Re: rcs -> cvs .....

[Derek R. Price]

Jody Brownell wrote:

> The only way I have been able to succeed it to copy files from the working
> RCS
> version into the working CVS directory. Then perform an update. The only
> problem is
> if the working copy of the CVS is the latest in the repository, no merge is
> performed.
> The only work around I could find would be to do a *dumb* commit on the file
> from a
> different working directory and redo the update/merge. This works enough to
> be able to
> remove any discrepencies between the RCS and CVS track, which I will later
> commit.
>
> Is there any other way to perform the merge/update from RCS to CVS with out
> the copying,
> *dumb* commit etc?

Do your dumb commit to a branch which is based on the last RCS version you
checked in and then merge the changes into the CVS trunk:

-------bTRUNK----------
  \
   \---------bRCS------------
    \-split  \-merge1    \-merge2

'split', 'merge1', and 'merge2' are freezepoints on the bRCS branch.  'split'
represents the point where the CVS development track (bTRUNK) and the RCS
development track (bRCS) were identical.  So after you do a dumb merge into
bRCS and tag it as 'merge1', you can let CVS handle the merge into the trunk:

    cvs update -A -jsplit -jmerge1
    cvs ci

The same will work the next time you merge the RCS development track as long as
you keep the tags up to date.  In the example, you will merge from 'merge1' to
'merge2':

    cvs update -A -jmerge1 -jmerge2
    cvs ci

Derek

--
Derek Price                      CVS Solutions Architect ( http://CVSHome.org )
mailto:address@hidden     OpenAvenue ( http://OpenAvenue.com )
--
I will not belch the National Anthem.
I will not belch the National Anthem.
I will not belch the National Anthem...

          - Bart Simpson on chalkboard, _The Simpsons_