Hi Kristis, 

> It seems that this capability could be implemented on 
> activity_commit as an additional policy. E.g. add in 
> glue.conf a policy configuration variable that stores the 
> commands that should be executed on the server after a 
> successful commit.

The thing is that this repository does not need a Bug ID in order to
change any of the files as they are all documents which we don't track
with bug IDs.  Hence I didn't want to link it into any of the bug ID
handling code.

> I had a look over your text document. It seems that you want 
> to execute the tool describe (that issued the activity) every 
> time you commit. This is because you assume all your 
> documents are PDF files. Perhaps this could be further 
> generalized to provide regular expressions that given a match 
> will execute a program on the daemon server (which in your 
> case would be '.*pdf'.

It's not quite the case.  We store all sorts of different doc types in
our document repository.  I then have a set of code plugged into a post
commit hook that checks several things to do with the file (one of which
is if it is of a supported conversion type) and if it needs converting
it will be copied out of subversion and then the command line message
called on the Daemon to convert the file.  After that the newly
generated pdf file is copied to a location viewable on our intranet.
(Don't want sales and support guys messing about with Subversion).

As you can imagine I have put a lot of logic into my hook script that
would not be applicable to any-one else, so I wanted to keep the
interface very generic so that it could be re-used by anyone just to run
a generic command (Removing any business logic from Scmbug for this
custom code).


> The first question that comes to mind is do we need to supply 
> arguments to that script ? What does your script do right 
> now, precisely ?

It isn't actually a script that is being called based on this call to
the Daemon, it is actually a PDF generation program that has a set of
arguments passed to it.

Hope this helps explain why I proposed to keep the Scmbug Daemon
interface very generic.

Thanks, Rob
_______________________________________________
scmbug-users mailing list
[email protected]
http://lists.mkgnu.net/cgi-bin/mailman/listinfo/scmbug-users

Reply via email to