This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: implement fdiagnostics-print-source-range-info
2009/8/4 Jakub Jelinek <jakub@redhat.com>:
>
> Shouldn't normal location_t be instead always a range
> (well, table index which would expand to a begin/end location pair
> rather than just one location)?
> I think adding location_range_t is a wrong direction.
>
Most of the time we just need a single location, ranges are for
diagnostics that mention several related things. On the other hand,
expressions could probably have a location range instead of a single
location, since they are likely to be "wide".
In any case, the change you are proposing is massive, touching
everywhere in the compiler, and given GCC's lack of modularity and
interfaces, I find it rather difficult, if not impossible. This patch
on the other hand is minimally invasive (although not perfect at all).
Cheers,
Manuel.