[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: Please test OPENLDAP_REL_ENG_2_3

Hallvard B Furuseth wrote:

tests/data/slapd-valsort.conf is missing, for test042.

tests/data/slapd-aci.conf is missing.
If it is copied from HEAD, slapd dumps core in test041:

#0  0x080a807b in parse_acl (be=0x82b75a8,
        fname=0x828b008 "./testrun/slapd.1.conf",
        lineno=54, argc=6, argv=0x82b48f0, pos=-1) at aclparse.c:1558
1558                                if( !is_at_syntax( b->a_aci_at->ad_type,

...because b->a_aci.at is NULL.

(It is configured with --enable-most things, including --enable-aci.)

(gdb) bt
#0  0x080a807b in parse_acl (be=0x82b75a8, fname=0x828b008 "./testrun/slapd.1.conf", lineno=54, argc=6, argv=0x82b48f0, pos=-1)
   at aclparse.c:1558
#1  0x080652e4 in config_generic (c=0x82b3778) at bconfig.c:1108
#2  0x0806fb18 in config_set_vals (Conf=0x8258db0, c=0x82b3778) at config.c:261
#3  0x0806ffda in config_add_vals (Conf=0x8258db0, c=0x82b3778) at config.c:329
#4  0x08070f2c in read_config_file (fname=0x828b008 "./testrun/slapd.1.conf", depth=0, cf=0x0, cft=0x8258d60) at config.c:660
#5  0x0806aef9 in read_config (fname=0x828b008 "./testrun/slapd.1.conf", dir=0x0) at bconfig.c:2738
#6  0x080df32f in slap_tool_init (progname=0x82065e0 "slapadd", tool=1, argc=6, argv=0xbffff334) at slapcommon.c:402
#7  0x080dd83b in slapadd (argc=6, argv=0xbffff334) at slapadd.c:70
#8  0x080636cd in main (argc=6, argv=0xbffff334) at main.c:526



False alarm; the SIGSEGV should now be fixed in HEAD (servers/slapd/init.c 1.88 -> 1.89); please merge into re23.
However, the test fails because in re23 #undef SLAP_ACL_HONOR_DISCLOSE; this flag causes "disclose" access to the search base to be tested (and required), and represents a change in ACL checking behavior. Either we decide to make this change with 2.3.7, or test041 should be disabled (or reworked to follow the old convention; I note, however, that by design ACIs don't quite meet the requirements for disclose checking as currently performed by access_allowed_mask(); see ITS#3963). In any case, better disable the test; I need it in HEAD for regressions on the work I'm making on ACIs, but since it's a low priority activity I don't think I can have all of the issues fixed in time, and I don't like the idea of carrying on two versions of the test.

p.



   SysNet - via Dossi,8 27100 Pavia Tel: +390382573859 Fax: +390382476497