This is the mail archive of the gcc-patches@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]
Other format: [Raw text]

Re: [PATCH] PR optimization/13472


On Tue, 23 Dec 2003, Eric Botcazou wrote:

> So I think the fix is to remove decl_constant_value_for_broken_optimization
> altogether from the C front-end.  But this is probably not doable on a 
> release branch."

I'd have expected that some stuff might break (albeit rather broken
anyway, linking in to general constant expressions issues) when simply
removing it like this without reinstating checks elsewhere.  But because
for technical reasons constant expression constraints should not depend on
-pedantic, the constant expressions testcases don't use it.  The most
relevant testcases are gcc.dg/c9[09]-const-expr-3.c.  Using -pedantic on
those testcases gets rid of four out of seven of the expected failures.  
Does your patch bring back any of those failures with -pedantic?  (If it
does, this may not be an argument against your patch, since it's still
broken to depend on -pedantic for *avoiding* diagnostics, and correct
diagnostics still should not need to impede optimisation.)

-- 
Joseph S. Myers
jsm@polyomino.org.uk


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