[OpenIndiana-discuss] mate system monitor does not work any more on recent versions of openindiana

[Udo Grabowski (IMK)]

On 19/04/2023 15:01, Marc Lobelle wrote:
> Hello Udo,
>
> I do not really understand what to do: uninstalling the old gcc 7 ? But I do not
> see the relation between fam and the system monitor program.
>
> What exactly should I do ?
>

You can't do anything. mate-system-monitor has to be recompiled
so that it don't load both gcc7 and gcc10 libraries. I emulated
that by changing the binary with elfedit, linking libraries
locally and preloading. But then you trap on the root cause,
which is the missing fam support in glib that causes most of
the gio stuff to fail. This is also seen in a report on the
GNU Hurd list: <https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=875915>

>
> On 4/19/23 14:46, Udo Grabowski (IMK) wrote:
>> On 19/04/2023 13:00, Stephan Althaus wrote:
>>> On 4/19/23 12:25, Marc Lobelle wrote:
>>>> Hello,
>>>>
>>>> I updated openindiana from 2021-10 to 2022-10 on my notebook and the mate
>>>> system monitor does not start if I click on the widget: the widget is still
>>>> there and displays cpu usage, but if I click on it to get the window with full
>>>> information, I see a hourglass for a few seconds then nothing: it seems to
>>>> crash before opening the window.
>>>>
>>>> Is there a workaround ?
>>>>
>>>> Thanks
>>>>
>>>> Marc
>>>
>>> Hello!
>>>
>>> This issue persists for several weeks now and is reproducible.
>>>
>>> Alas, i can't see the root caause from the 'pstack' of the core dump, maybe
>>> someone else is able to dig this out.
>>>
>>> Regards,
>>>
>>> Stephan
>>>
>>> $ pstack core
>>> core 'core' of 7849:    mate-system-monitor
>>> --------------------- thread# 1 / lwp# 1 ---------------------
>>>  00007fffaf40729a _lwp_kill () + a
>>>  00007fffaf39b1ee raise (6) + 1e
>>>  00007fffaf374c58 abort () + 58
>>>  00007fffab6e2b0e ???????? ()
>>>  00007fffab8708c1 __gxx_personality_v0 () + d1
>>>  00007fffaccd5756 _Unwind_RaiseException_Phase2 () + a6
>>>  00007fffaccd6126 _Unwind_Resume () + 126
>>>  00007fff7712b9d5 ???????? ()
>>>  000000000042f84b _ZN11PrettyTableC1Ev () + 1db
>>>  00000000004223e1 _ZN8ProcData12get_instanceEv () + 81
>>>  000000000044832c _ZL16procman_data_newP10_GSettings () + 1c
>>>  000000000044904d _ZN10ProcmanApp10on_startupEv () + 4d
>>>  00007fff770f9615 _ZN3Gio17Application_Class16startup_callbackEP13_GApplication
>>> ...
>>
>> This boils down essentially to two problems at once: Both gcc7 and gcc10
>> libstdc++.so and libgcc_s.so are loaded, which breaks exception handling.
>> Massaging dyn:runpath and preloading then uncovers the second problem:
>>
>> ro sunts /tmp # ./mate-system-monitor
>>
>> (<unknown>:11878): glibmm-CRITICAL **: 14:08:27.008:
>> unhandled exception (type Glib::Error) in signal handler:
>> domain: g-io-error-quark
>> code  : 0
>> what  : Unable to find default local file monitor type
>>
>> Segmentation fault
>>
>>
>> That's what I already wrote in the thread
>> <https://www.mail-archive.com/openindiana-discuss@openindiana.org/msg26671.html>
>>
>> The /usr/lib/gio/modules/ directory has the giomodule.cache file.
>> When you try to recreate it with gio-querymodules /usr/lib/gio/modules/,
>> it's gone!
>>
>> So probably the right fixes should be: Sort out the gcc7/10 confusion,
>> and compile glib with inotify/fam support.
>>
>> _______________________________________________
>> openindiana-discuss mailing list
>> openindiana-discuss at openindiana.org
>> https://openindiana.org/mailman/listinfo/openindiana-discuss


-- 
Dr.Udo Grabowski   Inst.f.Meteorology & Climate Research IMK-ASF-SAT
http://www.imk-asf.kit.edu/english/sat.php
KIT - Karlsruhe Institute of Technology           http://www.kit.edu
Postfach 3640,76021 Karlsruhe,Germany T:(+49)721 608-26026 F:-926026