Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.1 6/24/83; site wjvax.UUCP Path: utzoo!watmath!clyde!bonnie!akgua!sdcsvax!sdcrdcf!hplabs!pesnta!wjvax!ron From: ron@wjvax.UUCP (Ron Christian) Newsgroups: net.unix-wizards Subject: Re: Multiple file versions -- FLAME off!! Message-ID: <210@wjvax.UUCP> Date: Fri, 28-Sep-84 17:26:32 EDT Article-I.D.: wjvax.210 Posted: Fri Sep 28 17:26:32 1984 Date-Received: Mon, 1-Oct-84 04:13:18 EDT References: <2163@mit-hermes.ARPA> Organization: Watkins Johnson, San Jose, Calif. Lines: 25 () Well, I'm not a unix wizard exactly, but for a user's point of view: I implemented a 'miltiple files' gimmic on 4.2 BSD with a couple aliases. Not hard to do, you can append sequential numbers (or even the date) to new files one each during the edit/compile/run cycle. Or push them up a queue (foo.old foo.reallyold foo.reallyreallyold, etc) automagically. Had an alias 'backup' that would do this, then nested it into 'cc' alias. (You could put it in the makefile, too.) Anyway, I soon abandoned the practice after I discovered that when something DID barf, often as not I knew right away where I bungled, or, the error was so deeply embedded that I'd have to give up several nifty (working) features I had added to go far enough back to not have the bug. I still have the 'backup' command, but it only creates ONE backup, and I have to initiate it. (To guard against serious fumbles while in 'vi'.) New hacks to code are always marked (/*TEST TEST TEST*/ or some such) and I can almost always remove a couple generations of changes successfully. What more do you need? -- "Trivia is important." Ron Christian (syntax bug) Watkins-Johnson Co. San Jose, Calif. (...ios!wjvax!ron)