In one of my d-projects I get an access violation as soon as I compile th project with -inline. From the looks of it this seems to be a codegen issue and not a issue of my code. Any ideas how I can reduce this using dustmite? If I just check for any access violation I will most likely not get the result I want.

-- 
Kind Regards
Benjamin Thaut

On Thursday, 24 October 2013 at 05:06:23 UTC, Benjamin Thaut wrote:
> In one of my d-projects I get an access violation as soon as I compile th project with -inline. From the looks of it this seems to be a codegen issue and not a issue of my code. Any ideas how I can reduce this using dustmite? If I just check for any access violation I will most likely not get the result I want.

If you are on Linux/OS X, try using a script that runs the executable in GDB and then greps for the right functions to appear in the backtrace. It's obviously not a hundred percent precise, but has worked well for me in the past. I think there also is an example for this in the Dustmite docs.

David

Am 24.10.2013 09:50, schrieb David Nadlinger:
> On Thursday, 24 October 2013 at 05:06:23 UTC, Benjamin Thaut wrote:
>> In one of my d-projects I get an access violation as soon as I compile
>> th project with -inline. From the looks of it this seems to be a
>> codegen issue and not a issue of my code. Any ideas how I can reduce
>> this using dustmite? If I just check for any access violation I will
>> most likely not get the result I want.
>
> If you are on Linux/OS X, try using a script that runs the executable in
> GDB and then greps for the right functions to appear in the backtrace.
> It's obviously not a hundred percent precise, but has worked well for me
> in the past. I think there also is an example for this in the Dustmite
> docs.
>
> David

I'm on windows ;-)

-- 
Kind Regards
Benjamin Thaut