Manuel López-Ibáñez wrote:
The locations are as wrong now as they will be after enabling caret
diagnostics. The only difference is that no one pays attention right
now because that would require to enable -fshow-column and count
column numbers. So I am certainly not "breaking the compiler and
hoping that other developers will fix it".
OK, instead of "break", I will say that you are hoping to change the
user-visible behavior in a way that makes a current limitation obvious,
in the hope that will prod people to fix it. I don't think that's
appropriate.
Having cc1 print a statistic (when not passed "-quiet") saying "I got
error messages wrong 100 times in this translation unit" would be fine
(if it were possible to compute) since it wouldn't get in anyone's way.
Your change will inconvenience people.
Please attack the problem directly, by working with others to improve
the location information in the compiler. My experience has been that
if you get the ball rolling, other people will get behind and help, when
they see that the problem is tractable.
Thanks,
--
Mark Mitchell
CodeSourcery
[EMAIL PROTECTED]
(650) 331-3385 x713
