ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Julien Couvreur" <jcouvr...@redcart.com>
Subject RE: Controling the environment for exec tasks
Date Mon, 17 Jul 2000 17:41:46 GMT

Hi,

First, thanks to all you guys answering ! That's really great.


Now,

 >> Yes, that'd be the way to go, and maybe if value was omitted use
 >> System.getenv to get the current value?

Are the current values not passed by default. Under Unix the environment is
passed to the "son" created with "fork". Under Win32 I don't know. So "env"
or "setenv" or what ever it gets called would just overload and override the
environment.


Also, could somebody give me (or us) a hint why antRun is involved ? What is
it used for ? I'll go in Exec.java anyway, but having an overall picture is
always nice to catch up on the code : )


 >> Take a look at the modifications I've made to
 >> docs/index.html#writingowntask and tell me if that makes sense to you
 >> - and post any corrections please, English is not my native language.

About the createXXX and addXXX :
- this added documentation is mostly clear, as far as my english goes too !
- what is the difference between create and add. I went in
"MatchingTask.java" and there is only createInclude and createExclude, but
no addXXX.
- why not put a small example with createXXX, addXXX, and Class ReturnObject
? Or maybe just give a couple adequat classes reference in the Ant sources ?
- what constraints is there on the ReturnObject ? It seems to be constructed
as the task itself, with setters, but is it all ?


I will repost soon about the issue I had with "notavailable" and negation in
general.

And also, the building issue : chossing ant.build.dir=. could bring
something nicer than what it does (try it ! ;-).


See ya
Julien



Mime
View raw message