Pressing 'e' in eeschema causes crash frequently [OSX-specific]

[ocb-hd-ccb]

I though this was originally related to https://bugs.launchpad.net/kicad/+bug/1836673 but since learning how to check the crash logs it appears to be different.

I am pressing 'e' over field references in eeschema and sometimes it works and sometimes it crashes. But, once it crashes it crashes for a while. I have to do an unknown procedure to get it to work again.

#Kicad Info
```
Application: KiCad
Version: (5.1.0-1318-gb5cc49614), debug build
Libraries:
    wxWidgets 3.0.4
    libcurl/7.54.0 LibreSSL/2.6.5 zlib/1.2.11 nghttp2/1.24.1
Platform: Mac OS X (Darwin 18.6.0 x86_64), 64 bit, Little endian, wxMac
Build Info:
    wxWidgets: 3.0.4 (wchar_t,STL containers,compatible with 2.8)
    Boost: 1.69.0
    OpenCASCADE Community Edition: 6.9.1
    Curl: 7.54.0
    Compiler: Clang 9.0.0 with C++ ABI 1002

Build settings:
    KICAD_SCRIPTING=ON
    KICAD_SCRIPTING_MODULES=ON
    KICAD_SCRIPTING_PYTHON3=OFF
    KICAD_SCRIPTING_WXPYTHON=ON
    KICAD_SCRIPTING_WXPYTHON_PHOENIX=OFF
    KICAD_SCRIPTING_ACTION_MENU=ON
    BUILD_GITHUB_PLUGIN=ON
    KICAD_USE_OCE=ON
    KICAD_USE_OCC=OFF
    KICAD_SPICE=ON
```

