Hammerspoon: Problems after coroutine & modernise hs.application/window/uielement merge
@cmsj & @asmagill - I have now merged the latest Hammerspoon master into CommandPost-App. The first issue I've come across is:
2020-04-08 17:21:54: *** ERROR: ...ost.app/Contents/Resources/extensions/hs/window/init.lua:947: attempt to index a nil value (global 'uielement')
stack traceback:
...ost.app/Contents/Resources/extensions/hs/window/init.lua:947: in metamethod '__index'
...rishocking/Github/CommandPost/src/extensions/cp/init.lua:40: in main chunk
[C]: in function 'xpcall'
...Github/CommandPost/src/extensions/hs/_coresetup/init.lua:794: in function 'hs._coresetup.setup'
(...tail calls...)
latenitefilms
All 44 comments
We've also seemed to have lost hs.window.timeout().
latenitefilms
on 8 Apr 2020
Another one:
2020-04-08 17:33:18: ********
2020-04-08 17:33:18: 17:33:18 ERROR: LuaSkin: hs.timer callback error: ...ost.app/Contents/Resources/extensions/hs/window/init.lua:515: ERROR: incorrect type 'no value' for argument 2 (expected boolean)
stack traceback:
[C]: in method '_bringtofront'
...ost.app/Contents/Resources/extensions/hs/window/init.lua:515: in method 'focus'
...mandPost/src/plugins/finalcutpro/fullscreen/dockicon.lua:53: in function <...mandPost/src/plugins/finalcutpro/fullscreen/dockicon.lua:52>
2020-04-08 17:33:18: ********
@cmsj, the current branch includes changes to the underlying userdata for hs.window and hs.application, doesn't it?
Assuming this is so, since I'm seeing breakage with some of my own stuff that uses hs._asm.axuieleemnt, @latenitefilms I'm going to need to update axuielement, since it makes assumptions about the structures behind the userdata for these elements... it was easier than requiring hs._asm.axuielement to be recompiled everytime Hammerspoon was rebuilt.
I'll take a closer look at it later today.
asmagill
on 8 Apr 2020
@asmagill - Yes, it does - although I haven't spotted any hs._asm.axuielement errors yet.
Some other minor @stickler-ci things:
latenitefilms
on 8 Apr 2020
@asmagill - I'm wondering if now is a good time to just merge hs._asm.axuielement into Hammerspoon's master, either undocumented, or clearly documented that it's in beta and subject to change?
latenitefilms
on 8 Apr 2020
@asmagill it does, yeah. At the moment it doesn't do the equivalent of https://github.com/Hammerspoon/hammerspoon/blob/0.9.78/extensions/application/internal.m#L1257 and I'm not entirely sure if we want/need that. I thought we had that so that hs.application/window objects could be treated like hs.uielement objects, but if I understand the old code correctly, all it actually adds is these methods: https://github.com/Hammerspoon/hammerspoon/blob/0.9.78/extensions/uielement/internal.m#L388 and I think I'd rather re-implement those for each module than have them be weirdly intertwined like they were before.
cmsj
on 8 Apr 2020
Awesome, thanks @cmsj ! That's fixed a bunch of errors!
One error I'm now seeing a lot of is hs.application.frontmostApplication() returning nil, when it shouldn't ever really return nil - any ideas?
latenitefilms
on 8 Apr 2020
@latenitefilms could you merge b079110d into CP? Should help us narrow down where the nil is coming from, since there are multiple places it can happen.
cmsj
on 8 Apr 2020
Talking out loud... here's what the old code used to be:
/// hs.application.frontmostApplication() -> hs.application object
/// Function
/// Returns the application object for the frontmost (active) application. This is the application which currently receives input events.
///
/// Parameters:
/// * None
///
/// Returns:
/// * An hs.application object
static int application_frontmostapplication(lua_State* L) {
NSRunningApplication* runningApp = [[NSWorkspace sharedWorkspace] frontmostApplication];
if (runningApp) {
if (!new_application(L, [runningApp processIdentifier])) {
lua_pushnil(L);
}
} else {
lua_pushnil(L);
}
return 1;
}
Here's what the new code is:
+(HSapplication *)frontmostApplication {
HSapplication *frontmostApp = nil;
NSRunningApplication *runningApp = [[NSWorkspace sharedWorkspace] frontmostApplication];
if (runningApp) {
frontmostApp = [HSapplication applicationForNSRunningApplication:runningApp];
}
return frontmostApp;
}
...so it seems like the original hs.application.frontmostApplication() can technically return nil, but I never saw it happen. With this new code, I'm seeing it happen a lot, because I never allowed for it to return nil in any of our functions.
