[Bug 4350] x11-image gives weird error at END of bitbake

bugzilla-daemon at amethyst.openembedded.net bugzilla-daemon at amethyst.openembedded.net
Sun Jun 15 02:54:41 UTC 2008


http://bugs.openembedded.net/show_bug.cgi?id=4350


Mike Westerhof <mwester at mwester.net> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
         AssignedTo|openembedded-issues at lists.o |mwester at mwester.net
                   |penembedded.org             |
     Ever Confirmed|0                           |1




--- Comment #7 from Mike Westerhof <mwester at mwester.net>  2008-06-15 04:54:40 ---
Created an attachment (id=9792)
 --> (http://bugs.openembedded.net/attachment.cgi?id=9792)
Patch to ensure correct reading of an "ar" file

The root cause seems to be that the OE library that parses "ar" files will fail
if the size of an entry in that archive is of odd length.

More specifically, the file format for "ar" archives requires that all headers
in the archive must be aligned on a 16-bit boundary (i.e. even bytes).  In the
case where the size of the data associated with a header is of odd length, the
code would attempt to read one byte too early in the stream.

This problem has gone un-noticed for some time, probably because either the
padding character used was innocuous and caused no harm, or more likely because
the data content was always some even number of bytes.

With Fedora 9 the padding character used for that byte is the back-tick, and it
seems to be common for entries to be of odd-size.

This patch simply ensures that the each header is read from the correctly
aligned byte.

-- 
Configure bugmail: http://bugs.openembedded.net/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.




More information about the Openembedded-issues mailing list