Guard against crash during getmetadata on shared folder

This is what the crash looks like. It's reproducible on that folder
only, via cyradm (as cyrus-admin or a user), by doing a

'getmd "shared/Folder Example/TODOs Fabian@example.com"'

[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib64/glibc-hwcaps/power9/libthread_db-1.0.so".
Core was generated by `imapd '.
Program terminated with signal SIGSEGV, Segmentation fault.
1979 if (!userid[0] || !strcmp(userid, state->userid)) {
Missing separate debuginfos, use: yum debuginfo-install cyrus-sasl-lib-2.1.27-6.el8_5.ppc64le cyrus-sasl-plain-2.1.27-6.el8_5.ppc64le glibc-2.28-211.el8.ppc64le jansson-2.14-1.el8.ppc64le keyutils-libs-1.5.10-9.el8.ppc64le krb5-libs-1.18.2-22.el8_7.ppc64le libcom_err-1.45.6-5.el8.ppc64le libdb-5.3.28-42.el8_4.ppc64le libgcc-8.5.0-16.el8_7.ppc64le libical-3.0.3-3.el8.ppc64le libicu-60.3-2.el8_1.ppc64le libpq-13.5-1.el8.ppc64le libselinux-2.9-6.el8.ppc64le libstdc++-8.5.0-16.el8_7.ppc64le libuuid-2.32.1-39.el8_7.ppc64le libxcrypt-4.1.1-6.el8.ppc64le libxml2-2.9.7-15.el8_7.1.ppc64le libzstd-1.4.4-1.el8.ppc64le mariadb-connector-c-3.1.11-2.el8_3.ppc64le openldap-2.4.46-18.el8.ppc64le openssl-libs-1.1.1k-7.el8_6.ppc64le pcre-8.42-6.el8.ppc64le pcre2-10.32-3.el8_6.ppc64le sqlite-libs-3.26.0-17.el8_7.ppc64le xapian-core-libs-1.4.18-3.el8.ppc64le xz-libs-5.2.4-4.el8_6.ppc64le zlib-1.2.11-21.el8_7.ppc64le
(gdb) bt
0 rw_cb (mailbox=0x1305cbba0 "example.com!shared.Folder Example.TODOs Fabian", uid=0, entry=0x1305c4c05 "/vendor/kolab/folder-type", userid=0x0, value=0x7fffcd104ca0, mdata=0x7fffcd104c90, rock=0x1305d0e50) at imap/annotate.c:1979
1 0x00007fff83308528 in find_cb (rock=0x7fffcd106ed0, key=<optimized out>, keylen=<optimized out>, data=0x7fff7f15aa69 "", datalen=49) at imap/annotate.c:1137
2 0x00007fff82e1f4b4 in myforeach (db=0x13057b0b0, prefix=0x7fffcd105e10 "418d184e-f18f-41c5-8ba6-05e221ea7c60", prefixlen=62, goodp=0x7fff83306e10 <find_p>, cb=0x7fff83308220 <find_cb>, rock=0x7fffcd106ed0, tidptr=<optimized out>)

at lib/cyrusdb_twoskip.c:1651

3 0x00007fff82e0de98 in cyrusdb_foreach (db=<optimized out>, prefix=<optimized out>, prefixlen=<optimized out>, p=<optimized out>, cb=<optimized out>, rock=<optimized out>, tid=<optimized out>) at lib/cyrusdb.c:281
4 0x00007fff83308754 in _findall (data=data@entry=0x7fffcd106ea8, rock=rock@entry=0x7fffcd106ed0) at imap/annotate.c:1171
5 0x00007fff8330c920 in _findall (rock=0x7fffcd106ed0, data=0x7fffcd106ea8) at imap/annotate.c:1228
6 annotatemore_findall_full (pattern=pattern@entry=0x7fff83422e98 "", mailbox=0x130612fd0, mboxname=<optimized out>, mboxname@entry=0x0, uid=<optimized out>, entry=<optimized out>, since_modseq=since_modseq@entry=0,

proc=proc@entry=0x7fff83309290 <rw_cb>, rock=rock@entry=0x1305d0e50, flags=<optimized out>, flags@entry=0) at imap/annotate.c:1228

7 0x00007fff8330cc98 in annotation_get_fromdb (state=0x1305d0e50, entry=0x1305c5360) at imap/annotate.c:1993
8 0x00007fff83307b0c in _annotate_fetch_entries (state=state@entry=0x1305d0e50, proxy_check=proxy_check@entry=1) at imap/annotate.c:2648
9 0x00007fff8330e278 in annotate_state_fetch (state=0x1305d0e50, entries=0x7fffcd107220, attribs=0x7fffcd107230, callback=<optimized out>, rock=<optimized out>) at imap/annotate.c:2788
10 0x000000010e9baa48 in annot_fetch_cb (astate=<optimized out>, astate@entry=0x1305d0e50, rock=rock@entry=0x7fffcd107260) at imap/imapd.c:10092
11 0x000000010e9d76f8 in apply_mailbox_array (proc=0x10e9baa20 <annot_fetch_cb>, rock=0x7fffcd107260, mboxes=0x7fffcd1072c8, state=<optimized out>) at imap/imapd.c:10199
12 cmd_getmetadata (tag=tag@entry=0x130581fe0 "PROXY18") at imap/imapd.c:10571
13 0x000000010e9e96ec in cmdloop () at imap/imapd.c:1591
14 0x000000010e9eff0c in service_main (argc=<optimized out>, argv=<optimized out>, envp=<optimized out>) at imap/imapd.c:1005
15 0x000000010e9b1bc8 in main (argc=<optimized out>, argv=<optimized out>, envp=0x7fffcd10b748) at master/service.c:647

The crash is triggered by no userid being found for the folder (which is
subsequently dereferenced. I assume there is some data corruption
somewhere, this patch doesn't fix any of it, it simply prevents the
crash logging an error instead.