The new memory model is cool but ```BUILD SUCCESSFUL in 24m 37s 56 actionable tasks: 47 executed, 9 ...

# ios

[ribesg]

The new memory model is cool but

BUILD SUCCESSFUL in 24m 37s
56 actionable tasks: 47 executed, 9 up-to-date
18:31:05: Execution finished 'build'.

[ribesg]

That’s the compilation time of both the debug & release fat frameworks (arm64 & x64) of a 37kloc pure Kotlin iOS app with various KMP and cocoapods dependencies, so it’s a lot and not necessarily a task (build “all”) I really have to execute

[ribesg]

But still, in “normal usage”, so when I just want to launch my app in a simulator and it builds just the right x64 debug framework, it can take from 3 to 6min when the mac isn’t too overloaded, while it took from 1.5 to 3min before. Hopefully it gets better with time and with cache re-enabled soon

[dazza5000]

it's probably still shorter than writing the code twice or more though? assuming you are using kmp with 2+ platforms

[dazza5000]

at least once you get everything setup and rolling

[ribesg]

And what’s your point exactly?

[dazza5000]

that a longer build time is sometimes quicker than writing the same implementation in multiple languages

[ribesg]

This is an iOS app written entirely in Kotlin, not a multiplatform app (it’s not on android at all) and we’re kinda force to use the NMM because it’s being pushed/used by libraries, and using the NMM doubles/triples build time.

Now how does your answer add anything to anything, I don’t understand

[dazza5000]

I apologize for communicating with you. Have a good day.