Bastian Blank <[EMAIL PROTECTED]> writes: > On Thu, Feb 21, 2008 at 11:09:54AM -0500, James Antill wrote: >> It looks like my ENOMEM checks are failing. IIRC s390 is 31 bit? and >> those are designed for 32 bit ... so no big surprise I guess :(. > > No, s390 have 32 bit pointer but only a 31 bit address space.
Right... >> Probably the best instant fix is to just surround parts of >> T/ctst_99_64bit.c with #ifndef __s390__ or whatever. > > Can you please explain what this does? So it's basically two parts to that test case: 1) Makes sure you can't allocate more data than a Ustr can hold (Ie. just less than SIZE_MAX). 2. Bits using big_tst() ... this uses ftruncate() + mmap() to "fake" a Ustr that is just over (SIZE_MAX / 2) and then tests that adding it to itself doesn't work etc. ...I assume it's the big_tst() parts that are failing. -- James Antill -- [EMAIL PROTECTED] http://www.and.org/and-httpd/ -- $2,000 security guarantee http://www.and.org/ustr/ http://www.and.org/vstr/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]