🚀 go-pugleaf

In article <1992Nov2.150839.13091@brtph560.bnr.ca>, 
chrisf@brtph316.bnr.ca (Chris Fulmer) writes:
|> In article <1992Nov2.091110.4435@ericsson.se>, 
|> etxmesa@eos.ericsson.se (Michael Salmon) writes:
|> |> In article <1992Oct29.182305.14061@brtph560.bnr.ca>, 
|> |> chrisf@brtph316.NoSubdomain.NoDomain (Chris Fulmer) writes:
|> |> |> In article <1992Oct28.074024.14827@news.tu-graz.ac.at>,
|> |> |> tab@piis05.joanneum.ac.at (Behnam Tabatabai) writes:
|> |> |> > Is there a possibility to check out only a subset of files managed by
|> |> |> > RCS or SCCS by assigning a symbolic names to different subsets? 
|> |> |> > 
|> |> |> 
|> |> |> Yes there is in RCS at least.  Look at the -N and -n
|> |> |> options to the rcs command.  It's then possible to check
|> |> |> out a version of a file by specifying its symbolic name.
|> |> |> 
|> |> |> The only thing that you'd need to do is make a short shell
|> |> |> script which looped through all the files, checking out the
|> |> |> version specified by the symolic name of each file.
|> |> 
|> |> The original poster wanted to group files the other way, not by
|> |> revision but by name, so that if I say co xyz where xyz is a group name
|> |> then I will get abc, def, and ghi which I previously defined as
|> |> belonging to xyz. IMHO this is only useful in systems like VMS that
|> |> don't have make by default.
|> 
|> Yeah.  That was my point.  I was trying to say that you could mark
|> each node of a specific configuration.  You then make a shell script
|> which says something along the lines of:
|> 
|> foreach f (RCS/*,v)
|> co -r $1
|> end
|> 
|> Call this script "myco", and suddenly you can do a "myco xyz."

I am sorry if I sound a little dumb, I am somewhat tired today but I
can't see how this script works. Firstly co requires a file name and
secondly we have already established that we aren't interested in
symbolic revisions, more symbolic directories. It would appear that you
intend that if a symbolic revision doesn't exist then it won't be
fetched. I guess that this will work but you then lose the ability to
specify revisions and that is the whole point of RCS. If a parameter
must be used then state seems better i.e. merely horrible.

Relating CMS groups to symbolic directories however has lead me to
think that the solution to the original query could lie with symbolic
links. The idea is to create a directory with the group name in the
archive directory, following the above name RCS/xyz, in that directory
we create symbolic links RCS/xyz/abc,v -> RCS/abc,v etc. We can then
say co RCS/xyz/*, this is a little clunky but functional as RCS doesn't
create additional files. An attractive alternative is to use a group
file e.g. RCS/xyz,g which contains names of the members of the group
and then say co `co -p xyz,g`. This has the advantage of allowing the
group to change with time.

-- 

Michael Salmon

#include	<standard.disclaimer>
#include	<witty.saying>
#include	<fancy.pseudo.graphics>

Ericsson Telecom AB
Stockholm