...just saw your last post. Will compile b079110 into CommandPost-App now. Stay tuned.
latenitefilms
on 8 Apr 2020
Will keep playing, but just spotted:
2020-04-08 22:46:34: ********
2020-04-08 22:46:34: 22:46:34 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-08 22:46:34: ********
2020-04-08 22:46:36: ********
2020-04-08 22:46:36: 22:46:36 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-08 22:46:36: ********
@cmsj - Ok, so all the errors seem to be the same now:
2020-04-08 22:46:34: ********
2020-04-08 22:46:34: 22:46:34 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-08 22:46:34: ********
2020-04-08 22:46:36: ********
2020-04-08 22:46:36: 22:46:36 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-08 22:46:36: ********
2020-04-08 22:49:11: ********
2020-04-08 22:49:11: 22:49:11 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-08 22:49:11: ********
2020-04-08 22:49:11: ********
2020-04-08 22:49:11: 22:49:11 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-08 22:49:11: ********
2020-04-08 22:49:23: ********
2020-04-08 22:49:23: 22:49:23 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-08 22:49:23: ********
2020-04-08 22:49:23: ********
2020-04-08 22:49:23: 22:49:23 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-08 22:49:23: ********
@asmagill - Also, you're right, your fancy coroutine getMenuItems() function is no longer working using the latest Hammerspoon Master branch.
latenitefilms
on 8 Apr 2020
@latenitefilms It's strange - we're definitely bouncing between NSRunningApplication and a pid more than is needed, and it's certainly the case that not all NSRunningApplication's have a pid, but the old code shouldn't have worked for them either.
I just added 9c9ad75b which might at least point us in the direction of which application is considered frontmost, but I'm currently at a loss for why this is failing now when it wasn't before.
cmsj
on 8 Apr 2020
can you give me more info about the corouting getMenuItems() failure? I did make sure that all of the menu tests (including the "async" one) were passing, but perhaps I broke something more fundamental in the merge.
cmsj
on 8 Apr 2020
Will test out 9c9ad75 now.
Here's the code I'm using, however it might not mean much to you, so probably best to refer to @asmagill's original code here. Keep in mind though that it's using hs._asm.axuielement - so it's probably a bug in there, rather than in your changes.
latenitefilms
on 8 Apr 2020
@cmsj - I'm not seeing any new errors messages, same errors:
2020-04-08 23:07:35: ********
2020-04-08 23:07:35: 23:07:35 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-08 23:07:35: ********
2020-04-08 23:07:36: ********
2020-04-08 23:07:36: 23:07:36 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-08 23:07:36: ********
2020-04-08 23:07:46: ********
2020-04-08 23:07:46: 23:07:46 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-08 23:07:46: ********
2020-04-08 23:07:47: ********
2020-04-08 23:07:47: 23:07:47 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-08 23:07:47: ********
@asmagill since we seem to have gone into a little nexus of significant refactoring, how do you feel about us doing a push to integrate axuielement? :D
cmsj
on 8 Apr 2020
@cmsj - I'm going to try this code, to see if I can work out which app is failing:
+(HSapplication *)frontmostApplication {
LuaSkin *skin = [LuaSkin sharedWithState:NULL];
HSapplication *frontmostApp = nil;
NSRunningApplication *runningApp = [[NSWorkspace sharedWorkspace] frontmostApplication];
[skin logError:[NSString stringWithFormat:@"HSapplication::frontmostApplication failed for app: %@", runningApp.localizedName]];
if (runningApp) {
frontmostApp = [HSapplication applicationForNSRunningApplication:runningApp];
if (!frontmostApp) {
[skin logError:[NSString stringWithFormat:@"HSapplication::frontmostApplication failed for app: %@", runningApp.localizedName]];
}
} else {
[skin logError:@"Unable to fetch frontmost application"];
}
return frontmostApp;
}
@cmsj - Ok, so it seems to be failing when switching from another app to CommandPost. Could CommandPost (or Hammerspoon) be blocking itself somehow?
