dan@rna.UUCP (01/16/84)
Well to answer my own question, at least partially... Berkeley Mail does hand fiddle the atime of the mail file, either by utime() or by sleep(), open(), read(), as I had guessed. Csh know about this and gets new mail right. USG mail doesn't fiddle with atime, and (at the version I have of) the Bourne shell can't tell the difference between new mail and a partially deleted old mail file. Terrible. It just seems strange that you can't tell if the last program that touched a file did only writes, and that mail has to hand fiddle the atime of a file to get new mail to flag unambiguously. Cheers, Dan Ts'o ...cmcl2!rna!dan