home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.lang.c
- Path: sparky!uunet!spool.mu.edu!agate!doc.ic.ac.uk!warwick!dcs.warwick.ac.uk!rince
- From: rince@dcs.warwick.ac.uk (James Bonfield)
- Subject: Re: sccs eats good code -
- Message-ID: <1993Jan4.113418.27377@dcs.warwick.ac.uk>
- Sender: news@dcs.warwick.ac.uk (Network News)
- Nntp-Posting-Host: stone
- Organization: Department of Computer Science, Warwick University, England
- References: <1992Dec21.4396.29992@dosgate> <1992Dec22.090329.11162@dcs.warwick.ac.uk> <1992Dec26.062743.7520@robohack.UUCP>
- Date: Mon, 4 Jan 1993 11:34:18 GMT
- Lines: 47
-
- In <1992Dec26.062743.7520@robohack.UUCP> woods@robohack.UUCP (Greg A. Woods) writes:
-
- >> Do you notice anything blindingly wrong with this? Sccs has the intelligent
- >> idea of giving control characters special meaning within the history file.
- >> And it doesn't bother to escape any of the ones you happen to have typed in.
- >> This,naturally, means that if you enter a control char into your source then
- >> upon the next 'get' or 'edit' sccs will kindly remove the only existing copy
- >> you've got and *then* complain that it doesn't understand the history file.
- >> For this reason using sccs actually slowed down and hampered my development
- >> by a huge factor. Surely this isn't the scheme of such things? On the
- >> contrary the next time I was asked to make use of such a system I put RCS
- >> to the test
-
- >
- > Whew! What kind of code were you writing anyway?!?!? I've only twice
- > encountered a conflict with SCCS's keyword expansion control codes --
- > once in the same situation as the original poster noted, and once when
- > I wished to include an example SCCS keyword string in a source module.
- >
- > That's twice out of thousands of modules. I know of large projects
- > with millions of lines of code, in tens of thousands of modules, all
- > managed with the use of SCCS, and no conflicts that couldn't be easily
- > dealt with.
- >
- > RCS will do the same thing, though it has far fewer combinations of
- > strings with which it might find conflict.
-
- The insertion of a control character (^A I believe) was infact accidental. The
- point I'm making is that sccs didn't just replace it with something else (as
- in the previous complains of %whatevers), but actually mangled the history
- file and deleted the only correct copy of the source code.
-
- This bug does not exist is RCS. RCS makes use of the @ symbol in it's history
- file. If you insert an @ into your source then all RCS does is to escape this
- by using @@ in the history file. This is of course the correct and quite
- natural way to do things.
-
- Of course both systems have the expansion of characters to different
- characters. This is a feature and not a bug and is quite independant from the
- problems I faced.
-
- The next time I decide to use a control system I'll probably investigate CVS
- as I've heard so many good (and only good) reports.
- --
- James Bonfield (jkb@mrc-lmba.cam.ac.uk / rince@dcs.warwick.ac.uk)
- Medical Research Council Laboratory of Molecular Biology, Hills Road,
- Cambridge, CB2 2QH, England. Tel: 0223 402499 Fax: 0223 412282
-