[Pkg-db-devel] Bug#521860: Bug#521860: Processed: Re: Bug#521860: libpam-modules: libdb upgrade silently breaks pam_userdb.so

[Clint Adams]

severity 521860 grave
quit

On Fri, Apr 03, 2009 at 06:29:10AM +0000, Clint Adams wrote:
> -		if (PGNO(meta) == PGNO_BASE_MD && !F_ISSET(dbp, DB_AM_RECOVER))
> +		if (PGNO(meta) == PGNO_BASE_MD && meta->dbmeta.last_pgno > 0 && !F_ISSET(dbp, DB_AM_RECOVER))

> +		     hcp->hdr->dbmeta.last_pgno > 0 &&

So for me, on a v7 hash created with db3_load 3.2.9+dfsg-0.1, db4.7_dump will corrupt the database file
unless either -r or -R are specified.  With a v8 btree, a db4.7_dump -p will output correctly but a
db4.7_dump -d a will not.  In either case, dumping the v8 btree does not appear to lead to data loss.

a db4.7_verify on the v7 hash gives an error about an impossible max_buckets setting in the metadata
page.  It does this because it in the partially-quoted patch above, it clobbers the last_pgno setting
and compares 1, for instance, against 0 instead of 2.

I don't know if this fix is comprehensive and I don't know what it breaks.  I will attempt to escalate
this to Oracle.