Reinitialize NSRunningApplication at EVENT_HANDLER_APPLICATION_LAUNCHED

[donaldguy]

As per koekeishiya#923 (comment) this appears to resolve koekeishiya#920 crashes for me

(all the ones I have experienced so far anyway, there may be more like it)

I'm not asserting this is the correct way to handle this, but its a workaround at least

[donaldguy]

(^ Rebased on c35ef0a )

[xorpse]

In principle, since this fork started off as a workaround for issues not tackled in the main repository, I have no issue merging the fix if it works :).

However, I'd love to get to the bottom of the scripting additions issue. From your comments in #4 and koekeishiya#923, am I correct in assuming you've been able to get yabai running without issues on the M1 with scripting additions working using the bundled injector, not the Frida/standalone one?

[donaldguy]

However, I'd love to get to the bottom of the scripting additions issue. From your comments in #4 and koekeishiya#923, am I correct in assuming you've been able to get yabai running without issues on the M1 with scripting additions working using the bundled injector, not the Frida/standalone one?

Not quite. yabai starts and manages windows but is presenting

on init.

Running sudo yabai --install-sa fails silently with return of 2
Running sudo yabai --load-sa fails silently with return of 1

:(

maybe and maybe not of relevance, but to my surprise

pgrep -fl Dock.app
8221 /System/Library/CoreServices/Dock.app/Contents/MacOS/Dock
8224 /System/Library/CoreServices/Dock.app/Contents/XPCServices/com.apple.dock.extra.xpc/Contents/MacOS/com.apple.dock.extra
8225 /System/Library/CoreServices/Dock.app/Contents/XPCServices/com.apple.dock.external.extra.arm64.xpc/Contents/MacOS/com.apple.dock.external.extra.arm64

(XPC is a new one to me (though I see there is code in upstream yabai to filter out / ignore such processes ) - and I wonder if it doesn't offer a better approach to the current injection - but I imagine if it did, we'd have already gone down that path(?); maybe I've just never run pgrep -f Dock before (though that seems impossible) ; maybe also I just have some docklet running somehow for the first time)

but:

❯ git rev-parse HEAD
ed717687b2ed65e5db0bbb7bf6536b64f4f1af40

yabai on  canon [?] via 
❯ cd arm64-injector-standalone

yabai/arm64-injector-standalone on  canon [?]
❯ make clean
rm -f inject inject.o
rm -f sa_arm64e sa_arm64e.o sa_arm64e.h

yabai/arm64-injector-standalone on  canon [?]
❯ make
./extract.sh
clang -o inject inject.c -arch arm64e -std=c99 -Wall -DNDEBUG -O2 -fvisibility=hidden -mmacosx-version-min=10.13

yabai/arm64-injector-standalone on  canon [?]
❯ pgrep Dock
8221

yabai/arm64-injector-standalone on  canon [?]
❯ sudo ./inject 8221
Password:

yabai/arm64-injector-standalone on  canon [?] took 2s
❯ echo $?
0

but (re)starting yabai thereafter still gives

I haven't dug into the injector or the payload much at all, so idk if that implies that the above 0 return from inject.c's main is spurious - or if its successfully injecting a null (or malformed) payload but can't do the full/real one for some reason

[donaldguy]

Feel free to suggest things for me to look at / ask for me to report on some stuff - since you are without a 12.0 M1

[donaldguy]

also, possibly of relevance, I am not doing any code signing of any kind right now cause that bit wasn't clear to me. but I gather there may be relevant entitlements in the arm64e sitch

In particular, I'm not clear what the yabai-cert file in the existing sign Makefile target should be;

I saw something that looked relevant in my brief glance at a frida tutorial, but I gather we aren't doing that anymore?

(It's probs not relevant to this signature (which I imagine is either a self-signed bit or like a MDM cert?), but I am a registered apple developer for the time being)

[donaldguy]

Closing in favor of #7 (which maybe you don't want, if this is likely to go upstream)

but would love to still hear advice on playing more with the injector on Monterey