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] fortran/28548 -- Optional argument fails.


On Fri, Aug 04, 2006 at 06:27:43PM +0200, Paul Thomas wrote:
> 
> OK with the proviso that the error is turned into a warning; yes, I 
> think that Martin is correct about the standard stating that this is 
> only an error if the argument is missing.   Applying a runtime error 
> looked horribly messy, although a library function might be the way 
> around it.   I have suggested a new wording in a Bugzilla atachment to 
> the PR.  This means that the dg-error's in the original testcase must be 
> changed to dg-warning's.  Would it be an idea to make this conditional 
> on -pedantic too?

I just grabbed your patch.  I'll add the -pedantic conditional.
After regression testing, I apply the patch.

> I do apologise for not reading the section 12.4.1.5 carefully enough and 
> thank you, Steve, for covering my a**e whilst I was away.  An upside is 
> that we have all learned that attr.elemental is not always what it 
> appears to be!

This certainly applies to the conversion functions, which
are given the elemental attribute but not a formal argument
list.  I don't know if it applies to any other procedure.
Also, note that you do not want to remove elemental from
the conversion functions.  It gets ugly.

-- 
Steve


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