ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From James Duncan Davidson <dun...@x180.net>
Subject Re: Question on builds that fail
Date Thu, 04 Jan 2001 10:37:11 GMT
On 1/3/01 2:14 PM, "Conor MacNeill" <conor@ebinteractive.com.au> wrote:

> Of course, once we add this to javac, the question is why not add it to
> other tasks, say the <ant> task for example. I think that was perhaps the
> thrust of David Bailey's fail fast post. Therefore, perhaps this should be a
> general attribute of all tasks. That would move "proceed" processing into
> the ant core and I'm pretty sure Duncan has expressed an opinion that this
> is a task responsibility.

It is a task responsibility. I'm not sure whether this is one that should
just be enforced by convention, or whether it should be part of some default
kind of TaskAction base implementation that is part of what I was talking
about in the MT post I just made.

If it was part of base functionality, a proceed of yes would case "false"
returns from the execute method to be warned, not stop the build.

If we do enable this for *everything*, then there might be problems that
people get into. Of course, the major tasks that fail on a regular basis are
going to be exactly those ones that people want to see this on anyways
(compilation tasks). Given that, I might be willing to see this go into a
base class impl.

-- 
James Duncan Davidson                                        duncan@x180.net
                                                                  !try; do()


Mime
View raw message