Re: Solaris 10 and gcc update

[Curt Arnold]

The failure in test9 is likely happening on the following line:
  std::locale localeUS(LOCALE_US);
where LOCALE_US = "en_US".  Do you have any documentation on the  
names
recognized for std::locale constructors on Solaris?

I am pretty sure en_US is a valid locale.  I went to the /usr/lib/ 
locale and
there is a en_US directory.  Any other thoughts on what to try??

Nope.  That isn't an essential test, I'd just comment it out in the  
CPPUNIT_TEST block at the top and we will dig into it later.



When I tried to build without the -Dlogchar=utf8 -Dhas.wchar_t=0
command
lines, I got the following link error.
build-shortsocketserver:
    [mkdir] Created dir:
/export/disk10/cppProjects/logging-log4cxx/build/debug/shared/
shortsocketser
ver_obj
       [cc] 2 total files to be compiled.
       [cc] Starting link
       [cc] Undefined                   first referenced
       [cc]  symbol                         in file
       [cc] log4cxx::helpers::UnicodeHelper::encodeWide(unsigned int,
wchar_t*)/export/disk10/cppProjects/logging-log4cxx/build/debug/
shared/liblo
g4cxx.so
       [cc] log4cxx::helpers::UnicodeHelper::decodeWide(wchar_t
const*&,
wchar_t
const*)/export/disk10/cppProjects/logging-log4cxx/build/debug/shared/
liblog4
cxx.so
Any suggestions on what I could do to resolve this?


Those methods have been suppressed since you specified that the
compiler did not have a wchar_t type with the -Dhas.wchar_t=0.  I
assume that Solaris gcc actually has a wide character type and you
could turn it on and the problem would go away.  However, I should
review shortsocketserver to see why it doesn't build when wchar_t is
not present.
I got this error when I used the ant -Dapr.dir=/usr/local/apr
-Dcppunit.dir=/export/disk10/cppProjects/cppunit-1.10.2
Looking at the unicodehelper.cc, I'm not sure the __STD_ISO_10646__ is
defined.  I think this is why then encodeWide is not defined.  I did a
search through various header files and the STD_ISO_10646 was no  
where to be
found.


__STD_ISO_10646__ is defined by the compiler when the wchar_t  
contains UCS-4 code points (that is 4 byte unicode scalar values).   
encodeWide and decodeWide will be compiled for WIN32 (where wchar_t  
is UTF-16) and when __STD_ISO_10646__ is defined (where wchar_t is  
UCS-4).  Do you have a description of wchar_t is on that platform and  
how it can be detected.