Hi Ant developers,
I also have a second suggestion which I have already posted to this list on 
25.2. but I didn't receive any feedback. I don't know if this means that my 
suggestion has been rejected so I'm trying again. This time I will definitely 
consider a no-response reaction to be a rejection:)
Ok, the thing is that in some cases I needed to tell Ant that it is not a 
problem if in the buildfile there are any targets which depend on targets which 
are not present in the buildfile (or in any imported one) IF these tasks with 
missing dependencies are not going to be invoked in the current process. The 
reason for that was that I'm using Ant tasks (in fact Apache Ivy) which 
download Ant buildfile fragments from a repository. And these fragments are 
included using <import optional="true".../> so in the first run, when the 
fragments are being downloaded, the targets from these fragments are not 
available and there may be tasks in the main file that depend on them. However, 
there is no problem with that as these targets are not being invoked at the 
moment.
I've looked into the source code and found that this is absolutely easy to do 
because in fact there is a (well documented) redundant check of the 
dependencies of the targets which are not needed during the current build. I've 
added commandline option to disable this redundant check. So that's the first 
thing. Would you find it useful for submission to Ant?
Second thing is that I didn't like specifying this commandline option whenever 
it was necessary so I created a special Ant task which from inside a buildfile, 
where it is used in the toplevel, can disable this check (because fortunately 
top-level tasks in a buildfile are executed before dependency checking - this 
may, however, possibly change in the future). In general, it could be a way to 
configure Ant from within a buildfile which may be useful if some Ant options 
should be switched on/off for specific buildifiles. That's my second question, 
do you find this task useful?
Thanks for your opinions.
David 

Reply via email to