2020-04-08 23:17:30: ********
2020-04-08 23:17:30: 23:17:30 ERROR: LuaSkin: HSapplication::frontmostApplication failed for app: Safari
2020-04-08 23:17:30: ********
2020-04-08 23:17:30: ********
2020-04-08 23:17:30: 23:17:30 ERROR: LuaSkin: HSapplication::frontmostApplication failed for app: Safari
2020-04-08 23:17:30: ********
2020-04-08 23:17:30: ********
2020-04-08 23:17:30: 23:17:30 ERROR: LuaSkin: HSapplication::frontmostApplication failed for app: Safari
2020-04-08 23:17:30: ********
2020-04-08 23:17:35: -- Loading extension: hs.application
2020-04-08 23:17:35: ********
2020-04-08 23:17:35: 23:17:35 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-08 23:17:35: ********
2020-04-08 23:17:36: ********
2020-04-08 23:17:36: 23:17:36 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-08 23:17:36: ********
2020-04-08 23:17:36: ********
2020-04-08 23:17:36: 23:17:36 ERROR: LuaSkin: HSapplication::frontmostApplication failed for app: CommandPost
2020-04-08 23:17:36: ********
2020-04-08 23:17:36: ********
2020-04-08 23:17:36: 23:17:36 ERROR: LuaSkin: HSapplication::frontmostApplication failed for app: CommandPost
2020-04-08 23:17:36: ********
2020-04-08 23:17:36: ********
2020-04-08 23:17:36: 23:17:36 ERROR: LuaSkin: HSapplication::frontmostApplication failed for app: CommandPost
2020-04-08 23:17:36: ********
Ummm, just got different errors this time:
2020-04-09 00:09:33: ********
2020-04-09 00:09:33: 00:09:33 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: 24457
2020-04-09 00:09:33: ********
2020-04-09 00:09:33: ********
2020-04-09 00:09:33: 00:09:33 ERROR: LuaSkin: hs.application.watcher callback: ...ing/Github/CommandPost/src/extensions/cp/ui/notifier.lua:64: attempt to index a nil value (local 'app')
stack traceback:
...ing/Github/CommandPost/src/extensions/cp/ui/notifier.lua:64: in function <...ing/Github/CommandPost/src/extensions/cp/ui/notifier.lua:63>
2020-04-09 00:09:33: ********
2020-04-09 00:09:33: ********
2020-04-09 00:09:33: 00:09:33 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: 24457
2020-04-09 00:09:33: ********
2020-04-09 00:09:33: ********
2020-04-09 00:09:33: 00:09:33 ERROR: LuaSkin: hs.application.watcher callback: ...hrishocking/Github/CommandPost/src/extensions/cp/app.lua:1241: attempt to index a nil value (local 'hsApp')
stack traceback:
...hrishocking/Github/CommandPost/src/extensions/cp/app.lua:1241: in function <...hrishocking/Github/CommandPost/src/extensions/cp/app.lua:1240>
2020-04-09 00:09:33: ********
2020-04-09 00:09:33: ********
2020-04-09 00:09:33: 00:09:33 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: 24457
2020-04-09 00:09:33: ********
2020-04-09 00:09:33: ********
2020-04-09 00:09:33: 00:09:33 ERROR: LuaSkin: hs.application.watcher callback: .../Github/CommandPost/src/plugins/core/menu/menuaction.lua:156: attempt to index a nil value (local 'app')
stack traceback:
.../Github/CommandPost/src/plugins/core/menu/menuaction.lua:156: in function <.../Github/CommandPost/src/plugins/core/menu/menuaction.lua:141>
2020-04-09 00:09:33: ********
2020-04-09 00:09:33: ********
2020-04-09 00:09:33: 00:09:33 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: 24457
2020-04-09 00:09:33: ********
2020-04-09 00:09:33: ********
2020-04-09 00:09:33: 00:09:33 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: 24457
2020-04-09 00:09:33: ********
2020-04-09 00:10:09: -- Loading extension: hs.application
> hs.application.applicationForPID(24457)
2020-04-09 00:10:15: ********
2020-04-09 00:10:15: 00:10:15 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: 24457
2020-04-09 00:10:15: ********
nil
Here's a screen recording so you can see when the errors are happening:
https://latenitefilms.digitalpigeon.com/msg/7T0pIHmjEeqrzAb4tvG4vQ/0Ha4_vj2gLJ5raW-NFCBng
latenitefilms
on 8 Apr 2020
Some more errors:
2020-04-09 00:53:54: ********
2020-04-09 00:53:54: 00:53:54 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-09 00:53:54: ********
2020-04-09 00:53:54: ********
2020-04-09 00:53:54: 00:53:54 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-09 00:53:54: ********
2020-04-09 00:54:41: ********
2020-04-09 00:54:41: 00:54:41 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-09 00:54:41: ********
2020-04-09 00:54:41: ********
2020-04-09 00:54:41: 00:54:41 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-09 00:54:41: ********
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: 00:55:06 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: 40463
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: 00:55:06 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: 00:55:06 ERROR: LuaSkin: hs.application.watcher callback: ...ing/Github/CommandPost/src/extensions/cp/ui/notifier.lua:64: attempt to index a nil value (local 'app')
stack traceback:
...ing/Github/CommandPost/src/extensions/cp/ui/notifier.lua:64: in function <...ing/Github/CommandPost/src/extensions/cp/ui/notifier.lua:63>
