This is the mail archive of the
mailing list for the GCC project.
Re: Patch: fix column numbers in C parser errors
>>>>> "Daniel" == Daniel Jacobowitz <firstname.lastname@example.org> writes:
Daniel> On Tue, Nov 06, 2007 at 09:36:36PM -0700, Tom Tromey wrote:
>> I adopted a simple rule while writing this patch: all calls to error
>> functions (error, warning, and the pedwarn family) in c-parser.c must
>> provide a location using %H. This is simple to understand and forces
>> the author of the call to think about which location is appropriate.
Daniel> I'm sure this has come up before, but are there any error
Daniel> reporting functions that take the location directly without
Daniel> the %H?
Not that I know of.
Daniel> Should there be?
I'd prefer that this be the only kind of error reporting
function... but that is a lot of work.
Daniel> Not only will this invalidate a lot of existing translations,
Daniel> but it's also pretty confusing to read the new strings, IMVHO.
For the latter, there's always error ("%H" "The rest of the message").
That's confusing in another way, I suppose.