[GRASS-dev] GRASS & QGIS: the future

[Glynn Clements]

Radim Blazek wrote:

> > Can you try the attached patch? The new function should be used like:
> >
> >         if (setjmp(*G_fatal_longjmp(1))) {
> >             // this will be executed on fatal errors
> >         }
> 
> Works. Can I hope the patch will find a way to 7.0.0?

Committed in r60590.

The only aspect which might be invasive is that defs/gis.h now
includes <setjmp.h> for the jmp_buf declaration. Although this is a
standard C header, it's not necessarily as straightforward as other
headers, being tied to the compiler's internals. I would advise
checking whether there are any build issues on less-common platforms
(e.g. AIX) before promoting the change to release.

In the worst case, we can change the return type to void* and require
an explicit cast to jmp_buf* at the call site, meaning that <setjmp.h>
only needs to be included from lib/gis/error.c plus any file which
actually calls the function.

-- 
Glynn Clements <glynn at gclements.plus.com>