Mail Archives: cygwin/2000/02/17/17:53:18
I was able to hack up a solution that allowed me to catch a C++
exception thrown from inside a DLL. I put my technique up on the web:
http://www.cs.washington.edu/homes/yngblut/ehtest/
I hope other people find this useful.
Eric Youngblut
yngblut AT cs DOT washington DOT edu
PS Thanks to all of the Cygwin and Mingw32 developers! I appreciate
your work tremendously.
Mumit Khan <khan AT xraylith DOT wisc DOT EDU> wrote:
>
>Paul Henshaw <paul DOT henshaw AT jrc DOT it> writes:
>> Dear All,
>>
>> having managed now to build and link against a DLL, I have
>> encountered a problem with exception handling.
>>
>> I have a test program which deliberately provokes an exception,
>> expecting to catch it, and throw it to it's caller.
>>
>> When I run the DLLed version of this program, the exception
>> is constructed and displayed correctly, but the caller does
>> not catch the exception, and my SIGABORT handler is called.
>
>This is a problem that's been there since the beginning, and will be
>there for a while to come, sorry. Nobody is actively working on this
>part, at least publicly.
>
>The long term solution is rewrite the C++ exception mechanism for
>windows32 ports using SEH.
>
>> So is there is some magic incantation to chant over DLLs
>> to make exceptions work, or am I just being stupid (again)?
>
>I'll be grateful if someone does find a workaround for this issue.
>
>btw, the problem shows up on other platforms as well -- such as HPUX
>-- where exceptions thrown by shared libraries can't be caught. Not
>always an easy problem to solve, especially on window32, which GCC
>hackers just don't have much of an incentive to work on.
>
>Regards,
>Mumit
--
Want to unsubscribe from this list?
Send a message to cygwin-unsubscribe AT sourceware DOT cygnus DOT com
- Raw text -