Quoting John Darrington <[EMAIL PROTECTED]>: > Does anyone object if we change the line > > typedef int trns_proc_func (void *, struct ccase *, int); > > to read > > typedef int trns_proc_func (void *, struct ccase *, unsigned long); > > ?? > > It makes it compatible with the return type of casereader_cnum, and > avoids the possibility of overflow on very large casefiles.
That's fine with me. I've been mulling the possibility of creating a typedef for case numbers, so that we could easily change them to unsigned long long later--I suspect that >4G cases is not impossible, although uncommon. BTW, my responsiveness will be poor for a few days because of a hard disk failure last Friday. I don't expect I've lost any actual data, because I back up obsessively, but it'll take IBM a few days to ship me a replacement hard drive for my ThinkPad, and then reinstalling can take a little while too. _______________________________________________ pspp-dev mailing list [email protected] http://lists.gnu.org/mailman/listinfo/pspp-dev
