-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
"Molitor, Stephen" <[EMAIL PROTECTED]> writes:
> Kevin Burton writes:
> >> Hm. It could be possible to use both the 'ant' method and the 'java
> >> org.apache.tools.ant.Main' method. AKA if the command 'ant' does not
> >> exist within the PATH, try using the java method.
>
> There could also be an option to always use the 'java {...}ant.Main' method.
> If the bean shell was used, there wouldn't be a need to invoke a new JVM,
> which could speed things up significantly.
I spent some time researching this. From the way I can see it we have the
following problems.
Ant calls System.exit when it is done hence this cause the bsh to exit. The
solution here is to run it within a NoExitSecurityManager.
Annt would output to the *bsh* buffer and not the *compilation* buffer. We
would need to set `standard-output' so that this would exec from the *bsh*
buffer and output to the *compilation* buffer. I don't know enough about the
Emacs compilation system to determine if this would break anything.
> The cost would be that running Ant from within Emacs might behave differently
> than running Ant from the command line.
Hopefully the only difference will that it will be a lot faster :). I am going
to hack with it (when I get the time) and report my discoveries.
Kevin
- --
Need a good Engineer? Hire me! ( Java | P2P | XML | Linux | Open Source )
Kevin A. Burton [EMAIL PROTECTED], [EMAIL PROTECTED], [EMAIL PROTECTED] )
Jabber: [EMAIL PROTECTED] Cell: 415-595-9965 URL: http://relativity.yi.org
The worse thing in life is to fall short!
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: Get my public key at: http://relativity.yi.org/pgpkey.txt
iD8DBQE7xguUAwM6xb2dfE0RAqAvAKCvacAY/KV5+bpCLXKQ3YuJFbId+ACfeyD8
oPyN9r4B/ghDKi2WvXh93b0=
=wxX2
-----END PGP SIGNATURE-----