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: 00:55:06 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: 00:55:06 ERROR: LuaSkin: hs.application.watcher callback: ...hrishocking/Github/CommandPost/src/extensions/cp/app.lua:1241: attempt to index a nil value (local 'hsApp')
stack traceback:
...hrishocking/Github/CommandPost/src/extensions/cp/app.lua:1241: in function <...hrishocking/Github/CommandPost/src/extensions/cp/app.lua:1240>
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: 00:55:06 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: 00:55:06 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: 00:55:06 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: -1
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: 00:55:06 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: 40496
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: 00:55:06 ERROR: LuaSkin: hs.application.watcher callback: ...ing/Github/CommandPost/src/extensions/cp/ui/notifier.lua:64: attempt to index a nil value (local 'app')
stack traceback:
...ing/Github/CommandPost/src/extensions/cp/ui/notifier.lua:64: in function <...ing/Github/CommandPost/src/extensions/cp/ui/notifier.lua:63>
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: 00:55:06 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: 40496
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: 00:55:06 ERROR: LuaSkin: hs.application.watcher callback: ...hrishocking/Github/CommandPost/src/extensions/cp/app.lua:1241: attempt to index a nil value (local 'hsApp')
stack traceback:
...hrishocking/Github/CommandPost/src/extensions/cp/app.lua:1241: in function <...hrishocking/Github/CommandPost/src/extensions/cp/app.lua:1240>
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: 00:55:06 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: 40496
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: 00:55:06 ERROR: LuaSkin: hs.application.watcher callback: .../Github/CommandPost/src/plugins/core/menu/menuaction.lua:156: attempt to index a nil value (local 'app')
stack traceback:
.../Github/CommandPost/src/plugins/core/menu/menuaction.lua:156: in function <.../Github/CommandPost/src/plugins/core/menu/menuaction.lua:141>
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: 00:55:06 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: 40496
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: ********
2020-04-09 00:55:06: 00:55:06 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: 40496
2020-04-09 00:55:06: ********
> hs.application.applicationForPID(40496)
2020-04-09 00:55:53: ********
2020-04-09 00:55:53: 00:55:53 ERROR: LuaSkin: Unable to fetch NSRunningApplication for pid: 40496
2020-04-09 00:55:53: ********
nil
> hs.application.applicationForPID(40463)
hs.application: Final Cut Pro (0x6000005542c8)
For those playing at home, 32cd3cb seems to have solved the errors listed above. Woohoo!
Now the only thing left to do I think is to get hs._asm.axuielement playing nicely with everything again.
Thanks heaps @cmsj & @asmagill !!
latenitefilms
on 8 Apr 2020
To be clear, I didn't fix anything, I just switched some things from initialising HSapplication via PID when they had an NSRunningApplication. Something is clearly still wrong with the PID initialisation.
cmsj
on 8 Apr 2020
(but I can't trigger the PID failure unless I feed it an obviously incorrect PID, so it would be super helpful to have a small reproducer, if anyone is able to do that)
cmsj
on 8 Apr 2020
@latenitefilms try the latest at https://github.com/asmagill/hs._asm.axuielement
@cmsj I'm not quite ready to integrate axuielement, but with a little luck will be this weekend.
asmagill
on 9 Apr 2020
Thanks @asmagill - just merged in hs._asm.axuielement 0.7.4.
My existing code still doesn't seem to be working - so will have to do some more digging to work out why.
Also, just got this crash which I haven't seen before:
Process: CommandPost [48438]
Path: /Users/USER/*/CommandPost.app/Contents/MacOS/CommandPost
Identifier: org.latenitefilms.CommandPost
Version: 1.0.7-beta.6 (6015)
Code Type: X86-64 (Native)
Parent Process: ??? [1]
Responsible: CommandPost [48438]
User ID: 501
Date/Time: 2020-04-09 09:42:13.174 +1000
OS Version: Mac OS X 10.14.6 (18G3020)
Report Version: 12
Bridge OS Version: 3.0 (14Y908)
Anonymous UUID: 15304943-7B8E-F62C-F19F-B9B548D6D5BC
Sleep/Wake UUID: C959A4DF-DA7F-44B0-A6F4-7A02A127BDDF
Time Awake Since Boot: 720000 seconds
Time Since Wake: 2300 seconds
System Integrity Protection: enabled
Crashed Thread: 0 Dispatch queue: com.apple.main-thread
Exception Type: EXC_CRASH (SIGABRT)
Exception Codes: 0x0000000000000000, 0x0000000000000000
Exception Note: EXC_CORPSE_NOTIFY
Application Specific Information:
Assertion failed: (((n) < (L->top - L->ci->func)) && "not enough elements in the stack"), function luaD_precall, file /Users/chrishocking/Github/CommandPost-App/LuaSkin/lua-5.3.5/src/ldo.c, line 453.
