Subject: Re: [ecasound] segfault backtrace
From: Kai Vehmanen (k_AT_eca.cx)
Date: Wed Oct 16 2002 - 18:27:51 EEST
On Mon, 14 Oct 2002, Mario Lang wrote:
> After attaching gdb to the ecasound process which runs inside
> emacs, issueing `continue' at the gdb prompt and then entering preset-register
> in ecasound, gdb reports:
Hmm, this is a bit tricky. I suspect this is caused by a bug in
the C++ compiler or the C++ runtime. I'm not 100% sure, but there
doesn't seem to be any other good explanation for this. I've
received quite a few similar reports, where a bug is triggered
during the RESOURCE_FILE code path. As is the case in the trace you
posted, the actual crash usually happens inside the C++ runtime's
(libstdc++) string handling functions.
The problem is that in all the bug scenarios I've traced, ecasound doesn't
really do anything wrong. The libstdc++ just crashes... :(
> #1 0x4041ead9 in kvu_remove_surrounding_spaces (a=@0xbfffea04)
> at /usr/lib/gcc-lib/i386-linux/2.95.4/../../../../include/g++-3/std/bastring.h:229
> #2 0x4021c445 in RESOURCE_FILE::load (this=0xbfffedb4) at resource-file.cpp:60
Hmm, you could try temporarily renaming your ~/.ecasound/ecasoundrc to
some other name and see if this helps.
-- http://www.eca.cx Audio software for Linux!
This archive was generated by hypermail 2b28 : Wed Oct 16 2002 - 18:26:10 EEST