[v3] solaris c include model

Marc Glisse marc.glisse+gcc@normalesup.org
Wed Apr 11 12:18:00 GMT 2007

On Wed, 11 Apr 2007, Benjamin Kosnik wrote:

> In a nutshell, this introduces a new "C" header model, ie
> --enable-cheaders=c_std_solaris. In this patch, this is the default
> for solaris 2.7+, which should probably be tested or tweaked.

Thanks a lot for doing this work.
(precision: I won't have much access to any solaris system for a long
time, so I can only comment from what I remember of my earlier attempts)

> To get this to work, you'll need to:
> 1) patch cc1plus so that __cplusplus is the correct value. (Also the
> oldest libstdc++ PR)
> 2) patch gcc/ginclude/stddef.h
> 3) do some fixincludes on the solaris side, where extern "C" and
> extern "C++" decls clash with gcc but not, apparently, with sunpro.
> See the attached for a manual way to do this: I've not yet fought with
> fixincludes for this.
> 4) patch libstdc++ with the mega-patch attached.
> There are a lot of hacks in the libstdc++ bits here (strdup on
> solaris??? How do I get this back? stddef.h is not figured out, and
> it's apparent that better separation between C90 and C99 needs to take
> place in wchar/wctype), but oh well. I'll try to fix this up when I
> return.
> In addition, this includes the libstdc++ parts of the newlib work.
> Which is not finished yet either.
> To intrepid solaris hackers, I say "bonne chance." It will be
> interesting to see if you exist.

Some comments in random order:

* for qsort and bsearch, the problem comes from bug 2316. Your fix seems
   strange, you now declare (for recent __cplusplus) only the extern c++
   version of bsearch, which is defined nowhere. I would rather simply
   remove the extern "C++" declaration and keep the other one as is until
   2316 is fixed (not as violent as what I was mentionning in 30928. By
   the way 30928 shows an other approach that gives a more uniform result
   across solaris 8-11 but does not use the system headers as nicely as
   you are trying to do). If 2316 is fixed then an extern "C++" version of
   bsearch and qsort will need to be provided probably in libsupc++ or
   some such.

* in math.h (for instance), you could #include <cmath> and #include_next
   <math.h>, not simply the second one (not sure right now why I wanted
   that, maybe just so cos(2) works after including math.h)

* a number of things become useless in the headers since you assume
   solaris 7+ (like no stdlib.h or similar things)

* I am surprised you have not hit bug 30112 (only gives 4 failures in the
   testsuite) and more importantly bug 27340 which should make valarray

* for C99 functions you are going to have a hard time getting something
   nice until solaris updates its headers some time after the standard is

Good luck for continuing this work, and once again thank you.

Marc Glisse

More information about the Libstdc++ mailing list