http://dovecot.org/releases/dovecot-1.0.rc32.tar.gz http://dovecot.org/releases/dovecot-1.0.rc32.tar.gz.sig
Pretty late for changes if v1.0 is supposed to come out tomorrow, but I can't really leave these LDAP bugs unfixed. They shouldn't anyway break anything, so here's one more day for you people to find out about any bugs.
One thing that I'm a bit concerned about is the addition of #define LDAP_DEPRECATED. I know it fixes crashes with OpenLDAP 2.3 + 64bit systems, but I hope it doesn't break compiling with some old versions or non-OpenLDAP libraries..
- LDAP, auth_bind=no: Don't crash if doing non-plaintext ldap passdb
lookup for unknown user. This also broke deliver when userdb static
was used.
- LDAP, auth_bind=yes and userdb ldap: We didn't wait until bind was
finished before sending the userdb request, which could have caused
problems.
- LDAP: Don't break when compiling with OpenLDAP v2.3 library
- Convert plugin: Don't create "maildirfolder" file to Maildir root.Well, it builds OK for me on Solaris 8 32-bit SPARC with OpenLDAP 2.3, with less complaints from gcc (3.3.2), just:-
mail-index.c: In function mail_index_parse_extensions': mail-index.c:343: warning: comparison between signed and unsigned mail-index.c: In function mail_index_map_clone':
mail-index.c:1294: warning: comparison between signed and unsigned
idxview.c: In function dump_hdr': idxview.c:89: warning: comparison between signed and unsigned idxview.c: In function dump_cache_hdr':
idxview.c:179: warning: comparison between signed and unsigned
idxview.c:194: warning: comparison between signed and unsigned
idxview.c: In function dump_cache': idxview.c:243: warning: comparison between signed and unsigned idxview.c: In function dump_record':
idxview.c:345: warning: comparison between signed and unsigned
whereas before I had:-
db-ldap.c: In function db_ldap_search': db-ldap.c:201: warning: implicit declaration of function ldap_search'
db-ldap.c: In function db_ldap_bind_callback': db-ldap.c:425: warning: implicit declaration of function ldap_result2error'
db-ldap.c: In function db_ldap_bind': db-ldap.c:443: warning: implicit declaration of function ldap_bind'
db-ldap.c: In function db_ldap_connect': db-ldap.c:494: warning: implicit declaration of function ldap_init'
db-ldap.c:494: warning: assignment makes pointer from integer without a cast
db-ldap.c: In function ldap_conn_close': db-ldap.c:599: warning: implicit declaration of function ldap_unbind'
passdb-ldap.c: In function ldap_query_save_attr': passdb-ldap.c:84: warning: implicit declaration of function ldap_get_values'
passdb-ldap.c:84: warning: assignment makes pointer from integer without
a cast
passdb-ldap.c:103: warning: implicit declaration of function
ldap_value_free' passdb-ldap.c: In function handle_request_get_entry':
passdb-ldap.c:155: warning: implicit declaration of function
ldap_result2error' passdb-ldap.c: In function authbind_start':
passdb-ldap.c:268: warning: implicit declaration of function ldap_bind' userdb-ldap.c: In function ldap_query_get_result':
userdb-ldap.c:84: warning: implicit declaration of function
ldap_get_values' userdb-ldap.c:84: warning: assignment makes pointer from integer without a cast userdb-ldap.c:104: warning: implicit declaration of function ldap_value_free'
userdb-ldap.c: In function handle_request': userdb-ldap.c:146: warning: implicit declaration of function ldap_result2error'
as well.
Hopefully I'll manage to try some direct LDAP auth before tomorrow (we're using pam_ldap instead at the moment). I'll let you know if there are any issues.
Best Wishes, Chris
Timo Sirainen wrote:
One thing that I'm a bit concerned about is the addition of #define LDAP_DEPRECATED. I know it fixes crashes with OpenLDAP 2.3 + 64bit systems, but I hope it doesn't break compiling with some old versions or non-OpenLDAP libraries..
-- --+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+- Christopher Wakelin, c.d.wakelin@reading.ac.uk IT Services Centre, The University of Reading, Tel: +44 (0)118 378 8439 Whiteknights, Reading, RG6 2AF, UK Fax: +44 (0)118 975 3094
On Thu, 2007-04-12 at 18:58 +0100, Chris Wakelin wrote:
Well, it builds OK for me on Solaris 8 32-bit SPARC with OpenLDAP 2.3, with less complaints from gcc (3.3.2), just:-
mail-index.c: In function
mail_index_parse_extensions': mail-index.c:343: warning: comparison between signed and unsigned mail-index.c: In functionmail_index_map_clone': mail-index.c:1294: warning: comparison between signed and unsigned idxview.c: In functiondump_hdr': idxview.c:89: warning: comparison between signed and unsigned idxview.c: In functiondump_cache_hdr': idxview.c:179: warning: comparison between signed and unsigned idxview.c:194: warning: comparison between signed and unsigned idxview.c: In functiondump_cache': idxview.c:243: warning: comparison between signed and unsigned idxview.c: In functiondump_record': idxview.c:345: warning: comparison between signed and unsigned
These should all go away with a newer gcc.
whereas before I had:- db-ldap.c: In function
db_ldap_search': db-ldap.c:201: warning: implicit declaration of functionldap_search'
The LDAP_DEPRECATED define should have fixed these. Maybe your ldap.h doesn't have the deprecated functions at all for some reason? Or did I somehow mess up applying the deprecated-patch. :)
I'd anyway try fixing these warnings before trying to run the code. I guess they should work with 32bit systems without crashing, but I wouldn't bet on it.
The LDAP_DEPRECATED define should have fixed these. Maybe your ldap.h doesn't have the deprecated functions at all for some reason? Or did I somehow mess up applying the deprecated-patch. :)
rc32 builds cleanly for me (only a few minor warnings) and seems to be running well.
Thanks!
-Ben
Remember even if you don't get everything right by tomorrow there's always version 1.01. There will always be bugs and new features to add.
Marc Perkel schrieb:
Remember even if you don't get everything right by tomorrow there's always version 1.01. There will always be bugs and new features to add.
I do hope Timo will leave bugs and features to the HEAD branch and save the fixes for 1.0.X :-) [1]
Anyways, to Timo - good luck with the release - if time doesn't suffice, call it "Friday 23:59 UTC" which buys you three extra hours in Eastern European Daylight Savings Time. =:-)
[1] Ok, call me German for counting peas... I don't mind as that's a factual description. 8-)
On Thursday 12 April 2007 21:18, Marc Perkel wrote:
Remember even if you don't get everything right by tomorrow there's always version 1.01. There will always be bugs and new features to add.
1.01 is a bad version designation. Version strings are no floating-point numbers!
-- Magnus Holmgren holmgren@lysator.liu.se (No Cc of list mail needed, thanks)
participants (6)
-
Ben Beuchler
-
Chris Wakelin
-
Magnus Holmgren
-
Marc Perkel
-
Matthias Andree
-
Timo Sirainen