Thread 0 Crashed:: Dispatch queue: com.apple.main-thread
0 libsystem_kernel.dylib 0x00007fff5c51e2c2 __pthread_kill + 10
1 libsystem_pthread.dylib 0x00007fff5c5d9bf1 pthread_kill + 284
2 libsystem_c.dylib 0x00007fff5c488745 __abort + 144
3 libsystem_c.dylib 0x00007fff5c4886b5 abort + 142
4 libsystem_c.dylib 0x00007fff5c45120d __assert_rtn + 324
5 org.hammerspoon.LuaSkin 0x0000000105d1de8b luaD_precall.cold.1 + 35
6 org.hammerspoon.LuaSkin 0x0000000105d129ba luaD_precall + 983 (ldo.c:512)
7 org.hammerspoon.LuaSkin 0x0000000105d0ea3c luaV_execute + 708 (lvm.c:1134)
8 org.hammerspoon.LuaSkin 0x0000000105d12e08 unroll + 72 (ldo.c:568)
9 org.hammerspoon.LuaSkin 0x0000000105d14221 luai_objcttry + 28 (lobjectivec_exceptions.m:173)
10 org.hammerspoon.LuaSkin 0x0000000105d12b31 lua_resume + 232 (ldo.c:681)
11 org.hammerspoon.LuaSkin 0x0000000105d0a9ca auxresume + 85 (lcorolib.c:40)
12 org.hammerspoon.LuaSkin 0x0000000105d0a79d luaB_coresume + 43 (lcorolib.c:61)
13 org.hammerspoon.LuaSkin 0x0000000105d1288c luaD_precall + 681 (ldo.c:453)
14 org.hammerspoon.LuaSkin 0x0000000105d0ea3c luaV_execute + 708 (lvm.c:1134)
15 org.hammerspoon.LuaSkin 0x0000000105d129fa luaD_call + 64 (ldo.c:517)
16 org.hammerspoon.LuaSkin 0x0000000105d12a3b luaD_callnoyield + 21 (ldo.c:527)
17 org.hammerspoon.LuaSkin 0x0000000105d14221 luai_objcttry + 28 (lobjectivec_exceptions.m:173)
18 org.hammerspoon.LuaSkin 0x0000000105d12f38 luaD_pcall + 110 (ldo.c:746)
19 org.hammerspoon.LuaSkin 0x0000000105d0c688 lua_pcallk + 236
20 org.hammerspoon.LuaSkin 0x0000000105cf54cb -[LuaSkin protectedCallAndTraceback:nresults:] + 218 (Skin.m:332)
21 internal.so 0x000000010aefb458 -[HSTimer callback:] + 244 (internal.m:51)
22 com.apple.Foundation 0x00007fff327067cb __NSFireTimer + 80
23 com.apple.CoreFoundation 0x00007fff30477810 __CFRUNLOOP_IS_CALLING_OUT_TO_A_TIMER_CALLBACK_FUNCTION__ + 20
24 com.apple.CoreFoundation 0x00007fff304773bc __CFRunLoopDoTimer + 851
25 com.apple.CoreFoundation 0x00007fff30476f02 __CFRunLoopDoTimers + 330
26 com.apple.CoreFoundation 0x00007fff30458112 __CFRunLoopRun + 2130
27 com.apple.CoreFoundation 0x00007fff3045766e CFRunLoopRunSpecific + 455
28 com.apple.HIToolbox 0x00007fff2f6b61ab RunCurrentEventLoopInMode + 292
29 com.apple.HIToolbox 0x00007fff2f6b5ee5 ReceiveNextEventCommon + 603
30 com.apple.HIToolbox 0x00007fff2f6b5c76 _BlockUntilNextEventMatchingListInModeWithFilter + 64
31 com.apple.AppKit 0x00007fff2da4d77d _DPSNextEvent + 1135
32 com.apple.AppKit 0x00007fff2da4c46b -[NSApplication(NSEvent) _nextEventMatchingEventMask:untilDate:inMode:dequeue:] + 1361
33 com.apple.AppKit 0x00007fff2da46588 -[NSApplication run] + 699
34 com.apple.AppKit 0x00007fff2da35ac8 NSApplicationMain + 777
35 libdyld.dylib 0x00007fff5c3e33d5 start + 1
@asmagill - Unfortunately, with hs._asm.axuielement 0.7.4 I just get a crash pretty much straight away when I load CommandPost now, so hard to test.
Crash report is on CommandPost's Fabric too if helpful.
latenitefilms
on 9 Apr 2020
Yeah, just got one myself, but haven't had a chance to review the crash log yet... Ok, well, it's getting there, just in fits and spurts...
asmagill
on 9 Apr 2020
Ok, @latenitefilms check out my latest pull request to Hammerspoon and the most recent push to axuielement and let me know how it goes.
