Larry Jones wrote:
Derek R. Price writes:
It wouldn't, but it could break the RCS archive file contract that says that internal commit timestamps will be increasing.
Is there such a contract, given that RCS's ci has a -d option that lets
you set the date to anything you want (and its default is the file's
timestamp, not the current time)? On further consideration, I'm not
sure that that would cause any problems (unless CVS really does depend
on revisions being in date order).
Yes. RCS_getdatebranch() in rcs.c stops when it finds the first date that is too old. From RCS_getdatebranch() in rcs.c:
/* walk the next pointers until you find the end, or the date is too late */
while (p != NULL)
{
vers = p->data;
if (RCS_datecmp (vers->date, date) <= 0) cur_rev = vers->version;
else break;
/* if there is a next version, find the node */
if (vers->next != NULL) p = findnode (rcs->versions, vers->next);
else p = NULL;
}Derek
_______________________________________________ Info-cvs mailing list [EMAIL PROTECTED] http://lists.gnu.org/mailman/listinfo/info-cvs
