Skip to Content.
Sympa Menu

shibboleth-dev - Re: [Shib-Dev] log4shib compile error

Subject: Shibboleth Developers

List archive

Re: [Shib-Dev] log4shib compile error


Chronological Thread 
  • From: "E. Stuart Hicks" <>
  • To:
  • Subject: Re: [Shib-Dev] log4shib compile error
  • Date: Mon, 09 Jun 2008 13:32:27 -0400
  • Organization: OhioLINK

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

...and here's a working version of the patch that doesn't doesn't have
weird end-of-line errors in the Clock.cpp part. :-\


E. Stuart Hicks wrote:
| Attached is a patch file for Tru64. It seems to work on all 3 servers
| and all of the 'make check' tests pass; however, it's still kind of a
| hack job and there are probably better solutions. If anyone is
| interested in addressing this and would like to see the actual
| compilation output then let me know.
|
|
| ----------------------
| E. Stuart Hicks
| Access Manager / Systems Engineer
| OhioLINK
|

|
|
| Scott Cantor wrote:
| |> I get this error when I try and compile log4shib on a Tru64-
| |> based Alpha server with gcc 3.4.6:
| |
| | Tru64 isn't a supported platform, so if you're expecting anything to
| "just work", you'll need to adjust your expectations.
| |
| | Going further, it's:
| |
| | - not been attempted in years
| | - not related to any platform that is supported
| | - not something anybody involved has access to, so is impossible to
| support on an ongoing basis
| |
| |> g++: strict_ansi_errors: No such file or directory
| |> cc1plus: error: unrecognized command line option "-std"
| |
| | Apparently the makefile (or really configure) won't support it without
| changes. Possibly trivial ones that you should be able to see how to
| make pretty easily. If not, you're probably going to need help porting.
| |
| |> Is this something that I need to take up with the log4cpp people or
| |> could it have something to do with the log4shib fork itself?
| |
| | log4cpp doesn't build much of anywhere other than Linux, so you can do
| the math. That's why I forked.
| |
| |> And, for that matter, is it even worth trying to compile the 2.0 SP
on a
| |> 64-bit box yet?
| |
| | It's not a 64-bit problem.
| |
| | -- Scott
| |
| |
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.9 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAkhNaSsACgkQqmBbivzHElIqswCdGfPBniyIQJIadOGwa0elIlwB
ClkAoJgU7DwwSD+mYGqHEdJ716+QJhKd
=7icj
-----END PGP SIGNATURE-----
diff -ur log4shib-1.0/configure log4shib-1.0-tru64/configure
--- log4shib-1.0/configure 2007-11-10 00:01:38.000000000 -0500
+++ log4shib-1.0-tru64/configure 2008-06-09 13:24:22.000000000 -0400
@@ -21532,14 +21532,6 @@
LOG4SHIB_LIBDEPS="$LIBS"
LOG4SHIB_VERSION="$VERSION"

-# CXX fine tuning
-case "$host" in
- *-dec-osf*)
- CXXFLAGS="$CXXFLAGS -std strict_ansi_errors"
- ;;
- *)
- ;;
-esac


if test "$GCC" = yes; then
diff -ur log4shib-1.0/tests/Clock.cpp log4shib-1.0-tru64/tests/Clock.cpp
--- log4shib-1.0/tests/Clock.cpp 2007-10-21 17:24:34.000000000 -0400
+++ log4shib-1.0-tru64/tests/Clock.cpp 2008-06-09 13:26:24.000000000 -0400
@@ -6,9 +6,7 @@

#include <cstdlib>
#include <sys/time.h> // for struct timeval
-#ifdef __osf__
-# include <machine/builtins.h> // for __RPCC()
-#elif __linux__ && __i386__
+#ifdef __linux__ && __i386__
# define rdtscl(low) \
__asm__ __volatile__("rdtsc" : "=a" (low) : : "edx")
#endif
@@ -34,9 +32,7 @@
warn = false;
}

-#ifdef __osf__
- return (usec_t) __RPCC();
-#elif __linux__ && __i386__
+#ifdef __linux__ && __i386__
{
unsigned long tsc;

Attachment: log4shib-1.0-tru64.patch.sig
Description: Binary data




Archive powered by MHonArc 2.6.16.

Top of Page