gcc builds broken on dev: mbstate_t not in std

[Liviu Nicoara <ni...@roguewave.com>]

Martin,

As we discussed it yesterday the gcc builds are broken on dev because of
  a glitch introduced in [I believe] revision 366948.

AFAICT the else branch of the new SunOS conditional in rw/_mbstate.h
does not preserve the former logic.

The following is a tentative patch; could you please take a look at it
and see if I understood correctly the problem?

Thanks,
Liviu

Index: include/rw/_mbstate.h
===================================================================
--- include/rw/_mbstate.h       (revision 367667)
+++ include/rw/_mbstate.h       (working copy)
@@ -122,7 +122,19 @@
 #    endif   // _RWSTD_NO_USING_LIBC_IN_STD

 #  else   // if !defined (_RWSTD_OS_SUNOS)
+
 #    include _RWSTD_CWCHAR
+
+#    ifndef _RWSTD_NO_USING_LIBC_IN_STD
+
+namespace std {
+
+using ::mbstate_t;
+
+}   // namespace std
+
+#    endif // _RWSTD_NO_USING_LIBC_IN_STD
+
 #  endif   // _RWSTD_OS_SUNOS

 #elif !defined (_RWSTD_MBSTATE_T_DEFINED)

Re: gcc builds broken on dev: mbstate_t not in std

[Martin Sebor <se...@roguewave.com>]

Liviu Nicoara wrote:
> Martin,
> 
> As we discussed it yesterday the gcc builds are broken on dev because of
>   a glitch introduced in [I believe] revision 366948.

Yeah, I'm afraid I broke it for at least some platforms :(

> 
> AFAICT the else branch of the new SunOS conditional in rw/_mbstate.h
> does not preserve the former logic.
> 
> The following is a tentative patch; could you please take a look at it
> and see if I understood correctly the problem?

The cruft in this file is really brittle. I'm working on a hopefully
more robust patch. If I don't commit it tonight I should be done with
it sometime tomorrow.

Sorry about that.
Martin