Michael Ströder writes:

/home/michael/src/openldap/OPENLDAP_REL_ENG_2_3/openldap/servers/slapd/.libs/lt-slapd:

Symbol `ldap_int_global_options' has different size in shared object, consider re-linking

What ./configure options? Does make distclean and rebuild help?

On Dienstag, 12. Februar 2008, Michael Ströder wrote:

Hallvard B Furuseth wrote:

...

Michael Ströder writes:

...

/home/michael/src/openldap/OPENLDAP_REL_ENG_2_3/openldap/servers/s lapd/.libs/lt-slapd:

Symbol `ldap_int_global_options' has different size in shared object, consider re-linking

What ./configure options? Does make distclean and rebuild help?

I've attached the build script. It contains 'make distclean'. It fails no matter whether I compile with -O4 or -O0 (gcc version 4.2.1).

Might that be caused by the missing "make depend" step?

Michael Ströder writes:

...

Might that be caused by the missing "make depend" step?

If so make distclean (which he did use) would be broken. There are hosts which don't provide a way to do "make depend".

Any hint? Maybe something related to libtool?

Remove your CPPFLAGS and LDFLAGS. It's the -Wl,-R/usr/lib. That worked in OpenLDAP 2.2 but has been broken since 2.3.4 (the first 2.3 release). Links the libraries in /usr/lib instead of the ones you just built. Don't know if it's the Makefile or libtool.

I seem to remeber warnings in the gcc doc that you should not do -I/usr/include either, it overrides gcc's special version of some files. Or did, maybe they've worked around it since I saw that.

Passes all tests on CentOS4.6 and CentOS5.1 (i686 and virtual machines)

Ing. Luca Scamoni Responsabile Ricerca e Sviluppo

SysNet s.r.l. via Dossi, 8 - 27100 Pavia - ITALIA http://www.sys-net.it ----------------------------------- Office: +39 0382 573859 (137) Mobile: +39 347 1014425 Email: luca.scamoni@sys-net.it -----------------------------------

Are there some non-Linux folks here who can test it? :-)

On Feb 12, 2008, at 9:43 AM, Hallvard B Furuseth wrote:

Are there some non-Linux folks here who can test it? :-)

I was going to say "Sure!" since I have a NetBSD host handy that I use for LDAP experiments among other things, but it seems that I can't make test002-populate work, for any release. Error occurs when authenticating the root DN -

do_bind: version=3 dn="cn=Manager,dc=example,dc=com" method=128 ==> bdb_bind: dn: cn=Manager,dc=example,dc=com bdb_dn2entry("cn=manager,dc=example,dc=com") => bdb_dn2id("dc=example,dc=com") <= bdb_dn2id: get failed: DB_NOTFOUND: No matching key/data pair found (-30990)

I must be missing something - config looks fine, the rootdn is specified as above. Database is valid, though empty as I guess expected.

Anyway, I need to move on to something else right now, and will probably have to wait until tomorrow before I get back to this.

Donn Cave, donn@u.washington.edu

On NetBSD 3.1, test008-concurrency dies trying to allocate 8388608 bytes -

search_candidates: base="dc=example,dc=com" (0x00000001) scope=2 ch_malloc of 8388608 bytes failed assertion "0" failed: file "ch_malloc.c", line 57, function "ch_malloc"

