ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Guillaume Rousse <>
Subject Re: [JPackage-devel] FW: jakarta ant 1.4.1/1.5 ant script
Date Tue, 07 May 2002 12:49:33 GMT
Ainsi parlait Mardi 7 Mai 2002 12:38, Stefan Bodewig :
> > Moreover, i still it is an error to have ant developpers taking care
> > of this.
> Let's go a few steps back and see where we started from, OK?
> My main concern is that the ant script that shipped with the Ant RPMs
> in the past were too different from the ones in Ant's CVS.
This ant rpm is the one maintained by Henry outside of jpackage project. I 
think this rpm should be targeted as a standalone rpm, providing all optional 
tasks (including those relying on commercial libs), and not relying on any 
external dependencies, thus using the default ant script. Otherwise it is 
clearly redundant compared to jpackage one.

> We got some bug reports that we couldn't really understand until we
> finally got a snippet of the wrapper script the people were using -
> and the we've sent the same people away with "make your administrator
> stop messing with our scripts" as we didn't know it were yours.
> We advice people to put libraries into ANT_HOME/lib or set the
> ANT_ARGS environment variable and so on ...
> We agreed that things would be a lot better if our scripts were closer
> to each other and in my and Nicolas' opinion the best thing would be
> if you didn't need to use a wrapper script of your own at all.  It
> seems as if we really could make that happen.
I still don't think it is totally possible (no reason to use file location 
test for a packaged ant, for instance), neither a really good idea: you don't 
want to depend on future jpackage reorganisation (ex: moving ant jars to 
/usr/share/ant/lib instead of current /usr/share/java as all other libs), nor 
do we want to depend on future release of ant to change any filename, if 
needed. Moreover, having a single script for two different context brings 
additional uneeded complexity.

However, having a common behaviour is achievable, and the best way so far 
would be sharing java functions. So far, Nicolas commited all needed changes, 
and we could even grant you direct commit access if needed.

Also, i'm still OK to revert jpp setup to a real ANT_HOME stragegy anyway, to 
allow default ANT_HOME/lib behaviour, and to provide support for original ant 
env vars in our script.
Guillaume Rousse <>
GPG key

To unsubscribe, e-mail:   <>
For additional commands, e-mail: <>

View raw message