Bug#1091358: mumble: crashes on startup with pipewire backtrace

[Philipp Kern]

Control: tag -1 - moreinfo

Hi,

On 12/24/24 4:48 PM, Diederik de Haas wrote:
> Control: tag -1 moreinfo

This is not an appropriate use of moreinfo. "unreproducible" if you 
can't reproduce it, maybe. But there's not a question of what's broken - 
and retesting with a version that is not in the archive is not "This bug 
can't be addressed until more information is provided by the submitter. 
The bug will be closed if the submitter doesn't provide more information 
in a reasonable (few months) timeframe. This is for bugs like "It 
doesn't work". What doesn't work?"[1].

> On Tue Dec 24, 2024 at 4:10 PM CET, Philipp Kern wrote:
>> Package: mumble
>> Version: 1.5.517-2.1
>> Severity: important
>>
>> Installed mumble without any prior config and it crashes on startup:
>>
>> [...]
>> <W>2024-12-24 16:07:13.611 Wayland does not support QWindow::requestActivate()
>> <W>2024-12-24 16:07:13.613 SocketRPC: Removing old socket on "/run/user/1000/MumbleSocket"
>> <W>2024-12-24 16:07:13.617 AudioInput: Opus encoder set for high quality speech
>> <W>2024-12-24 16:07:13.617 AudioInput: 40000 bits/s, 48000 hz, 480 sample
>> [W][95537.973114] pw.conf      | [          conf.c: 1214 try_load_conf()] can't load config client.conf: No such file or directory
>> [E][95537.973154] pw.conf      | [          conf.c: 1243 pw_conf_load_conf_for_context()] can't load config client.conf: No such file or directory
>> [New Thread 0x7fffd1ab36c0 (LWP 214051)]
>> [W][95537.973513] pw.conf      | [          conf.c: 1214 try_load_conf()] can't load config client.conf: No such file or directory
>> [E][95537.973521] pw.conf      | [          conf.c: 1243 pw_conf_load_conf_for_context()] can't load config client.conf: No such file or directory
>> Thread 1 "mumble" received signal SIGSEGV, Segmentation fault.
>> [...]
>> (gdb) bt
>> #0  pw_thread_loop_stop (loop=0x0) at ../src/pipewire/thread-loop.c:351
>> #1  0x00005555558546d3 in PipeWireEngine::stop (this=0x55555686b200) at ./src/mumble/PipeWire.cpp:292
>> #2  PipeWireInput::~PipeWireInput (this=<optimized out>, this=<optimized out>) at ./src/mumble/PipeWire.cpp:335
>> #3  PipeWireInput::~PipeWireInput (this=<optimized out>, this=<optimized out>) at ./src/mumble/PipeWire.cpp:336
>> #4  0x00005555556cb685 in boost::detail::sp_counted_base::release (this=0x5555569d5770) at /usr/include/boost/smart_ptr/detail/sp_counted_base_gcc_atomic.hpp:120
>> #5  boost::detail::sp_counted_base::release (this=0x5555569d5770) at /usr/include/boost/smart_ptr/detail/sp_counted_base_gcc_atomic.hpp:116
>> #6  boost::detail::shared_count::~shared_count (this=<optimized out>, this=<optimized out>) at /usr/include/boost/smart_ptr/detail/shared_count.hpp:432
>> #7  boost::shared_ptr<AudioInput>::~shared_ptr (this=<optimized out>, this=<optimized out>) at /usr/include/boost/smart_ptr/shared_ptr.hpp:336
>> #8  boost::shared_ptr<AudioInput>::reset (this=<synthetic pointer>) at /usr/include/boost/smart_ptr/shared_ptr.hpp:727
>> #9  Audio::stop () at ./src/mumble/Audio.cpp:234
> 
> In https://bugs.debian.org/1089246 there are some patches which upgrade
> mumble to a newer version, which is also available here:
> https://salsa.debian.org/diederik/mumble/-/tree/update-to-1.5.735
> 
> There is an unstream commit included in v1.5.613 (and newer):
> b5a67c05fb54 ("REFAC: Fix tons of warnings and non-portable code")
> 
> That commit adds ``static_cast< std::uint8_t >`` to various statementsin
> PipeWire.cpp, among many other things, so that could be relevant here.
> 
> So, could you try if the problem is still there with a newer version?

That patch - only applied to PipeWire.cpp because it doesn't cleanly 
apply - does not suffice.

If you upload a new version to experimental, I'm happy to test with that.

Kind regards
Philipp Kern

[1] https://www.debian.org/Bugs/Developer#tags