ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Peter Donald <dona...@apache.org>
Subject Re: [DISC] details of task library concept
Date Wed, 23 May 2001 15:00:37 GMT
At 10:39 AM 5/23/01 -0400, Daniel Barclay wrote:
>Peter Donald wrote:
>> 
>....
>> Ages ago I suggested a manifest attribute
>> 
>> Ant-Class-Path: foo.jar
>> 
>> This would search all the entries in "ant.lib.path" and "ant.tsk.path" to
>> find foo.jar. This was the only mod I needed to do all complicated
>> inheritance etc. (though I had to manually throw an exception if the
>> dependency DAG had classes that depended on incompatible versions etc).
>
>See if the Java package versioning system can be used for that.
>(See http://www.java.sun.com/j2se/1.3/docs/guide/versioning/ .)
>
>It provides for providing an abstract name and version numbers for an
>abstract package (not a Java package name but chunk of things in a Jar 
>file). 
>
>I thought it provided for listing depended-on packages, but maybe I'm
>confusing it with the extension loading mechanism and Jar file manifest
>Class-path: headers.
>
>
>Ant could use the regular class path (maybe with $ANT_HOME/ext/* included)
>to find Jar files, and use package versioning to identify what was found, 
>and then use dependencies (in package versioning if it does support 
>dependencies, or your proposed manifest attribute modified to use abstract 
>package names instead of concrete Jar file names) to load everything
>needed.

Thats one way of doing it but it requires excessive scanning (especially if
we have small jars with few tasks in it) and is not intuitive. I suppose it
is better theoretically - it is just practically I don't think it is pretty
(especially as most people don't add versioning data into jar - which would
mean you could never depend on it).

Cheers,

Pete

*-----------------------------------------------------*
| "Faced with the choice between changing one's mind, |
| and proving that there is no need to do so - almost |
| everyone gets busy on the proof."                   |
|              - John Kenneth Galbraith               |
*-----------------------------------------------------*


Mime
View raw message