#Crash Report
```
Crashed Thread: 0 Dispatch queue: com.apple.main-thread

Exception Type: EXC_BAD_INSTRUCTION (SIGILL)
Exception Codes: 0x0000000000000001, 0x0000000000000000
Exception Note: EXC_CORPSE_NOTIFY

Termination Signal: Illegal instruction: 4
Termination Reason: Namespace SIGNAL, Code 0x4
Terminating Process: exc handler [16850]

Application Specific Information:
Crashing on exception: -[NSAlert runModal] may not be invoked inside of transaction begin/commit pair, or inside of transaction commit (usually this means it was invoked inside of a view's -drawRect: method.)

Thread 0 Crashed:: Dispatch queue: com.apple.main-thread
0 com.apple.AppKit 0x00007fff47071a67 -[NSApplication _crashOnException:] + 109
1 com.apple.AppKit 0x00007fff46d6ec84 __65+[CATransaction(NSCATransaction) NS_setFlushesWithDisplayRefresh]_block_invoke + 841
2 com.apple.CoreFoundation 0x00007fff49645928 __CFRUNLOOP_IS_CALLING_OUT_TO_AN_OBSERVER_CALLBACK_FUNCTION__ + 23
3 com.apple.CoreFoundation 0x00007fff4964585d __CFRunLoopDoObservers + 451
4 com.apple.CoreFoundation 0x00007fff495e7f80 __CFRunLoopRun + 1136
5 com.apple.CoreFoundation 0x00007fff495e78be CFRunLoopRunSpecific + 455
6 com.apple.HIToolbox 0x00007fff488d396b RunCurrentEventLoopInMode + 292
7 com.apple.HIToolbox 0x00007fff488d35ad ReceiveNextEventCommon + 355
8 com.apple.HIToolbox 0x00007fff488d3436 _BlockUntilNextEventMatchingListInModeWithFilter + 64
9 com.apple.AppKit 0x00007fff46c6d987 _DPSNextEvent + 965
10 com.apple.AppKit 0x00007fff46c6c71f -[NSApplication(NSEvent) _nextEventMatchingEventMask:untilDate:inMode:dequeue:] + 1361
11 libwx_osx_cocoau-3.0.0.4.0.dylib 0x000000010ac5167c wxGUIEventLoop::OSXDoRun() + 220
12 libwx_osx_cocoau-3.0.0.4.0.dylib 0x000000010ab4c0e1 wxCFEventLoop::DoRun() + 49
13 libwx_osx_cocoau-3.0.0.4.0.dylib 0x000000010aa99b42 wxEventLoopBase::Run() + 162
14 _eeschema.kiface 0x000000011432ce11 DIALOG_SHIM::ShowQuasiModal() + 673
15 _eeschema.kiface 0x00000001141467b4 SCH_EDIT_TOOL::editComponentFieldText(SCH_FIELD*) + 420
16 _eeschema.kiface 0x000000011414772f SCH_EDIT_TOOL::Properties(TOOL_EVENT const&) + 1263
17 _eeschema.kiface 0x0000000114156cd5 int std::__1::__invoke_void_return_wrapper<int>::__call<std::__1::__bind<int (SCH_EDIT_TOOL::*&)(TOOL_EVENT const&), SCH_EDIT_TOOL*, std::__1::placeholders::__ph<1> const&>&, TOOL_EVENT const&>(std::__1::__bind<int (SCH_EDIT_TOOL::*&)(TOOL_EVENT const&), SCH_EDIT_TOOL*, std::__1::placeholders::__ph<1> const&>&&&, TOOL_EVENT const&&&) + 645
18 _eeschema.kiface 0x0000000114156999 std::__1::__function::__func<std::__1::__bind<int (SCH_EDIT_TOOL::*&)(TOOL_EVENT const&), SCH_EDIT_TOOL*, std::__1::placeholders::__ph<1> const&>, std::__1::allocator<std::__1::__bind<int (SCH_EDIT_TOOL::*&)(TOOL_EVENT const&), SCH_EDIT_TOOL*, std::__1::placeholders::__ph<1> const&> >, int (TOOL_EVENT const&)>::operator()(TOOL_EVENT const&) + 57
19 _eeschema.kiface 0x00000001144c803e std::__1::function<int (TOOL_EVENT const&)>::operator()(TOOL_EVENT const&) const + 142
20 _eeschema.kiface 0x00000001144c7f8d COROUTINE<int, TOOL_EVENT const&>::callerStub(long) + 109
21 _eeschema.kiface 0x000000011440cd2a make_fcontext + 42
22 ??? 0x00007ffee56bdd90 0 + 140732747472272

Thread 1:
0 libsystem_pthread.dylib 0x00007fff757093f0 start_wqthread + 0

Thread 2:
0 libsystem_pthread.dylib 0x00007fff757093f0 start_wqthread + 0

Thread 3:
0 libsystem_pthread.dylib 0x00007fff757093f0 start_wqthread + 0

Thread 4:: com.apple.CFSocket.private
0 libsystem_kernel.dylib 0x00007fff7565261a __select + 10
1 com.apple.CoreFoundation 0x00007fff49616572 __CFSocketManager + 635
2 libsystem_pthread.dylib 0x00007fff7570a2eb _pthread_body + 126
3 libsystem_pthread.dylib 0x00007fff7570d249 _pthread_start + 66
4 libsystem_pthread.dylib 0x00007fff7570940d thread_start + 13

Thread 5:: com.apple.NSEventThread
0 libsystem_kernel.dylib 0x00007fff7564b22a mach_msg_trap + 10
1 libsystem_kernel.dylib 0x00007fff7564b76c mach_msg + 60
2 com.apple.CoreFoundation 0x00007fff495e8bee __CFRunLoopServiceMachPort + 328
3 com.apple.CoreFoundation 0x00007fff495e815c __CFRunLoopRun + 1612
4 com.apple.CoreFoundation 0x00007fff495e78be CFRunLoopRunSpecific + 455
5 com.apple.AppKit 0x00007fff46c756a6 _NSEventThread + 175
6 libsystem_pthread.dylib 0x00007fff7570a2eb _pthread_body + 126
7 libsystem_pthread.dylib 0x00007fff7570d249 _pthread_start + 66
8 libsystem_pthread.dylib 0x00007fff7570940d thread_start + 13

Thread 6:
0 libsystem_pthread.dylib 0x00007fff757093f0 start_wqthread + 0

Thread 0 crashed with X86 Thread State (64-bit):
  rax: 0x00006000028bc230 rbx: 0x000060000329a220 rcx: 0x00007fa6edc471f0 rdx: 0x00007fff73d5ef18
  rdi: 0x00007fa6edc47000 rsi: 0x00007fff477f0bed rbp: 0x0000000111a607b0 rsp: 0x0000000111a607a0
   r8: 0x0000000111a60340 r9: 0x0000000000000000 r10: 0x00007fffa4980390 r11: 0x00007fa6defb7190
  r12: 0x0000600000cfd860 r13: 0x00007fa6ed7612c0 r14: 0x00007fff73d3a680 r15: 0x00007fff49652f83
  rip: 0x00007fff47071a67 rfl: 0x0000000000010206 cr2: 0x000000010b31b3cc

Logical CPU: 3
Error Code: 0x00000000
Trap Number: 6
```

