[SyncEvolution] Syncing to Memotoo w SyncEvolution post-Fedora 37 upgrade (fwd)

[Max Pyziur]

On Tue, 21 Feb 2023, Patrick Ohly wrote:

> Max Pyziur <pyz at brama.com> writes:
>
>> On Tue, 21 Feb 2023, Patrick Ohly wrote:
>>
>>> Max Pyziur <pyz at brama.com> writes:
>>>> On Sun, 19 Feb 2023, Patrick Ohly wrote:
>>>> Here is the command that I run in a separate shell:
>>>> $ syncevolution --run  --sync refresh-from-local memotoo
>>>>
>>>>> The relevant output is the one from valgrind.
>>>>
>>>> This is the output that I receive:
>>>> pyz at pegasus ~/projects/HomeComputers/Pegasus> valgrind
>>>> /usr/libexec/syncevo-dbus-server
>>>> ==6181== Memcheck, a memory error detector
>>>> ==6181== Copyright (C) 2002-2022, and GNU GPL'd, by Julian Seward et al.
>>>> ==6181== Using Valgrind-3.20.0 and LibVEX; rerun with -h for copyright
>>>> info
>>>> ==6181== Command: /usr/libexec/syncevo-dbus-server
>>>> ==6181==
>>>> --6181-- WARNING: unhandled amd64-linux syscall: 434
>>>> --6181-- You may be able to write your own handler.
>>>> --6181-- Read the file README_MISSING_SYSCALL_OR_IOCTL.
>>>> --6181-- Nevertheless we consider this a bug.  Please report
>>>> --6181-- it at http://valgrind.org/support/bug_reports.html.
>>>
>>> And that's all? It doesn't crash anymore?
>>
>> It crashes.
>>
>>> To be clear, you need to run `syncevolution --run --sync
>>> refresh-from-local memotoo` while valgrind is in control of
>>> syncevo-dbus-server.
>>
>> I did; it was in the reply that I sent, please see above.
>>
>> I'll repeat the steps as I know them; I am also expanding the 'valgrind'
>> command.
>>
>> #1
>> pyz at pegasus ~> ps waux | grep sync
>> pyz         5511  0.0  0.3 527508 28416 ?        Ssl  07:09   0:00
>> /usr/libexec/syncevo-dbus-server
>> pyz         5756  0.0  0.0 222156  2124 pts/2    R+   07:12   0:00 grep
>> --color=auto sync
>> pyz at pegasus ~> killall syncevo-dbus-server
>>
>> #2
>> Start valgrind thusly:
>> pyz at pegasus ~> valgrind --leak-check=full /usr/libexec/syncevo-dbus-server
>
>
> Something doesn't add up here. You should only need to do this once,
> i.e. this invocation here should keep running until you do step #3.
>
>> #3
>> Run in a separate window and receive the following output:
>> pyz at pegasus ~> syncevolution --run  --sync refresh-from-local memotoo
>> [INFO] memo: inactive
>> [INFO] todo: inactive
>> [ERROR syncevo-dbus-server] child process quit because of signal 11
>> [ERROR] The connection is closed
>> pyz at pegasus ~>
>
> Now the valgrind invocation above should report the crash and exit.
>
>> #4
>> I receive the following output from the valgrind command:
>> pyz at pegasus ~> valgrind --leak-check=full /usr/libexec/syncevo-dbus-server
>> ==5779== Memcheck, a memory error detector
>> ==5779== Copyright (C) 2002-2022, and GNU GPL'd, by Julian Seward et al.
>> ==5779== Using Valgrind-3.20.0 and LibVEX; rerun with -h for copyright
>> info
>> ==5779== Command: /usr/libexec/syncevo-dbus-server
>> ==5779==
>> [ERROR syncevo-dbus-server] could not obtain D-Bus name - already
>> running?
>
> That's not what should happen. When you run valgrind directly after
> killall, this instance should keep running.

I'm following the steps that you've given me, as best I understand them. 
Here, I've enumerated them (1 to 4).

What else do I have to do to detect the correct error for you?

syncevolution + memotoo are integral to my PIM setup.

I'm using this on a Dell XPS L322X (I think that it's a 2012 vintage 
laptop).

I've installed and upgraded Fedora on this machine since Fedora 25 or so 
until the recent version 37.

W/o clear verification, it is only post F37 that syncevolution crashs when 
I try to sync with Memotoo.

fyi,

M