> -----Original Message-----
> From: Diane Holt [mailto:[EMAIL PROTECTED]
> Sent: Tuesday, May 15, 2001 12:51 PM
> To: [EMAIL PROTECTED]
> Subject: Re: [PATCH] changes to ant start script for UNIX
>
[snip]
>
> I'm not completely sure the test is all that useful (if
> JAVACMD ends up
> being set to just plain 'java', then unless the user is in
> the directory
> where 'java' lives, the test is going to fail -- so it seems
> to really be
> more a test for JAVA_HOME being set [and should we really be
> requiring it
> is?]).
In the ant script, there already is a test for the JAVA_HOME variable (see
line 111 of version 1.6).
> But rather than take the test back out, we might just
> want to stop
> telling people to set JAVACMD this way and instead add the
> memory-increase
> flag to ANT_OPTS, since I suspect that's actually the more
> correct place
> to put that sort of thing.
Ah, this is another one of those "undocumented features", just like JAVACMD
was.
> Could you try doing that and see
> if that works
> for you?
I've just tried this, and ANT_OPTS works just fine. Thanks.
> (Or I suppose we could change the test to something like:
> [ ! -x "`echo $JAVACMD | awk '{print $1;}'`" ])
>
> Diane
>
I guess I formally remove my patch from the submission list.
Thanks Diane
JDGlanville
