Re: [PATCH 2/2] gdb: MI stopped events when unwindonsignal is on

[Andrew Burgess <aburgess@redhat.com>]

Thiago Jung Bauermann <thiago.bauermann@linaro.org> writes:

> Hello Andrew,
>
> Andrew Burgess <aburgess@redhat.com> writes:
>
>> Thiago Jung Bauermann <thiago.bauermann@linaro.org> writes:
>>
>>> Hello Andrew,
>>>
>>> Andrew Burgess <aburgess@redhat.com> writes:
>>>
>>>> Thiago Jung Bauermann <thiago.bauermann@linaro.org> writes:
>>>>
>>>>> This test fails:
>>>>>
>>>>> Expecting: ^([
>>>>> ]+)?(terminate called after throwing an instance of 'int'
>>>>> ~"\\nProgram"
>>>>> ~" received signal SIGABRT, Aborted\.\\n"
>>>>> ~"0x[0-9A-Fa-f]+ in [^
>>>>> ]+"
>>>>> \*stopped,reason="signal-received",signal-name="SIGABRT"[^
>>>>> ]+frame=\{addr="0x[0-9A-Fa-f]+",[^
>>>>> ]+\}[^
>>>>> ]+
>>>>> &"Error in testing condition for breakpoint 2:\\n"
>>>>> &"The program being debugged was signaled while in a function called from GDB\.\\n"
>>>>> &"GDB remains in the frame where the signal was received\.\\n"
>>>>> &"To change this behavior use \\"set unwindonsignal on\\"\.\\n"
>>>>> &"Evaluation of the expression containing the function\\n"
>>>>> &"\(cond_throw\(\)\) will be abandoned\.\\n"
>>>>> &"When the function is done executing, GDB will silently stop\.\\n"
>>>>> =breakpoint-modified,bkpt={number="2",type="breakpoint",[^
>>>>> ]+times="1",[^
>>>>> ]+}[
>>>>> ]+[(]gdb[)] 
>>>>> [ ]*)
>>>>> terminate called after throwing an instance of 'int'
>>>>> ~"\nProgram"
>>>>> ~" received signal SIGABRT, Aborted.\n"
>>>>> ~"__pthread_kill_implementation (threadid=281474842417696, signo=signo@entry=6, no_tid=no_tid@entry=0) at ./nptl/pthread_kill.c:44\n"
>>>>> &"44\t./nptl/pthread_kill.c: No such file or directory.\n"
>>>>> *stopped,reason="signal-received",signal-name="SIGABRT",signal-meaning="Aborted",frame={addr="0x0000fffff7c5f200",func="__pthread_kill_implementation",args=[{name="threadid",value="281474842417696"},{name="signo",value="6"},{name="signo@entry",value="6"},{name="no_tid",value="0"},{name="no_tid@entry",value="0"}],file="./nptl/pthread_kill.c",fullname="./nptl/./nptl/pthread_kill.c",line="44",arch="aarch64"},thread-id="1",stopped-threads="all",core="78"
>>>>> &"Error in testing condition for breakpoint 2:\n"
>>>>> &"The program being debugged was signaled while in a function called from GDB.\n"
>>>>> &"GDB remains in the frame where the signal was received.\n"
>>>>> &"To change this behavior use \"set unwindonsignal on\".\n"
>>>>> &"Evaluation of the expression containing the function\n"
>>>>> &"(cond_throw()) will be abandoned.\n"
>>>>> &"When the function is done executing, GDB will silently stop.\n"
>>>>> =breakpoint-modified,bkpt={number="2",type="breakpoint",disp="keep",enabled="y",addr="0x0000aaaaaaaa087c",func="foo()",file="/home/tcwg-build/workspace/tcwg_gnu_5/abe/snapshots/gdb.git~master/gdb/testsuite/gdb.mi/mi-condbreak-throw.cc",fullname="/home/tcwg-build/workspace/tcwg_gnu_5/gdb/gdb/testsuite/gdb.mi/mi-condbreak-throw.cc",line="29",thread-groups=["i1"],cond="cond_throw ()",times="1",original-location="mi-condbreak-throw.cc:29"}
>>>>> (gdb) 
>>>>> FAIL: gdb.mi/mi-condbreak-throw.exp: unwind_on_exception=off: wait for stop (unexpected output)
>>>>>
>>>>> It looks like the testcase isn't expecting these lines:
>>>>>
>>>>> ~"__pthread_kill_implementation (threadid=281474842417696, signo=signo@entry=6, no_tid=no_tid@entry=0) at ./nptl/pthread_kill.c:44\n"
>>>>> &"44\t./nptl/pthread_kill.c: No such file or directory.\n"
>>>>
>>>> Thanks for testing this, and highlighting this failure.
>>>>
>>>> I've included an updated patch below.  I've made the test pattern a
>>>> little more general so it should (I hope) handle the lines you are
>>>> seeing.
>>>>
>>>> I'd be grateful if you could retest this new version and let me know if
>>>> you are still seeing the failure.
>>>
>>> Thank you for the new version. Unfortunately I still see the failure.
>>> I'm attaching the gdb.log file.
>>
>> Ok, I think I have it this time :) Maybe...
>>
>> ... would you be so kind as to give the version below a try please.
>
> Of course. Thank you for the new version. This one worked well.
>
> Running gdb.git~master/gdb/testsuite/gdb.mi/mi-condbreak-throw.exp ...
> PASS: gdb.mi/mi-condbreak-throw.exp: unwind_on_exception=off: breakpoint at main
> PASS: gdb.mi/mi-condbreak-throw.exp: unwind_on_exception=off: mi runto main
> PASS: gdb.mi/mi-condbreak-throw.exp: unwind_on_exception=off: set unwind-on-terminating-exception
> PASS: gdb.mi/mi-condbreak-throw.exp: unwind_on_exception=off: insert conditional breakpoint
> PASS: gdb.mi/mi-condbreak-throw.exp: unwind_on_exception=off: get number for breakpoint
> PASS: gdb.mi/mi-condbreak-throw.exp: unwind_on_exception=off: wait for stop
>
>                 === gdb Summary ===
>
> # of expected passes            6
>
> Sorry for the delay in testing this one.

Not a problem.  Thanks for all the testing.

I've gone ahead and pushed this series, as this was blocking me from
posting the latest version of my inferior function call series[1].

If there's any post-commit feedback, I am always happy to address it.

Thanks,
Andrew

[1] https://inbox.sourceware.org/gdb-patches/cover.1686131880.git.aburgess@redhat.com/