[ocb-hd-ccb (ocb-hd-ccb)]

I don't actually think this will be helpful but here is what happens. I had the MCU selected, I deselected it and hit 'e' between the 'field ref' and 'field val' and I picked 'field ref' then crash...

https://imgur.com/a/zF4wehd

[Jeff Young (jeyjey)]

I've made this happen a couple of times, but now can't make it happen again.

@Seth, the "[NSAlert runModal] may not be invoked inside of transaction begin/commit pair, or inside of transaction commit (usually this means it was invoked inside of a view's -drawRect: method.)" rings a bell. Didn't we have another one of those? Any memory of what it turned out to be?

[Jeff Young (jeyjey)]

I just got it to happen with only the Kicad Manager and Eeschema running (so it doesn't appear to be related to cross-probing), and without any Clarify Selection menus having appeared (so it doesn't appear to be related to them either).

Some comments online suggest that this is Apple's bug (possibly High Sierra-specific), but we'll still need to figure out a way to work around it.

[ocb-hd-ccb (ocb-hd-ccb)]

I guess it could be High Sierra and beyond but I am running Mojave...

[Jeff Young (jeyjey)]

I've put a hack in which *might* work around this. I say "might" because this bug seems to go in and out of reproducibility, so the fact that I can't make it happen with this hack doesn't actually mean it's fixed.

[KiCad Janitor (kicad-janitor)]

Fixed in revision 4d4adf07538f87536cf3196cd65b859efc05b079
https://git.launchpad.net/kicad/patch/?id=4d4adf07538f87536cf3196cd65b859efc05b079

[Jeff Young (jeyjey)]

@ocb-hd-ccb, can you try this out when you get a chance and see if it helps any?

[Jeff Young (jeyjey)]

(After the next nightly is built, that is....)

[ocb-hd-ccb (ocb-hd-ccb)]

@Jeff

Yea! I will give it a whirl on Monday. Hopefully the nightly is updated by then.

👍

[ocb-hd-ccb (ocb-hd-ccb)]

Issue still happens on 20190721 ... Don't know if this version would have had the fix or not...

```
Application: KiCad
Version: (5.1.0-1341-g0d8ab12ad), debug build
Libraries:
    wxWidgets 3.0.4
    libcurl/7.54.0 LibreSSL/2.6.5 zlib/1.2.11 nghttp2/1.24.1
Platform: Mac OS X (Darwin 18.6.0 x86_64), 64 bit, Little endian, wxMac
Build Info:
    wxWidgets: 3.0.4 (wchar_t,STL containers,compatible with 2.8)
    Boost: 1.69.0
    OpenCASCADE Community Edition: 6.9.1
    Curl: 7.54.0
    Compiler: Clang 9.0.0 with C++ ABI 1002

Build settings:
    KICAD_SCRIPTING=ON
    KICAD_SCRIPTING_MODULES=ON
    KICAD_SCRIPTING_PYTHON3=OFF
    KICAD_SCRIPTING_WXPYTHON=ON
    KICAD_SCRIPTING_WXPYTHON_PHOENIX=OFF
    KICAD_SCRIPTING_ACTION_MENU=ON
    BUILD_GITHUB_PLUGIN=ON
    KICAD_USE_OCE=ON
    KICAD_USE_OCC=OFF
    KICAD_SPICE=ON
```

[Jeff Young (jeyjey)]

@Adam, can you tell me whether or not the above build has 4d4adf07538f87536cf3196cd65b859efc05b079 in it or not? Thanks!