#0 0xbd9980fb in kill () from /usr/lib/libc.so.12 #1 0xbda1917f in abort () from /usr/lib/libc.so.12 #2 0xbd9e309a in __assert13 () from /usr/lib/libc.so.12 #3 0x08077bce in ch_malloc (size=8388608) at ch_malloc.c:57 #4 0x080c18e6 in search_stack (op=0x8435000) at search.c:1008 #5 0x080c1c4d in search_candidates (op=0x8435000, rs=0xb8fffd58, e=0xb8f3fa58, locker=70, ids=0xb8f7fa88, scopes=0xb8f3fa88) at search.c:1094 #6 0x080c13d3 in bdb_search (op=0x8435000, rs=0xb8fffd58) at search.c:573 #7 0x0806390f in fe_op_search (op=0x8435000, rs=0xb8fffd58) at search.c:368 #8 0x08063195 in do_search (op=0x8435000, rs=0xb8fffd58) at search.c:217 #9 0x080617a8 in connection_operation (ctx=0xb8fffe28, arg_v=0x8435000) at connection.c:1084 #10 0x080619ae in connection_read_thread (ctx=0xb8fffe28, argv=0xa) at connection.c:1211 #11 0x0814f21c in ldap_int_thread_pool_wrapper (xpool=0x830a380) at tpool.c:625 #12 0xbda4917d in pthread_create () from /usr/lib/libpthread.so.0

search.c:1008 ret = ch_malloc( bdb->bi_search_stack_depth * BDB_IDL_UM_SIZE * sizeof( ID ) );

source acquired via cvs ca 10 AM PST - same failure with rel_eng_2_3 and rel_eng_2_4.

Also, same failure with test039-glue-ldap-concurrency - could be a pattern here.

Donn Cave, donn@u.washington.edu

On Feb 14, 2008, at 2:32 AM, Howard Chu wrote:

Donn Cave wrote:

...

On NetBSD 3.1, test008-concurrency dies trying to allocate 8388608 bytes -

 search_candidates: base="dc=example,dc=com" (0x00000001) scope=2
 ch_malloc of 8388608 bytes failed
 assertion "0" failed: file "ch_malloc.c", line 57, function

"ch_malloc"

...

Also, same failure with test039-glue-ldap-concurrency - could be a pattern here.

Yes, both tests would have similar behavior. It seems to me that this isn't anything unusual though. Does NetBSD default to a particular ulimit on datasize?

$ ulimit -a core file size (blocks, -c) unlimited data seg size (kbytes, -d) 131072 file size (blocks, -f) unlimited max locked memory (kbytes, -l) 1279544 max memory size (kbytes, -m) 3838632 open files (-n) 64 pipe size (512 bytes, -p) 1 stack size (kbytes, -s) 2048 cpu time (seconds, -t) unlimited max user processes (-u) 160 virtual memory (kbytes, -v) 133120

The tests complete OK on 2_3 and 2_4, with data seg up to 1048576, max memory size unlimited, stack size 32768, virtual memory 1081344.

Donn Cave, donn@u.washington.edu

On Tue, 12 Feb 2008, Hallvard B Furuseth wrote:

Are there some non-Linux folks here who can test it? :-)

Tested both RE23 and RE24 on FreeBSD 6.2.

Buchan Milne wrote:

On Tuesday 12 February 2008 19:43:28 Hallvard B Furuseth wrote:

...

Are there some non-Linux folks here who can test it? :-)

I consider myself a Linux person, but I do have a Solaris 10 / sparc64 box under my desk, if no-one else is testing Solaris 10 / sparc, I can go through the pain of getting it to compile OpenLDAP (or seeing if I can convince my solaris sparc build of rpm to build my usual SRPM).

I still have temporary access to a Solaris 10 T5120. I'll give it a whirl. (Each core is pretty slow, but you can get pretty good build times with "make -j32")

Buchan Milne wrote:

On Tuesday 12 February 2008 19:43:28 Hallvard B Furuseth wrote:

...

Are there some non-Linux folks here who can test it? :-)

I consider myself a Linux person, but I do have a Solaris 10 / sparc64 box under my desk, if no-one else is testing Solaris 10 / sparc, I can go through the pain of getting it to compile OpenLDAP (or seeing if I can convince my solaris sparc build of rpm to build my usual SRPM).

Of course I'm testing RE24 at the moment. Aside from ITS#5379 it built fine on Solaris 10 sparc64 with BDB 4.6.21. Will try RE23 when this round of tests completes.