[PATCH] caret diagnostics (was: broken FE diagnostics wrt complex expressions)

[Joseph S. Myers]

On Thu, 14 Aug 2008, Manuel López-Ibáñez wrote:

> It is controlled by -fdiagnostics-show-caret. See the diff for gcc/opts.c.
> 
> The configure options are meant to enable/disable all code related to
> caret printing in a similar way as it was done with mapped locations.
> This was requested the first time I sent this patch because it was
> considered too experimental to have it even with
> -fno-diagnostics-show-caret as the default.

Mapped locations were sitting around on trunk for a long time as a 
bitrotten feature; I don't think that's a good example here.  When the 
code works and passes review it should go on trunk, default remaining the 
present diagnostic style according to the GNU Coding Standards, no 
configure option.

I don't think the option should necessarily just be boolean; once choice 
that may make sense would be caret diagnostics for the first diagnostic 
from an input file only, to avoid blowing up the output size when one 
mistake causes a cascade of diagnostics.  (This is a matter of designing 
the option as e.g. -fdiagnostics-show-caret={no,yes,first} rather than as 
-f/fno-, not a matter of needing such a feature implemented in the first 
version going on trunk.)

-- 
Joseph S. Myers
joseph@codesourcery.com