[Jeff Young (jeyjey)]

Other evidence suggests (https://bugs.launchpad.net/kicad/+bug/1836267) the builds are not up-to-date so I'm going to go ahead and mark this fixed for now.

[ocb-hd-ccb (ocb-hd-ccb)]

The 'e' issue is still present in todays nightly 20190722

```
Application: KiCad
Version: (5.1.0-1347-ga18fcdd6f), debug build
Libraries:
    wxWidgets 3.0.4
    libcurl/7.54.0 LibreSSL/2.6.5 zlib/1.2.11 nghttp2/1.24.1
Platform: Mac OS X (Darwin 18.6.0 x86_64), 64 bit, Little endian, wxMac
Build Info:
    wxWidgets: 3.0.4 (wchar_t,STL containers,compatible with 2.8)
    Boost: 1.69.0
    OpenCASCADE Community Edition: 6.9.1
    Curl: 7.54.0
    Compiler: Clang 9.0.0 with C++ ABI 1002

Build settings:
    KICAD_SCRIPTING=ON
    KICAD_SCRIPTING_MODULES=ON
    KICAD_SCRIPTING_PYTHON3=OFF
    KICAD_SCRIPTING_WXPYTHON=ON
    KICAD_SCRIPTING_WXPYTHON_PHOENIX=OFF
    KICAD_SCRIPTING_ACTION_MENU=ON
    BUILD_GITHUB_PLUGIN=ON
    KICAD_USE_OCE=ON
    KICAD_USE_OCC=OFF
    KICAD_SPICE=ON
```

[ocb-hd-ccb (ocb-hd-ccb)]

@Jeff

Bug is still in 20190725 10-14 nightly

```
Application: KiCad
Version: (5.1.0-1361-g807bd7889), debug build
Libraries:
    wxWidgets 3.0.4
    libcurl/7.54.0 LibreSSL/2.6.5 zlib/1.2.11 nghttp2/1.24.1
Platform: Mac OS X (Darwin 18.7.0 x86_64), 64 bit, Little endian, wxMac
Build Info:
    wxWidgets: 3.0.4 (wchar_t,STL containers,compatible with 2.8)
    Boost: 1.69.0
    OpenCASCADE Community Edition: 6.9.1
    Curl: 7.54.0
    Compiler: Clang 9.0.0 with C++ ABI 1002

Build settings:
    KICAD_SCRIPTING=ON
    KICAD_SCRIPTING_MODULES=ON
    KICAD_SCRIPTING_PYTHON3=OFF
    KICAD_SCRIPTING_WXPYTHON=ON
    KICAD_SCRIPTING_WXPYTHON_PHOENIX=OFF
    KICAD_SCRIPTING_ACTION_MENU=ON
    BUILD_GITHUB_PLUGIN=ON
    KICAD_USE_OCE=ON
    KICAD_USE_OCC=OFF
    KICAD_SPICE=ON
```

[Jeff Young (jeyjey)]

@Seth, any ideas on comment #2?

[Seth Hillbrand (sethh)]

@Jeff- That was the footprint wizard activation issue.

f214cfae12ea6745c92e647f572a43e89be92405 in the v5 branch.

[KiCad Janitor (kicad-janitor)]

Fixed in revision 21b90de829df1d205a760102746991bb227aac59
https://git.launchpad.net/kicad/patch/?id=21b90de829df1d205a760102746991bb227aac59

[Jeff Young (jeyjey)]

@ocb-hd-ccb, I've merged another attempt to fix it. Let me know how you get on with it.

