Re: Using rlog to look for files associated with a particular label

[Hugh Sasse]

On Thu, 16 Aug 2007, Cannon, Alastair, VF UK, Partner wrote:

> Hi,
> 
> I am porting (or trying to) a ksh unix script to perl. In the ksh script
> rlog is used in the generation of a list of file names and versions
> associated with a particular label. It basically seems to do it by
> rlog'ing *every* file in the repository and looking to extract the
> version from the line containing the label in question.
> 
> I am very new to RCS - is there a more elegant way of doing this? has
> anyone out there done it in perl who could give me some pointers?

rlog, well RCS in general, doesn't consider file collections as a
whole.  It really only versions single files, and keeps the
information for each file in an associated ,v file.  Thus to find
out if the history of a file in your code base has something you
must 'query' that ,v file, i.e.  you must search all of them to
cover the whole base.

Possibly programs like glimpse could help, as they index whole
directories, but rlog 'knows' how to parse ,v files, filtering out
the details you don't seem to need.  Showing us the relevant
chunk(s) of code may lead to suggestions on how to convert this
thing to Perl effectively.

It may be that other source code management systems can do better,
but then you create a new problem of how to migrate from RCS to the
other system.
> 
> Any advice or guidance would be most welcome
> 
> best regards
> 
> Alastair Cannon
>