Fwd: Re: [oe] [RFC] move to opkg / md5 error on big-endian machine

Holger Freyther zecke at openmoko.org
Sat Mar 22 01:09:37 CET 2008


On Thursday 20 March 2008 13:20:01 Thomas Wood wrote:
> On Wed, 2008-03-19 at 19:25 +0100, Marcin Juszkiewicz wrote:

> This issue came as a result of updating the md5 sum check, as the old
> one gave difficult compiler warnings.

Hmm, to be future proof you should use the flags exported by gcc [1] instead 
of doing the autoconf test.

Reasoning: E.g. with universal binaries on OSX you will have one autoconf run 
but use gcc -arch i386 ppc to compile for two different architectures in one 
run. In one of the two compiler runs the endian is wrong.


z.

[1] 
http://developer.apple.com/documentation/MacOSX/Conceptual/universal_binary/universal_binary_tips/chapter_5_section_35.html#//apple_ref/doc/uid/TP40002217-CH239-BCICJCDG




More information about the opkg-devel mailing list