Does SCCS allow you to "un-delta" a version???

Does SCCS allow you to "un-delta" a version???

Post by Bill Irw » Tue, 28 May 1991 16:31:37



I am just getting acquainted with the SCCS system commands and
have found that sometimes you can think a modification is
complete and delta it, only to discover later that there are
still some bugs in it.  Is it possible to Get the most current
version back, make some changes, then delta the same version
number again?

Or is the delta step a final one, where further mods have to
produce a new version?
--
Bill Irwin    -       The Westrheim Group     -    Vancouver, BC, Canada
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
uunet!van-bc!twg!bill     (604) 431-9600 (voice) |     Your Computer  

 
 
 

Does SCCS allow you to "un-delta" a version???

Post by Martin Gemm » Tue, 28 May 1991 22:03:15



Quote:>                        Is it possible to Get the most current
>version back, make some changes, then delta the same version
>number again?

Try sccs fix

If your last checked-in version of prog.c was 2.4 then
do 'sccs fix -r2.4 prog.c'.  The delta of 2.4 will be removed
from the history file and the new delta will be 2.4.
Personally I think there are two drawbacks: you _must_ supply
the -r option, and you can't fix the first version you checked
in.

Now while we're at it, does anybody think SCCS is better than
RCS or vice versa? Which system is newer/better? Is there a version
of make that supports RCS? If not, why?
RCS stands for Revision Control System by Walter F. Tichy.

--
Martin

 
 
 

Does SCCS allow you to "un-delta" a version???

Post by Chris Lew » Wed, 29 May 1991 06:23:20



>I am just getting acquainted with the SCCS system commands and
>have found that sometimes you can think a modification is
>complete and delta it, only to discover later that there are
>still some bugs in it.  Is it possible to Get the most current
>version back, make some changes, then delta the same version
>number again?

Do a "get -k s.<file>" to get the file back out in edittable form
(writeable, and keywords unexpanded), and then do a "rmdel -rA.B s.<file>",
where "A.B" was what the "get -k" reported the revision to be.
This will remove the delta you just applied.  Then, to get the SCCS
file ready for the delta after you've fixed it, do a:
        get -e -p s.<file> > /dev/null
(which gets out the previous version for editting, but throws it away)

Then you can change <file> and delta it again - it will end up with the
same revision level as the original delta that you *ed up.

If the mistakes were made in the lines that I had altered in the
delta, I just chmod 644 the s.<file>, edit it, then "admin -z s.<file>"
(to fix the checksum) and then chmod 444 s.<file>.  This is dangerous
(of course), and should only be done on the lines that were *added*
by the delta that was *ed up, otherwise you lose historical
tracking and/or SCCS header validity.
--

UUCP: ...!cunews!latour!ecicrl!clewis; Ferret Mailing List:


 
 
 

Does SCCS allow you to "un-delta" a version???

Post by Len Re » Thu, 30 May 1991 02:32:54



>Now while we're at it, does anybody think SCCS is better than
>RCS or vice versa? Which system is newer/better?

RCS is easiser to use than SCCS and can be bought (from MKS) for
DOS and OS/2.  It's mechanism for inserting the checkin log into
the file text is nice ($Log$), and the fact that you don't have to
work with unsubstituted "what" commands in the editable file is preferable.
I like being able to say "co file.c; vi file.c"
instead of "get -e s.file.c; vi file.c"--I don't want to type "s.", I can
cause RCS to always check out editable versions.  The "s." nonsense is
real annoying with csh or ksh.  In csh, I can say

co file.c
vi !$

No such luck SCCS.

As to make, get dmake from Waterloo or Gnu-make.  Both support RCS rules
as part of a generalization of implicit rules.  Both are far better
than regular make.  Both have make-time ifs and pattern substitution
and lots else.  Both are free.

--
Len Reed
Holos Software, Inc.
Voice: (404) 496-1358
UUCP: ...!gatech!holos0!lbr