Subject: Re: [ecasound] ecasound hangs when adding chain.
Date: Fri Jun 30 2000 - 18:46:00 EEST
>Yes, this is a __very__ common problem. I've received numerous
>problem reports (-a:something causing crashing), but so far
>I haven't been able to reproduce the error. Needless to say,
>I'm quite anxious to resolve this.
>> I did a trace with gdb and find that the program appears to be
>> hanging in a libstdc++ call:
>> (gdb) bt
>> #0 0x4038dc80 in basic_string<char, string_char_traits<char>,
>> __default_alloc_template<true, 0> >::Rep::clone () from
>Yep, this looks bad. Trying out different versions of libstdc++ would be the
>first thing I'd try myself. Compiling libstdc++ from sources
>is quite painful, so I suggest to use binary-RPMs or debs.
So I noticed. It appears that Im using the latest libstdc++ for my
So its a little hard to replace the libstdc++.a with an older version.
Im also using egcs-2.91.66
It appears that the only newer version is libstdc++-3 (2.90.8)
from the GNU site.
Unfortunately my c++ knowledge is limited to probably only being able
to put cout's before and after all the clone lines I can find in the
If you need some compile time on my machine while Im online, to help debug it
I'll give you a login on my machine. Since the problem is reproducible here.
Also I cant see where add_new_chains calls clone... I guess im just showing
my c++ inability.
I've also attempted to use the binary distribution. Though it is linked
with ncurses 4, I assumed was the libc5 version. I appear to only have
ncurses 5 on my machine.
( I had some nasty effects when ncurses 4 was on the machine, as it
also wanted libc5 dynamically linked having both libc5 and libc6 wanting
to be linked in the single binary tends to cause core dumps)
-- To unsubscribe send message 'unsubscribe' in the body of the message to <email@example.com>.
This archive was generated by hypermail 2b28 : Fri Jun 30 2000 - 18:46:56 EEST