ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefan Bodewig <>
Subject Re: Task Submissions
Date Thu, 19 Jul 2001 08:18:28 GMT
On Thu, 19 Jul 2001, Steve Loughran <> wrote:

> Yes, a facade would work, although there is always the issue of how
> to have non standard options.

Yep, an issue all these facade tasks share.

The easiest (but incomplete) solution is to allow arbitrary additional
arguments to be specified, I'm going to provide this for javac unless
somebody gets there faster than me.  That way, you can specify
non-standard options if you know which implementation will be chosen -
I don't think we'll have some means other than magic properties or
finally defining and implementing the user preferences system right

> Buts its key weaknesses are
>  -does not (yet) do dependency based invocation.
>  -doesnt use any javac choice (a feature of the jasperc library
>   which I dont want to look at)

I think both are important features - depending on the number of JSPs
of course.  Users will also want that the chosen javac is the same the
<javac> task would use.

> One issue to worry about first: what use cases are we addressing
> with this task 
> 1. compilation of jsp pages purely as a syntax checker before
> bothering to deploy
> 2. precompilation of jsp pages into java files /.class files for
> inclusion in war files

I agree with you, first (1) and do it properly, then (2).  I think (2)
may also need knowledge about the target container as there is no
standard location for compiled JSPs AFAIK.  I'm not sure whether we
can assume that the jspc implementation will always be coupled to a
specific target container.


View raw message