On Fri, Dec 15, 2006 at 02:02:58AM +0100, Julien Cristau wrote: >On Fri, Dec 15, 2006 at 00:59:38 +0100, Samuel Thibault wrote: > >> Hi, >> >> Looks like the culprit line is update.c:726. >> >> unsigned char checksum[16]; >> >> An unsigned char[16] won't necessarily be aligned by gcc. However, it >> is used as a checksum buffer, on which md5_read_ctx() performs 32bit >> operations. Hence the SIGBUS on sparc. I guess the best correction >> would be to just express the correct type: >> >> md5_uint32 checksum[4]; >> >I can confirm that the attached patch fixes this problem for me.
Excellent, thanks for the help in tracking this down, Julien and Samuel! I'll add this and upload this weekend, and clearly upstream need to know too. -- Steve McIntyre, Cambridge, UK. [EMAIL PROTECTED] You raise the blade, you make the change... You re-arrange me 'til I'm sane...
signature.asc
Description: Digital signature