[Jeff Young (jeyjey)]

@Seth, thanks. I've tried something similar only waiting for Idle events to be seen before continuing. But this one seems a little more resilient to fixes. (Perhaps we're breeding resistant strains of bugs....)

[ocb-hd-ccb (ocb-hd-ccb)]

@Jeff cool. I will let you know.

[ocb-hd-ccb (ocb-hd-ccb)]

@Jeff still an issue in 20190727

```
Application: KiCad
Version: (5.1.0-1376-geaeb96ea5), debug build
Libraries:
    wxWidgets 3.0.4
    libcurl/7.54.0 LibreSSL/2.6.5 zlib/1.2.11 nghttp2/1.24.1
Platform: Mac OS X (Darwin 18.7.0 x86_64), 64 bit, Little endian, wxMac
Build Info:
    wxWidgets: 3.0.4 (wchar_t,STL containers,compatible with 2.8)
    Boost: 1.69.0
    OpenCASCADE Community Edition: 6.9.1
    Curl: 7.54.0
    Compiler: Clang 9.0.0 with C++ ABI 1002

Build settings:
    KICAD_SCRIPTING=ON
    KICAD_SCRIPTING_MODULES=ON
    KICAD_SCRIPTING_PYTHON3=OFF
    KICAD_SCRIPTING_WXPYTHON=ON
    KICAD_SCRIPTING_WXPYTHON_PHOENIX=OFF
    KICAD_SCRIPTING_ACTION_MENU=ON
    BUILD_GITHUB_PLUGIN=ON
    KICAD_USE_OCE=ON
    KICAD_USE_OCC=OFF
    KICAD_SPICE=ON
```

[Jeff Young (jeyjey)]

I also tried moving editComponentFieldText() back to SCH_EDIT_FRAME (in case it was something about the coroutine stuff that was triggering it). That didn't work either.

And I tried firing a second ACTION from the first so that we go back around the event loop. Again, no joy.

At this point I'm out of ideas.

[Jeff Young (jeyjey)]

@Seth, if you want to give this a crack that would be great.

[Seth Hillbrand (sethh)]

Oh boy. :) Well, at least it's not a Windows bug.

@ocb-hd-ccb: Could you attach a few more of your crash logs? I'd like to verify we're always seeing the same behavior.

[ocb-hd-ccb (ocb-hd-ccb)]

@Seth I have been checking the crash logs each time and the reason I haven’t posted them was because it was always the NSAlert -drawRect thing. Same as the first log I posted.

When I get back to the office Monday night/ Tuesday I can upload another.

[ocb-hd-ccb (ocb-hd-ccb)]

I don't know how much of the crash log you need but it appears to all be the same

This is from the last crash
```
Application Specific Information:
Crashing on exception: -[NSAlert runModal] may not be invoked inside of transaction begin/commit pair, or inside of transaction commit (usually this means it was invoked inside of a view's -drawRect: method.)
```

[Jeff Young (jeyjey)]

Ditto for me. When I /can/ reproduce it, it's always the same crash.

[Jeff Young (jeyjey)]

@Seth, this was really bothering me so I took another look from the perspective of "what's different about this dialog"?

Anyway, I think I've fixed it.

[KiCad Janitor (kicad-janitor)]

Fixed in revision c03535343c67c17ddf0ac18621743d9cf762eb05
https://git.launchpad.net/kicad/patch/?id=c03535343c67c17ddf0ac18621743d9cf762eb05

[ocb-hd-ccb (ocb-hd-ccb)]

Yippee! I will pull and test the next nightly I see.

[ocb-hd-ccb (ocb-hd-ccb)]

It did not crash!

Thanks!

I need to find better screen recording software that shows what keys are being hit as well as records system sounds. There is some slightly odd behavior but I cannot easily describe it. When I find a better than macOS default screen recording I will open a new issue.

[Seth Hillbrand (sethh)]

Nicely done Jeff!