[aspectc-user] Execution advice for variable argument list functions
aspectc at stefan-marr.de
aspectc at stefan-marr.de
Wed May 14 08:30:10 CEST 2008
Hi Olaf,
> the bug is not fixed yet. The reason is that it is not very easy to fix
> it.
>
> <snip/>
> code would have to be found and removed. It is not impossible but
> difficult.
Ok thanks, for now I've changed the functions to use a sufficent known
number of arguments. Is not nice but will work.
>
> BTW, you can turn off the "annoying warnings" if you know what they
> mean and have decided that it is not relevant for you.
Since it is a C program with OO emulation, the compiler generates a lot
warnings about classes defined by macros. Unfortunatly, I have not found the
right switch at the gnu gcc manual page.
Thanks a lot for this nice tool, works great for me :)
Are there any plans to build a new offical release (may be RC) from the
current SVN sources?
Kind Regards
Stefan
>
> Best regards,
>
> Olaf
>
> Stefan Marr wrote:
> > Hello,
> >
> > I'm currently using AspectC++ to modularize a C program.
> > After some small adjustments it compiles fine with g++ and
> fortunately
> > even with Ag++ :)
> > First, I got some problems with the used macros which emulate object
> > orientation in C with virtual tables and stuff, but now I switched to
> > the latest SVN source of AspectC++ and it work except some annoying
> > warnings, cause the object layout is defined using macros.
> >
> > But, now I'm trying to use aspects on functions with variable number
> > of arguments and I got into trouble.
> >
> > Have found the following bug report
> > http://www.aspectc.org/bugzilla/show_bug.cgi?id=106
> > mentioning that execution advices for functions with variable
> argument
> > lists is not working.
> >
> > Is this still true? And if so, would it be easy to fix? If you give
> me
> > a helping hand I would try to fix this.
> > At the moment I would need this feature, since some of the functions
> > we are using have variable argument list and are called through our
> OO
> > emulation and function pointers in a vtable. Call advices aren't
> > applied there.
> >
> > The execution advice seams to be applied in the right place, but the
> > generated code is wrong. Main problem is, that tjp seams to be
> defined
> > as an "int" and therefore tjp->proceed(); fails.
> >
> > Kind Regards
> > Stefan Marr
> >
> > _______________________________________________
> > aspectc-user mailing list
> > aspectc-user at aspectc.org
> > http://www.aspectc.org/mailman/listinfo/aspectc-user
> >
More information about the aspectc-user
mailing list