Jelly in itself doesn't, but Maven which depends on Jelly does, thanks to another piece of OSS called werkz (see werken.com) which has the notion of Goal, analog to Ant's targets. You'd have to ask on the Jelly list (commons list in fact) for more details I think. I didn't realize you needed the execution of dependent targets in what you have in mind, just that you wanted to take advantage of the execution of Bean-like Java code from XML, that both Jelly and Ant provide, in different ways though.
In any case, Jelly can use Ant tasks, so you can always go the Ant way, and later switch to Jelly if need be. --DD -----Original Message----- From: Scott Walters [mailto:[EMAIL PROTECTED]] Sent: Friday, January 17, 2003 3:45 PM To: Ant Users List Subject: Re: ant as a workflow engine I will look into it. The idea of using Ant for this appealed to me due to it's ability to execute java components in a particular order based on depends attributes coded in the ant targets. Does Jelly provide this capability without having to write script code to enforce the dependencies? Scott Dominique Devienne wrote: > Have you looked into Jelly? You may find that it suits your needs better in > some ways. Ant was design as a build tool and thus its keepers resist > scripting capabilities that you might need. Ant is more mature, but Jelly is > already used by a few projects, notably Maven (which incidentally used to be > built on Ant, and is now built on Jelly...). Hope that helps, --DD > > -----Original Message----- > From: Scott Walters [mailto:[EMAIL PROTECTED]] > Sent: Friday, January 17, 2003 12:45 PM > To: [EMAIL PROTECTED] > Subject: ant as a workflow engine > > I'm in the process of architecting a rewrite of our external system > interface modules. In general, these modules extract specific data from > our database, send it somewhere, poll for a response and then import the > response back into the database. They do this with several vendors, > each with their own data requirments and transport mechanisms. > > There's a lot of code reuse possibilities since many of the vendors use > very similar data formats and HTTP POST transport mechanisms. > > Given this, I'm seriously considering implementing the individual pieces > of this project as ant tasks and processing each request by > programatically invoking a build script that controls the execution of > these tasks. Requests would come into a servlet, which would invoke the > ant script in a new thread, specifying the ant target that implements > the particular interface requested. > > I'm certainly no ant expert, however. All I've done is create some > fairly simple scripts to build and jar java code. My specific questions > are... > > 1) Can I pipe the xml output of one task into the next task without > using a file? > > 2) Do tasks within a target always run in the order they're specified > in the build script? > > Any other advice or comments are welcome. > > Thanks, > Scott > > > > > -- > To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> > For additional commands, e-mail: <mailto:[EMAIL PROTECTED]> > > -- > To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> > For additional commands, e-mail: <mailto:[EMAIL PROTECTED]> > > -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]> -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
