Hi, here is a tidied ebuild which fixes econf for the nasty "mio unable to listen" errors that jabberd reports when ipv6 is compiled-in but not usable. I've added two new debugging USE flags. jabberd-1.4.4-r3 always segfaults when a user tries to connect, although jabberd-1.4.3-r5 works fine. I've tried tweaking jabberd.xml to no avail, and having the jabber user recreated. This is with the USE flags: -ipv6 -mysql -postgres -ssl This bug is relevant to the tracker bug #149358.
Created attachment 98287 [details] jabberd-1.4.4-r3.ebuild
Modified version commited, few comments: * S="${WORKDIR}/${P}" <- useles, I missed it too ;) * mv "${D}"/etc/jabber/jabber{,d}.xml and mv "${D}"/etc/jabber/jabber{,d}.xml.dist <- I know few bash tricks too, but it doesn't help readability * ewarn about separate transports - this is since some time, so no need to ewarn, einfo is enought. Apart from that thanks for help. About segfaulting: compile with debug enabled and run it under gdb. I'll try to reproduce it.
The backtrace: (gdb) run -c /etc/jabber/jabberd.xml Starting program: /usr/bin/jabberd -c /etc/jabber/jabberd.xml 20060928T10:16:31: [notice] (-internal): initializing server Program received signal SIGSEGV, Segmentation fault. 0xa7de2a40 in _pool_cleanup_append (p=0x81cae28, pf=0x8116fc8) at pool.c:203 203 pool.c: No such file or directory. in pool.c (gdb) bt #0 0xa7de2a40 in _pool_cleanup_append (p=0x81cae28, pf=0x8116fc8) at pool.c:203 #1 0xa7de2b26 in _pool_heap (p=0x81cae28, size=1765438053) at pool.c:257 #2 0xa7de2c8e in pmalloc (p=0x81cae28, size=8) at pool.c:312 #3 0xa7de83da in spool_add (s=0x81cae00, str=0x80c63c8 "/var/spool/jabber") at str.c:217 #4 0xa7de849b in spooler (s=0x81cae00) at str.c:246 #5 0xa7f70e21 in xdb_file_full (create=0, p=0x81cace0, spl=0x80c63c8 "/var/spool/jabber", host=0x81caed1 "localhost", file=0x81caec8 "nelchael", ext=0xa7f71e30 "xml", use_subdirs=0) at xdb_file.c:255 #6 0xa7f70f9e in xdb_file_phandler (i=0x80c3228, p=0x81caeb0, arg=0x80c6390) at xdb_file.c:283 #7 0xa7ddea52 in deliver_instance (i=0x80c3228, p=0x81caeb0) at deliver.c:707 #8 0xa7ddde78 in deliver (p=0x81caeb0, i=0x80bfb40) at deliver.c:483 #9 0xa7ddca88 in xdb_deliver (i=0x80bfb40, xc=0x8138290) at xdb.c:113 #10 0xa7ddcd8a in xdb_get (xc=0x80bfde0, owner=0x8115bc0, ns=0xa7c337cf "jabber:iq:auth") at xdb.c:218 #11 0xa7c1be12 in js_user (si=0x80bfd88, id=0x8115ae0, ht=0x8117418) at users.c:179 #12 0xa7c183bb in _js_authreg_register (p=0x81159c8) at authreg.c:123 #13 0xa7c1876e in js_authreg (arg=0x81159c8) at authreg.c:157 #14 0xa7ddc361 in mtq_main (arg=0x8116f08) at mtq.c:153 #15 0xa7d7f3f9 in pth_exit () from /usr/lib/libpth.so.20 #16 0xa7c72b74 in makecontext () from /lib/libc.so.6 #17 0x00000000 in ?? ()
28 Sep 2006; Krzysiek Pawlik <nelchael@gentoo.org> +files/jabberd-1.4.4-genhash.c-rev1253.patch, +files/jabberd-1.4.4-mio-bug-rev1291.patch, +files/jabberd-1.4.4-str.c-rev1103.patch, +files/jabberd-1.4.4-xdb_file.c-rev1107.patch, +files/jabberd-1.4.4-xdb_sql.c-rev1211.patch, jabberd-1.4.4-r3.ebuild: Add various fixes from upstream. I've added various patches from upstream SVN - it fixes not only this bug, but (as you can see from the ammount of them) also few others. I was able to succesfuly register. Can you please test it? Just sync portage (in about 1 hour) and re-emerge jabberd-1.4.4-r3
Yup, those patches fix it :)
(In reply to comment #5) > Yup, those patches fix it :) Great :)