Basically, HSuicore creates three classes for hs.application, hs.window, and hs.uielement but then allows some of the methods to be invoked from lua functions/methods rather than just through callbacks. It's those which can be invoked by lua functions/methods that require state to be passed, so I've added that to the necessary methods. It should be reviewed more to see if others need to be adjusted as well, but it passes my current usage of the three modules as well as my usage of hs._asm.axuielement.
asmagill
on 9 Apr 2020
Awesome, thank you! Building now.
latenitefilms
on 9 Apr 2020
Woohoo! That fixed it!
However, now I'm seeing this crash:
Process: CommandPost [53520]
Path: /Users/USER/*/CommandPost.app/Contents/MacOS/CommandPost
Identifier: org.latenitefilms.CommandPost
Version: 1.0.7-beta.6 (6017)
Code Type: X86-64 (Native)
Parent Process: ??? [1]
Responsible: CommandPost [53520]
User ID: 501
Date/Time: 2020-04-09 12:33:52.054 +1000
OS Version: Mac OS X 10.14.6 (18G3020)
Report Version: 12
Bridge OS Version: 3.0 (14Y908)
Anonymous UUID: 15304943-7B8E-F62C-F19F-B9B548D6D5BC
Sleep/Wake UUID: C959A4DF-DA7F-44B0-A6F4-7A02A127BDDF
Time Awake Since Boot: 730000 seconds
Time Since Wake: 12000 seconds
System Integrity Protection: enabled
Crashed Thread: 0 Dispatch queue: com.apple.main-thread
Exception Type: EXC_BAD_ACCESS (SIGSEGV)
Exception Codes: KERN_INVALID_ADDRESS at 0x00002b0308f50b30
Exception Note: EXC_CORPSE_NOTIFY
Termination Signal: Segmentation fault: 11
Termination Reason: Namespace SIGNAL, Code 0xb
Terminating Process: exc handler [53520]
VM Regions Near 0x2b0308f50b30:
CoreAnimation 000000011f054000-0000000120896000 [ 24.3M] rw-/rwx SM=PRV
-->
MALLOC_NANO 0000600000000000-0000600018000000 [384.0M] rw-/rwx SM=PRV
Thread 0 Crashed:: Dispatch queue: com.apple.main-thread
0 libobjc.A.dylib 0x00007fff5ac0864c objc_release + 28
1 org.latenitefilms.CommandPost 0x0000000109575e21 -[HSwindow dealloc] + 33
2 internal.so 0x000000010ecad86b userdata_gc + 169
3 org.hammerspoon.LuaSkin 0x00000001098d788c luaD_precall + 681 (ldo.c:453)
4 org.hammerspoon.LuaSkin 0x00000001098d79ee luaD_call + 52 (ldo.c:515)
5 org.hammerspoon.LuaSkin 0x00000001098d7a3b luaD_callnoyield + 21 (ldo.c:527)
6 org.hammerspoon.LuaSkin 0x00000001098d9221 luai_objcttry + 28 (lobjectivec_exceptions.m:173)
7 org.hammerspoon.LuaSkin 0x00000001098d7f38 luaD_pcall + 110 (ldo.c:746)
8 org.hammerspoon.LuaSkin 0x00000001098c926c GCTM + 252 (lgc.c:824)
9 org.hammerspoon.LuaSkin 0x00000001098c9078 runafewfinalizers + 53 (lgc.c:848)
10 org.hammerspoon.LuaSkin 0x00000001098d071e lua_pushlstring + 102 (lapi.c:487)
11 org.hammerspoon.LuaSkin 0x00000001098bea1a -[LuaSkin pushNSObject:withOptions:alreadySeenObjects:] + 1297
12 org.hammerspoon.LuaSkin 0x00000001098c0028 -[LuaSkin pushNSDictionary:withOptions:alreadySeenObjects:] + 534
13 org.hammerspoon.LuaSkin 0x00000001098beab1 -[LuaSkin pushNSObject:withOptions:alreadySeenObjects:] + 1448
14 org.hammerspoon.LuaSkin 0x00000001098c0028 -[LuaSkin pushNSDictionary:withOptions:alreadySeenObjects:] + 534
15 org.hammerspoon.LuaSkin 0x00000001098beab1 -[LuaSkin pushNSObject:withOptions:alreadySeenObjects:] + 1448
16 org.hammerspoon.LuaSkin 0x00000001098c0028 -[LuaSkin pushNSDictionary:withOptions:alreadySeenObjects:] + 534
17 org.hammerspoon.LuaSkin 0x00000001098beab1 -[LuaSkin pushNSObject:withOptions:alreadySeenObjects:] + 1448
18 org.hammerspoon.LuaSkin 0x00000001098c0028 -[LuaSkin pushNSDictionary:withOptions:alreadySeenObjects:] + 534
