I have been told that TWiki (either stable release or latest development release, I assume) doesn't have support for using CVS as a storage backend for revisions and content, despite alot of talk about the subject a couple of years ago. I'm a bit familiar in Perl, so would be willing to try and hack it into my local copy and then give it back to the dev community. However, I need some pointers like what files I need to start looking at and what issues I need to keep in mind. I don't have alot of tiem, so can't afford to learn the entire Twiki source tree. Thanks

-- AdamTheo - 23 Nov 2002

TWiki currently support RCS and a lite Perl version called RcsLite. Both of these are implemented as Perl objects. Module TWiki::Store passes calls through. The implementation being used is deterimined by TWiki::storeTopicImpl, which is defined in TWiki.cfg. I'm not sure how well CVS can be mapped to the semantics being used by TWiki. I think you'll find this take a fair amount of time. For good TWiki compatiblity I think you'll want to be able to convert to/from RCS files as transparently as possible - RcsLite might help you here as RCS binaries are not required.

-- JohnTalintyre - 24 Nov 2002

OK, thanks. It looks like it's alot more complicated than I was hoping, and will probably have to be a long term goal that I work on over the next couple of months unless the core team or someone else implements this into the CVS version, in which case I'll get the CVS and crash test it.

-- AdamTheo - 30 Nov 2002

How's it going, Adam?

-- GrantBow - 25 Jan 2003

Hello, Adam!

In my workplace we have much html pages stored in our CVS 
server. We would like to import them to twiki.
I tried to do import script and discovered that it is
something which you are planning to do, but your plan
to integrate CVS to TWiki seamlessly is more smart, but 
more difficult to do too.

I think that it is quite easy to do two scripts: One
which import data from CVS -server to TWiki -directory 
and other which export data from TWiki -directory to CVS
-server.
These script should be called from cron, for example, at
every midnight. This is "quick and dirty" way to use
CVS-server as a TWiki backend. More elegant way is to run those
script every time after twiki calls /usr/bin/ci and
/usr/bin/co

- - -

function import()
{
   echo $SOURCE
   rm $SOURCE* -rf
   h1="%META:TOPICINFO{author="guest" date="1076081480" format="1.0" version="1.1"}%"
   h2="%META:TOPICPARENT{name="HomePage"}%"
   echo $h1>SOVA/tmp.txt
   echo $h2>>SOVA/tmp.txt
   echo "
">>SOVA/tmp.txt
   cvs checkout $SOURCE   
   cat  $SOURCE >>SOVA/tmp.txt
   echo "
">>SOVA/tmp.txt
   mv SOVA/tmp.txt $SOURCE
   /usr/bin/ci -q -l -m'none' -t-none -w'guest' /opt/TWiki/alpha2003-12-01-1258/twiki/data/$SOURCE
   chmod o+x SOVA -R
}

export SOURCE=SOVA/Sovellusvalikko.txt
import
SOURCE=SOVA/SOVA.AsiakasTiedote.txt
import
# TWiki doesnt accept filenames where is two dots. Therefore
# first dot has to replace with underline.
mv $SOURCE SOVA/SOVA_AsiakasTiedote.txt
exit

-- VeijoRyhanen - 10 Feb 2004