[I am not subscribed to debian-arm; please be sure that you retain at least "firstname.lastname@example.org" in your replies.] On Mon, Jan 03, 2005 at 12:38:12AM +0100, Nicolas George wrote: > I would almost sayt it is the opposite of an ABI change. The exact > description of this problem is that the Xlib header fails to describe > properly the ABI of the Xlib. More precisely, it make assumptions about the > memory layout of structures that are left implementation-defined by the C > standard, and are not met by gcc on the ARM processor with the default > options. Okay. I spoke with Jim Gettys on IRC about this. I tend to put a lot of stock in his judgement on issues like this, because he's one of the founders of the X Window System, and has an extremely valuable longitudnal view of its development. He is also, it must be said, a better C programmer than I am. :) Here are his remarks, recast a bit from IRC-speak into something more conventional. GCC on ARM is doing something different from every other C compiler I've seen. It may not deviate from what the C specification allows, but it appears to deviate from common practice. The ARM folks would find code would work with fewer problems if they fixed GCC to behave like other C compilers do. Having an array of structs of a byte each usually forces 16 bit alignment only on other compilers I've seen. I would like the input of the Debian ARM list and the Debian GCC maintenance team on this issue. I'm also curious to know if xterm has always had this problem on Debian ARM, or if it has cropped up only with recent revisions of GCC. Can anyone tell me? For further background, please see the logs of this bug at: http://bugs.debian.org/285396 -- G. Branden Robinson | You don't just decide to break Debian GNU/Linux | Kubrick's code of silence and then email@example.com | get drawn away from it to a http://people.debian.org/~branden/ | discussion about cough medicine.
Description: Digital signature