19 org.hammerspoon.LuaSkin 0x00000001098beab1 -[LuaSkin pushNSObject:withOptions:alreadySeenObjects:] + 1448
20 org.hammerspoon.LuaSkin 0x00000001098c0028 -[LuaSkin pushNSDictionary:withOptions:alreadySeenObjects:] + 534
21 org.hammerspoon.LuaSkin 0x00000001098beab1 -[LuaSkin pushNSObject:withOptions:alreadySeenObjects:] + 1448
22 org.hammerspoon.LuaSkin 0x00000001098c0028 -[LuaSkin pushNSDictionary:withOptions:alreadySeenObjects:] + 534
23 org.hammerspoon.LuaSkin 0x00000001098beab1 -[LuaSkin pushNSObject:withOptions:alreadySeenObjects:] + 1448
24 org.hammerspoon.LuaSkin 0x00000001098bc37c -[LuaSkin pushNSObject:withOptions:] + 119
25 internal.so 0x000000010ed3e754 json_read + 509
26 org.hammerspoon.LuaSkin 0x00000001098d788c luaD_precall + 681 (ldo.c:453)
27 org.hammerspoon.LuaSkin 0x00000001098d39bf luaV_execute + 583 (lvm.c:1149)
28 org.hammerspoon.LuaSkin 0x00000001098d79fa luaD_call + 64 (ldo.c:517)
29 org.hammerspoon.LuaSkin 0x00000001098d7a3b luaD_callnoyield + 21 (ldo.c:527)
30 org.hammerspoon.LuaSkin 0x00000001098d9221 luai_objcttry + 28 (lobjectivec_exceptions.m:173)
31 org.hammerspoon.LuaSkin 0x00000001098d7f38 luaD_pcall + 110 (ldo.c:746)
32 org.hammerspoon.LuaSkin 0x00000001098d1688 lua_pcallk + 236
33 org.hammerspoon.LuaSkin 0x00000001098ba4cb -[LuaSkin protectedCallAndTraceback:nresults:] + 218 (Skin.m:332)
34 org.hammerspoon.LuaSkin 0x00000001098ba592 -[LuaSkin protectedCallAndError:nargs:nresults:] + 55 (Skin.m:346)
35 watcher.so 0x000000010ec0d06f -[AppWatcher callback:withEvent:] + 452 (watcher.m:110)
36 watcher.so 0x000000010ec0d2a4 -[AppWatcher applicationActivated:] + 65 (watcher.m:134)
37 com.apple.CoreFoundation 0x00007fff304bb346 __CFNOTIFICATIONCENTER_IS_CALLING_OUT_TO_AN_OBSERVER__ + 12
38 com.apple.CoreFoundation 0x00007fff304bb2c0 ___CFXRegistrationPost_block_invoke + 63
39 com.apple.CoreFoundation 0x00007fff304bb22a _CFXRegistrationPost + 404
40 com.apple.CoreFoundation 0x00007fff304c36d8 ___CFXNotificationPost_block_invoke + 87
41 com.apple.CoreFoundation 0x00007fff3042c064 -[_CFXNotificationRegistrar find:object:observer:enumerator:] + 1642
42 com.apple.CoreFoundation 0x00007fff3042b417 _CFXNotificationPost + 732
43 com.apple.Foundation 0x00007fff326b2a7b -[NSNotificationCenter postNotificationName:object:userInfo:] + 66
44 com.apple.AppKit 0x00007fff2dc9353e applicationStatusSubsystemCallback + 758
45 com.apple.LaunchServices 0x00007fff31a5618f ___LSScheduleNotificationFunction_block_invoke_2 + 47
46 com.apple.CoreFoundation 0x00007fff30494f14 __CFRUNLOOP_IS_CALLING_OUT_TO_A_BLOCK__ + 12
47 com.apple.CoreFoundation 0x00007fff30458637 __CFRunLoopDoBlocks + 394
48 com.apple.CoreFoundation 0x00007fff30458394 __CFRunLoopRun + 2772
49 com.apple.CoreFoundation 0x00007fff3045766e CFRunLoopRunSpecific + 455
50 com.apple.HIToolbox 0x00007fff2f6b61ab RunCurrentEventLoopInMode + 292
51 com.apple.HIToolbox 0x00007fff2f6b5ee5 ReceiveNextEventCommon + 603
52 com.apple.HIToolbox 0x00007fff2f6b5c76 _BlockUntilNextEventMatchingListInModeWithFilter + 64
53 com.apple.AppKit 0x00007fff2da4d77d _DPSNextEvent + 1135
54 com.apple.AppKit 0x00007fff2da4c46b -[NSApplication(NSEvent) _nextEventMatchingEventMask:untilDate:inMode:dequeue:] + 1361
55 com.apple.AppKit 0x00007fff2da46588 -[NSApplication run] + 699
56 com.apple.AppKit 0x00007fff2da35ac8 NSApplicationMain + 777
57 libdyld.dylib 0x00007fff5c3e33d5 start + 1
At the end of the crashlog on your computer, it should list the loaded libraries and their address ranges... which internal.so contains 0x000000010ecad86b? I'm going to guess hs.window, but could you confirm?
