http://bugzilla.novell.com/show_bug.cgi?id=548053
User wstephenson@novell.com added comment
http://bugzilla.novell.com/show_bug.cgi?id=548053#c2
Will Stephenson changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |wstephenson@novell.com
AssignedTo|wstephenson@novell.com |dmueller@novell.com
--- Comment #2 from Will Stephenson 2009-10-22 06:30:52 MDT ---
So I find the selftests pass on x86_64 and fail on i586.
x86_64 is being built with -D_LARGEFILE64_SOURCE
i586 is being built with D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64
I wrote a simple testcase to extract the kleopatra selftests (attached), and
when including the gpgme_engine_check_version() call from libgpgme++, triggered
an #error in gpgme.h for not building with -D_FILE_OFFSET_BITS=64 like gpgme
had been built with.
When this is set, the testcase fails.
When this is not set, the testcase passes.
I don't know enough about largfile support to say if _FILE_OFFSET_BITS=64
should be used on i586.
So the following alternate hypotheses could be true:
1) my testcase passes without this set because gpgme with _F_O_B=64 behaves in
an undefined way when called from code without it
2) _F_O_B=64 is wrong on i586 and makes the kleopatra selftest fail there
3) the tests are correct and there's another error inside gpgme that makes it
not support gpgconf despite having been configured with it.
Dirk, can you shed some light here?
--
Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.