Re: [Bug-gnubg] bad __attribute__ handling breaks w/newer standards, newer gcc, & glibc and causes infinite loop

[Michael Petch]

On 2015-10-07 16:32, Michael Petch wrote:

> Scrap this since Bison's original code technically isn't redefining
> __attribute__ but declaring YY_ATTRIBUTE that may use __attribute__ (or
> not).
> 
I could modify it to do it as Russ suggests, but there may be another
way. Autoconf archive has this
http://git.savannah.gnu.org/gitweb/?p=autoconf-archive.git;a=blob_plain;f=m4/ax_c___attribute__.m4
:

    AX_C___ATTRIBUTE__

I could use this in our configure process. If a platform/compiler
supports __attribute__ it defines HAVE___ATTRIBUTE__ . Then in our code
we could simply do:

#ifndef HAVE___ATTRIBUTE__
#define __attribute__(Spec) /* Empty */
#endif /* HAVE___ATTRIBUTE__ */

-- 
Michael Petch
CApp::Sysware Consulting Ltd.
OpenPGP FingerPrint=D81C 6A0D 987E 7DA5 3219 6715 466A 2ACE 5CAE 3304