bug#58226: 29.0.50; multisession-tests.el fails to finish on macOS (10.13.6)

[Gerd Möllmann]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Stefan Kangas <stefankangas@gmail.com> writes:
>
>> If I try to get the ELisp backtrace from ert with "b", Emacs crashes
>> with the below rather unhelpful backtrace. Any ideas for how to debug
>> further from here are welcome.
>>
>> Thread 2 received signal SIGABRT, Aborted.
>> 0x00007fff6d8a3b66 in ?? () from /usr/lib/system/libsystem_kernel.dylib
>> (gdb) bt
>> #0  0x00007fff6d8a3b66 in ?? () from /usr/lib/system/libsystem_kernel.dylib
>> #1  0x00007fff6da6e080 in pthread_kill () from
>> /usr/lib/system/libsystem_pthread.dylib
>> #2  0x00007fff6d7ff24d in __abort () from /usr/lib/system/libsystem_c.dylib
>> #3  0x0032000000000015 in ?? ()
>> #4  0x0000000000000000 in ?? ()
>
> I'm almost completely blank on running the debugger on Macos, but Gerd
> is an expert here

That's how quickly you get a bad reputation :-).

> -- Gerd, any advice?

It's to a degree similar to GDB.  I'm using the cheat sheet from
https://lldb.llvm.org/use/map.html if in doubt.  That's how experts do.

cd src
lldb --local-lldbinit emacs

You should see "Emacs support has been loaded".  If not, you can load it
later with

command script import ../etc/emacs_lldb.py

To start Emacs

run <args...>

where <args...> are command-line options for Emacs.  When back in lldb
because of the signal, use bt, up, down, and p as you would in GDB.
With emacs_lldb.py loaded, you can also use xbacktrace and xdebug_print.