(I've occasionally wondered if it would be worth the hassle of going through and uniquely re-naming every internal.m and watcher.m for this very reason, but it would mean changing every modules init.lua, every luaopen_* function, the entire build setup in XCode to match the new names, and probably some tweaks to package.cpath as well...)
asmagill
on 9 Apr 2020
Full crash log here: https://gist.github.com/latenitefilms/eabb739a2b5dc5f1ba347753f419f7eb
(I've occasionally wondered if it would be worth the hassle of going through and uniquely re-naming every internal.m and watcher.m for this very reason, but it would mean changing every modules init.lua, every luaopen_* function, the entire build setup in XCode to match the new names, and probably some tweaks to package.cpath as well...)
That is actually a really good idea. Obviously a lot of changes, but I think could definitely be worth thinking about - @cmsj ?
latenitefilms
on 9 Apr 2020
See if 0.7.5.1 for axuielement fixes it... it looks like the axuielement created for HSapplication objects is retained in it's initializers; the initializer for HSwindow assumes it's retained outside (because it usually is with a create or copy method prior to invoking the initializer, rather than hs.axuielement's way of slipping one in it already has the CFobject).
asmagill
on 9 Apr 2020
Woohoo! Yep, that seems to have fixed the crashes! You're an absolute genius and a legend. Thank you!
latenitefilms
on 9 Apr 2020
@asmagill - Are you using hs.application.watcher at all? I'm currently using it to detect when I'm changing apps - and it seems a bit slower than previously. Is that expected?
latenitefilms
on 9 Apr 2020
That I couldn't say... I'll take a closer look at it tomorrow and see what examples or tests I can run.
asmagill
on 9 Apr 2020
Actually, I have no idea if what I said is accurate or not - it could have always been this speed, and I've never noticed before because I wasn't looking for it - or my system could just be behaving differently. It could also be due to the fact that I have several hs.application.watcher objects in the code, so something in my own code could be slowing things down.
I'm going to close this issue, as I believe, for now, we've solved all the problems that I've run into.
Thanks for all your amazing work @cmsj & @asmagill !
latenitefilms
on 9 Apr 2020
@asmagill yeah I've considered doing The Great Renaming several times in the past couple of years. I think we should, but I think we should be very deliberate about it. We've had more pain than we might have done by landing the coroutines and application/window/uielement changes closely to each other, because they're kinda intertwined a bit.
If we do The Great Renaming, I think we should decide on what the outcome looks like, clear out anything else first that will interfere, and then get it all done.
cmsj
on 9 Apr 2020
@cmsj I agree that we should wait a bit before embarking on the great renaming... if you're still serious about locking down v1.0 in the near(ish) future (on even days I think its a good idea, on odd ones I'm not as sure since we're all just contributing "as and when we can"), then maybe make it part of v2?
And @latenitefilms you'll probably also be interested in this: I've added some documentation to the Hammerspoon wiki; first I moved my hs.canvas examples to the main wiki (and added a section in the menu for modules because I think there are some others than deserve examples as well) because I recently discovered that the hs.canvas doc strings still reference the one hosted on my github site... I'll correct this in a future pull, but it's a minor detail so not worthy of its own pull request.
I also added a section in the side menu for Developers and added a main page for my discussions about coroutines. While the old info is still relevant, and probably even important to module developers going forward, the more interesting document is the one about Leveraging Coroutines in Hammerspoon that you can get to from the main Coroutine page. It takes a barebones generic lua function and shows how to rewrite it to leverage coroutine.applicationYield as well as discuss some variants and limitations. I consider it very draftish at the moment, as I haven't tested all of the examples (especially the ones using hs.watchable to pass data in and out) fully yet, so let me know if you find any issues before I do... I'll be thoroughly testing it myself over the next couple of days, but figured even in its early state it would be of interest to the two of you.
asmagill
on 11 Apr 2020
Ok, I think I've worked out most of the bugs in my examples... let me know if I've missed anything!
asmagill
on 11 Apr 2020
Ok, I'm done tweaking the Leveraging documentation... for some reason it completely slipped my mind when considering callback functions -- they can allow passing new data into the coroutine as well, so I rewrote the section about passing data in and out. I'm done for tonight (on this at least), I promise!
asmagill
on 11 Apr 2020
Amazing - thank you! I look forward to reading this all in more detail over the coming days!
latenitefilms
on 11 Apr 2020
Related issues
agzam
路
3Comments
jasonrudolph
路
4Comments
reestr
路
3Comments
Pancia
路
4Comments
latenitefilms
路
4Comments