On Aug 10, 2009, at 1:44 PM, Donohoe, Mark wrote:
Laser, Mary wrote:
I agree. We did it mostly as an example to show our
potential. But now I'm thinking that we shouldn't have
released it until we actually use the data. Right now it's a
cost with no benefit unless people are doing direct db
queries or are counting on a UI to use it in the future. The
only feedback I've gotten so far is from folks who do not
know what this agent does.
So far we have the following options:
1) Do nothing. Leave as is.
2) Make pkgmetagetta a separate package (requires minor code
changes and pkgmetagetta will never do selftest even if installed)
3) Leave packaging alone, but make same minor code changes as
in 2, so that pkgmetagetta is NEVER run.
I don't feel strongly about this but am asking for feedback
to find out if users would like a change.
Votes
taggart: option 1 or 2
bobg: option 1, 2, or 3 ;-)
Mary: option 2 or 3
I'd like to counter propose:
2) investigate what it will take to make pkgmetagetta a separate
package. If the work is not too bad (2 person weeks?) then let's just
do it in 1.2. If the investigation shows it's ugly, then let's do
option 3 for 1.2 and make optional agents the next release.
FYI, your proposal is actually what is on the task list.
Matt says the (debian) packaging is easy.
Code changes include:
1) move Scheduler.conf to db
2) a little bit of UI code to get the available agents from the db
instead of hardcoded
3) change scheduler and selftest to use db
Sounds easy to me.
Bob Gobeille
[email protected]
_______________________________________________
fossology mailing list
[email protected]
http://fossology.org/mailman/listinfo/fossology
