Bugzilla – Bug 488
sasl build fails on rh8.0 when running cvs-build
Last modified: 2005-12-06 17:00:20
You need to log in before you can comment on or make changes to this bug.
When building sasl using cvs-build (from a checkout of HEAD or 2.2.2 branches), the cyrus-sasl build fails. I've only seen this behavior on red hat 8.0, which has the gcc version 3.2 compiler: gcc version 3.2 20020903 (Red Hat Linux 8.0 3.2-7) Basically, the configure step of sasl fails to find a number of headers (such as pwd.h) checking for getopt.h... no checking for unistd.h... no checking for crypt.h... no checking for pwd.h... no checking for shadow.h... no checking for paths.h... no These headers are all in /usr/include. Since these headers aren't found, config.h doesn't define the associated defs (HAVE_PWD_H) which causes compile errors. The configure step checks for the headers by compiling a simple program which includes the header. The following error shows the problem. The program compiles, but gives an error because /usr/local/include is a non-system directory (??), which I assume has been hardcoded into the pre-processor when it was built. configure:5040: gcc -E -I/usr/local/include conftest.c >/dev/null 2>conftest.out cpp0: warning: changing search order for system directory "/usr/local/include" cpp0: warning: as it has already been specified as a non-system directory configure: failed program was: #line 5035 "configure" #include "confdefs.h" #include <pwd.h> The configure script attempts to verify that no output was printed to stderr, which isn't the case because of the warnings, so the check for the header fails overall. The only solution I can see for this would be to remove the -I/usr/local/include addition to CPPFLAGS in sasl's configure.in. Not sure this kind of fix won't break on other platforms.
*** Bug 490 has been marked as a duplicate of this bug. ***
It has been suggested that upgrading from gcc 3.2 to gcc 3.2.1 cures this.
I just verified that upgrading to gcc 3.2.1 fixes this problem.
*** Bug 672 has been marked as a duplicate of this bug. ***
*** Bug 679 has been marked as a duplicate of this bug. ***
There's also a one-line patch to gcc 3.2 that will fix this if you'd rather not upgrade to 3.2.1. We have a patched gcc 3.2 RPM from SuSE that includes this patch; I don't know whether it's generally available. If you're interested in details on the gcc 3.2 patch, contact me, kst@sdsc.edu; if there's any interest I'll post more information here.
*** Bug 448 has been marked as a duplicate of this bug. ***
I've been using gcc-3.2.2-1 from RedHat's rawhide distribution without trouble for the past few weeks.
*** Bug 840 has been marked as a duplicate of this bug. ***
*** Bug 1109 has been marked as a duplicate of this bug. ***
Just to be completely clear on this, the easy solution to this bug is to upgrade your GCC to a version that is not gcc-3.2. For example, gcc-3.2.2 works.
*** Bug 1125 has been marked as a duplicate of this bug. ***
*** Bug 1314 has been marked as a duplicate of this bug. ***
Keith had the following post at globus bugzilla. "There's also a one-line patch to gcc 3.2 that will fix this if you'drather not upgrade to 3.2.1. We have a patched gcc 3.2 RPM from SuSEthat includes this patch; I don't know whether it's generally available.If you're interested in details on the gcc 3.2 patch, contact me,kst@sdsc.edu; if there's any interest I'll post more information here." I am using RH 8.0 and have gcc 3.2 rpm installed. RH 9.0 has gcc 3.2.2, but when I compiled it on RH 8.0, it failed. Do anyone have an easy solution to fix my gcc-3.2-7 rpm? Does anybody have a source rpm for RH 8.0? Thank you very much. Dantong
I'm working on a workaround for this problem. It's a small shell script that can be installed as "gcc"; it invokes the real gcc and filters out the warning message (while maintaining gcc's other output on stdout and stderr, as well as gcc's exit status). I'll post details here when it's done.
What about this patch: ftp://ftp.nordugrid.org/patches/globus/globus_cyrus-sasl_nousrlocal.patch We've been using it in NorduGrid for quite some time now. Anders
My workaround script is at <http://www.sdsc.edu/~kst/gcc-wrapper/>.
*** Bug 1658 has been marked as a duplicate of this bug. ***
*** Bug 1701 has been marked as a duplicate of this bug. ***
*** Bug 1818 has been marked as a duplicate of this bug. ***
*** Bug 1869 has been marked as a duplicate of this bug. ***
*** Bug 3768 has been marked as a duplicate of this bug. ***