mb@ttidca.TTI.COM (Michael Bloom) (03/08/88)
In the sccs source, in get() in get.c, the p-file gets updated prior to generating the new g-file. Is this ordering absulutely necessary? i.e would something break if I were to reverse this order? For a given SID, i'd like to be certain that the date symbolically stored in the p-file is never less than than the stat-able mod time of the g-file until a change has been made to the g-file. This would allow automatic detection of modules that need to be checked back in. Perhaps better than reversing the order might be to simply use utime() to set the g-file's mod date to "Timenow". This is particularly appealing in that it's only a one line addition. What might break with this approach? Finally, is some obvious solution already present that i've been blindly overlooking? - michael
davidsen@steinmetz.steinmetz.UUCP (William E. Davidsen Jr) (03/11/88)
If I were doing sccs over, I'd make two changes... when I added a
(delta) file I'd reset the modified time of the s file to the time of
the file added. When I extracted (get) a file I'd reset the time of
that, too.
sccs can really be a PAIN when used with make, unless you keep the
extracted files handy. Of course, having make pull writably copies of
all files from SCCS doesn't help.
--
bill davidsen (wedu@ge-crd.arpa)
{uunet | philabs | seismo}!steinmetz!crdos1!davidsen
"Stupidity, like virtue, is its own reward" -me