Re: Where to start in unix/matlab error message?
From: Peter Boettcher (boettcher@ll.mit.edu)
Date: 04/08/03
- Next message: Chris Cox: "Re: bash: WHY???"
- Previous message: Simon P. Grasser: "Re: solaris as a router"
- In reply to: Fred Ma: "Re: Where to start in unix/matlab error message?"
- Next in thread: Fred Ma: "Re: Where to start in unix/matlab error message?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
From: Peter Boettcher <boettcher@ll.mit.edu> Date: Tue, 08 Apr 2003 17:45:39 -0400
Fred Ma <fma@doe.carleton.ca> writes:
> Fred Ma wrote:
>
>> Hello,
>>
>> I'm running the math/science computing package Matlab(R13)
>> with my own C++ callable functions. I'm working in solaris 8
>> using gcc 3.2.1, gdb 5.3. I'm going by the procedure of
>> http://www.mathworks.com/support/solutions/data/31319.shtml.
>>
>> I'm getting an error message that is pretty cryptic to say the
>> least:
>>
>> Program received signal SIGUSR1, User defined signal 1.
>> 0xfded9794 in __sigprocmask () from /usr/lib/libthread.so.1
>>
>> This is after I set a break point on the top level function of my
>> dynamically callable function ("break mexFunction" in step 6
>> of the procedure) and try to continue. When I get the above
>> message, I am certainly not at the breakpoint. A backtrace
>> shows lots of stuff (attached).
>>
>> Of course, no one can debug this for me, but is there anyway
>> to get started in figuring out what this error is about? I'm not
>> all that keen on going off on a tangent to learn about vast
>> body of knowledge on unix under-the-hood, because I have
>> to stay focused on my thesis research. But if it is a small
>> detour to pick up some handy tools for my kind of work, I'd
>> certainly like to take a look at it.
>
> I should clarify, the program runs fine when I'm not using the
> debugger. Fine meaning no crashes, illegal memory accesses,
> etc..
This might be the mechanism used by dbmex to signal the debugger to
interrupt execution long enough to set the breakpoint. What if you
just try continuing from there? Can you continue long enough to turn
off dbmex again?
You might be able to disable gdb's trapping of SIGUSR1. From the gdb
prompt, try: handle SIGUSR1 nostop noprint
Finally, this might be a compiler version mismatch, which is likely
to cause problems when combined with C++.
-- Peter Boettcher <boettcher@ll.mit.edu> MIT Lincoln Laboratory MATLAB FAQ: http://www.mit.edu/~pwb/cssm/
- Next message: Chris Cox: "Re: bash: WHY???"
- Previous message: Simon P. Grasser: "Re: solaris as a router"
- In reply to: Fred Ma: "Re: Where to start in unix/matlab error message?"
- Next in thread: Fred Ma: "Re: Where to start in unix/matlab error message?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Relevant Pages
|
