Jeremy Dean wrote:
After rebuilding with these macros, everything seems to go fine, the library
builds ok. Now we are getting strange errors with the examples: alg4.cpp
It is failing on line 156 of include/set
The error is unable to instantiate from non templated ....
Unfortunately this is from a secure site, so not able to get exact error
message.
I'm afraid this isn't nearly enough detail for me to go on.
Have you been able to reproduce the error in your environment?
One thing you might want to configuring the library on your
machine and making the generated config header available to
your customer. He should then be able to compare it with his
and let you know what the differences are (and use your file
to try to resolve his problems). If the files are the same
chances are that the errors are due to a difference between
his compiler and yours.
Martin
Getting similar error with example tele.cpp failing on line 349 ofr
include/map
Jeremy
-----Original Message-----
From: Martin Sebor [mailto:[EMAIL PROTECTED]
Sent: Tuesday, January 31, 2006 9:58 AM
To: [email protected]
Subject: Re: Problem building on Solaris
Jeremy Dean wrote:
I am building on Solaris 10 with the Sun Studio 10 compiler and
getting this
error:
CC -D_RWCONFIG=15dECB -I./../../../include -I./../../.. -I.. -I.
-D_RWBUILD_std -g -D_REENTRANT -mt -PIC -library=%none -c
../bitset.cpp
CC -D_RWCONFIG=15dECB -I./../../../include -I./../../.. -I.. -I.
-D_RWBUILD_std -g -D_REENTRANT -mt -PIC -library=%none -c ../char.cpp
CC -D_RWCONFIG=15dECB -I./../../../include -I./../../.. -I.. -I.
-D_RWBUILD_std -g -D_REENTRANT -mt -PIC -library=%none -c
../codecvt.cpp
"./../../../include/loc/_rw_setlocale.h", line 53: Error: _RWSTD_LC_ALL
is
not defined.
1 Error(s) detected.
*** Error code 1
make: Fatal error: Command failed for target `codecvt.o'
This is on a secure machine and can not install any locales on it:
locale -a
POSIX
C
iso_8859_1
Is there any work around for this problem?
The most likely cause is that the LOCALE_NAME_FMAT.cpp configuration test
fails to #define the macro (the test is designed to define these even if no
locales are installed so what you're seeing is definitely a bug). I created
STDCXX-128 for it: http://issues.apache.org/jira/browse/STDCXX-128
The easiest workaround is to grab the macro definitions from a test run a
machine that does have some locales installed. Here's what I get on one of
our Solaris 10 boxes:
#define _RWSTD_LC_COLLATE 3
#define _RWSTD_LC_CTYPE 0
#define _RWSTD_LC_MONETARY 4
#define _RWSTD_LC_NUMERIC 1
#define _RWSTD_LC_TIME 2
#define _RWSTD_LC_MESSAGES 5
#define _RWSTD_LC_ALL 6
#define _RWSTD_LC_MAX _RWSTD_LC_ALL
#define _RWSTD_LC_MIN _RWSTD_LC_CTYPE
#define _RWSTD_CAT_3(pfx) { 3, "LC_COLLATE", pfx::_C_collate } #define
_RWSTD_CAT_0(pfx) { 0, "LC_CTYPE", pfx::_C_ctype } #define _RWSTD_CAT_4(pfx)
{ 4, "LC_MONETARY", pfx::_C_monetary } #define _RWSTD_CAT_1(pfx) { 1,
"LC_NUMERIC", pfx::_C_numeric } #define _RWSTD_CAT_2(pfx) { 2, "LC_TIME",
pfx::_C_time } #define _RWSTD_CAT_5(pfx) { 5, "LC_MESSAGES",
pfx::_C_messages } // #define _RWSTD_NO_SETLOCALE_ENVIRONMENT #define
_RWSTD_NO_CAT_NAMES #define _RWSTD_CAT_SEP "/" #define _RWSTD_NO_CAT_EQ //
#define _RWSTD_NO_INITIAL_CAT_SEP // #define _RWSTD_NO_CONDENSED_NAME
Martin
