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: fix regparm testcases


Michael Matz <matz@suse.de> writes:

> Hi,
>
> my patch for checking regparm attribute matching for different function
> decls triggered some errors in the testsuite, which I didn't notice
> because I've run the testsuite only on amd64.  Not very clever for a
> feature of this nature :-/
>
> The below patch fixes them.  Okay?  (tested on 3.4 and head on i686 this 
> time ;) )

Do we *want* to require people to write the regparm attribute in both
places?  I would think the natural thing would be to object only to
inconsistent specifications; if the definition doesn't have any
regparm attribute, just bring in the setting from the declaration.

zw


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