[fluid-dev] Fun dbus crash

[Corbin Simpson]

Hey all,

I know I've been super-silent on the FS front, but I got the bug in me
today, so I revived my Python FS bindings.

While hacking, I discovered this delightful crash:

$ python test.py
fluidsynth: warning: Failed to pin the sample data to RAM; swapping is possible.
*** glibc detected *** python: invalid fastbin entry (free): 0xb0200708 ***
======= Backtrace: =========
/lib/libc.so.6[0x48c51d22]
/lib/libdbus-1.so.3(dbus_free+0x20)[0x49311ed0]
/lib/libdbus-1.so.3[0x49312785]
/lib/libdbus-1.so.3[0x492feb74]
/lib/libdbus-1.so.3(dbus_message_iter_append_basic+0x266)[0x49300b66]
/lib/libdbus-1.so.3(dbus_message_new_error+0x111)[0x49302841]
/lib/libdbus-1.so.3[0x49305ae5]
/lib/libdbus-1.so.3(dbus_connection_send_with_reply+0x105)[0x492f4605]
/lib/libdbus-1.so.3(dbus_connection_send_with_reply_and_block+0xb4)[0x492f48f4]
/lib/libdbus-1.so.3(dbus_bus_register+0x122)[0x492ef382]
/lib/libdbus-1.so.3[0x492ef6b3]
/usr/lib/libfluidsynth.so.1(fluid_rtkit_make_realtime+0x3f)[0x493ea3bf]
/usr/lib/libfluidsynth.so.1(fluid_thread_self_set_prio+0x70)[0x493f5040]
/usr/lib/libfluidsynth.so.1[0x493f5086]
/lib/libglib-2.0.so.0[0x48e88f45]
/lib/libpthread.so.0[0x48d90cd3]
/lib/libc.so.6(clone+0x5e)[0x48cce83e]

fluid_rtkit_make_realtime() calls some dbus stuff, which appears to
trigger a bad free() in libdbus. Anybody else seen this before? Am I
doing something stupid or is this dbus's fault?

In the meantime, is there an easy way to hack around this in settings,
or do I need to rebuild FS without dbus support?

~ C.

-- 
When the facts change, I change my mind. What do you do, sir? ~ Keynes

Corbin Simpson
<address@hidden>