This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: branch bootstrap breakage in new DECL_RTL


 > From: Mark Mitchell <mark@codesourcery.com>
 > 
 > >>>>> "Kaveh" == Kaveh R Ghazi <ghazi@caip.rutgers.edu> writes:
 > 
 >     Kaveh> Its frustrating to me that people are ignoring the freakin'
 >     Kaveh> warnings from gcc.  Here's another example of a bootstrap
 >     Kaveh> failure that didn't need to happen.  :-(
 > 
 > And shame on me, since I'm the one that caused it.
 > 
 > Your criticism is fair -- but I will note that it was one warning
 > among many.  Despite your efforts, the warning count is creeping
 > upwards, again.
 > 
 > This is precisely where automated pre-commit testing would win: I made
 > a relatively innocent mistake, not so much out of carelessness as out
 > of the inability to test properly, not having a non-GCC compiler
 > available at the time.  A test server would have tested, found the
 > problem, and it would have taken me only minutes to correct the patch
 > -- all without making poor Robert have to track it down.
 > 
 > Perhaps we should make -pedantic-errors the default.  What are the
 > chances of surviving a bootstrap with that switch on most systems?
 > 
 > That said, there's no excuse.  I apologize.


Thanks, I appreciate that.  I'm very interested in an automated
solution. :-)

I've outlined in other emails in the past that basically -Werror and
-pedantic-errors are not workable solutions and will actually
*increase* bootstrap problems trying to get to zero warnings on all
platforms (at all times.)

Instead, the best thing we could do would be to hook
contrib/warn_summary into Geoff's regression checker so that we exert
downward pressure on warnings and spank people who introduce new ones
without actually causing breakage in bootstrapville.  If someday we
actually get to zero, so much the better.

Geoff mentioned he'd look into this, but I haven't heard whether he
was successful yet or not.

--
Kaveh R. Ghazi			Engagement Manager / Project Services
ghazi@caip.rutgers.edu		Qwest Internet Solutions


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]