifetimes or fight the borrow checker" also great
* phaleth quit (Quit: phaleth)
* phaleth joined #primate
blueHi phaleth
phalethhi blue
blueclanker is sleeping
bluewhat's up?
phalethlooking into inko
phalethno idea what's up with clanker, but inko seems like ideal lang for clankers
liolochey guys, can you use flog for monorepos?
liolocblue: ^
liolocnvm i see that it does (workspaces), would be nice if `flog init` had a flag for workspace maybe so that it creates a monorepo, but not sure if that makes sense?
liolocwhat do you think?
lioloclike `flog init --workspaces=apps,packages`
liolocwooooo nice new primate.run website !
lioloclooks sick
bluelioloc: sounds good, could you create a task on flog project?
bluewe use flog for primate and rcompat, which are both monorepos. flog is primarily developed for monorepos
liolocsounds good, I noticed that i had to run flog init inside the apps/<project> as well, this is normal right?
blueI believe flog init just creates a package.json
blueand potentially flog.lock
bluewhat is inko, phaleth?
bluehonestly any new language that forces you to do blue and red functions is something I won't adopt
blueasync should be integrated into the language, like in go
blueyou should look into cit, phaleth, that's the future
bluetbh even java doesn't have async
phaleththe syntax is not the problem, simply take your time to read the article I've linked above
phaleththere seems to be a lot of good ideas in inko's compiler that just take care of all the nerdy nonsense
phalethalso java is out of question for us anyway, too big of a memory hog
phalethor any JVM lang for that matter
bluethe point is, the answer's been at our door all along, it's c. we take c, we add the most minimal stuff atop, remove the atrocities like pointers, and voila, we're done
bluewe don't need all this newage nonsense
bluelook at cit, it's beautiful
phaleththat doesn't solve concurrency problems at all, which is the hardest part
phalethalso golang has go keyword, instead of async
phalethheh, how is that newage nonsense, it's just continuation of whatever also erlang and BEAM languages have been trying
bluebecause anything that uses or relies on rust is nonsense
bluerust is a complexity that should have never happened
bluecomplexity is one of the two final bosses of SE
blueI will not ever accept rust in my software stack
blueI'd take java before rust :P
phalethtoo late as a lot of what you use is written in rust, heh
bluethere are areas I do accept rust, like my terminal (alacritty) or wm (niri)
bluebut I will not run a product that relies on rust
bluethat's just too much of a gamble. I cannot run code I don't understand
phalethwell, using c++ for a compiler is also a problem
blueyes, which is why cit needs to selfhost ASAP
bluebut cit is just getting started, and dog is the only user
bluealso cit doesn't need libc as per your suggestion, which is kinda awesome
phalethyeah, but I don't think writing a new programming language makes sense either
bluedid you look at the ownership model of cit?
blueit's kinda cool
phalethand it's just lately everything is written in rust, cause people are tired of c's undefined behaviours and every c compiler behaving slightly different, every platform requiring platform specific code and so on
blueyeah I understand, but whatever faults c is, rust is definitely not the answer
phalethI don't have access to that, and also I'm not arguing for rust anyway
blueI thought zig would be the answer for a while, but then that happened
blueit's not like I want to create a new programming language
blueit's just no one is delivering
phalethyeah, but neither rust, c, zig or cit solve concurrency problems on at least two platforms
phalethmainly the two we are going to target
bluecit could solve that, we'd need to see how we could design it
bluewe have the chance to design a programming around our needs
bluewhich is the best way to design anything
phalethno, that is like a huge undertanking and clanker really has no chance
blueall those people designing things do them in a vacuum, that's why when you take them for a run, they never work like you need them
phaleththat is an infinite rabbit hole
bluecit is already working. and I'm slowly improving it, because it's not a high need situation, repopack is already working very well
phalethyeah, cause it's an infinitely complex problem to not only have the language right, but also have the tooling right
blueI'm hardly spending any time on it on the moment
phalethand by tooling I also mean standard lib
blueyes
bluemost of my time is spent on repopack and primate
blueI wish I had time to work on cit, but I don't
phalethjust don't worry about cit, node is running fine
phalethor dog or whatever
bluedog is the future :P
bluebut honestly, I'm not spending more than 1% of my time on any of this, so don't worry about it
phalethI mean don't worry about dog as well
bluemy time is currently I'd say, maybe 50% repopack, 40% primate, 10% flog
blueflog is mostly done for now, the linting/formatting is like a sideburner
blueand those 40% in primate are mostly requests coming from the repopack team
phalethflog just needs some more flags and it's done
bluelike transactions and so on
phalethrepopack requires shit ton of effort
blueyes
bluebut it can already outdo fly in many aspects
bluemost of the things I do with repopack, I can no longer do with fly
blueanyway, we're very close to a breakthrough with repopack chat
blueI think that's the moment people can 'get' it
phaleththere are still a lot of features missing compared to fly
blueright now it's like "what is this about"
phalethand then there is the e-commerce part completely non-existent
blueyes but you need investors for all of this
blueit's a huge undertaking, you gotta convince one
blueand you only convince people by showing them it's useful for them
phalethin such domain it's possible to sell unfinished product, fly is also still unfinished
blueyes, exactly
phalethanyway, I highly recommend to read the article as a whole, there's lot of complaints about rust there as well
liolocfor me repopack tasks only started making sense after I used it, and also having had experience contributing to traditional open source projects, but before using it I didn't get it
liolocFor an investor that isn't technical, that will be tricky