Menu

#120 <relentless> task

closed-fixed
nobody
logic (23)
5
2006-08-09
2004-12-21
No

In followup to my original message to
ant-contrib-developers (see
http://sourceforge.net/mailarchive/forum.php?thread_id=6187123&forum_id=2912\)
I got a message from Peter saying to submit by email,
and a message from Curt to submit by bug (since he
would submit a bug as soon as I emailed anyway). So
I'm submitting the bug.

Anyway, <relentless> is a task that executes all its
contained tasks, regardless of whether one or more of
them fails. If all the contained tasks succeed,
<relentless> will succeed. If any of the contained
tasks fail, <relentless> will fail.

I've attached our current implementation for perusal.
Please be aware of the following:
- javadoc is not complete
- internal documentation via comment is lacking
- package is not correct for ant-contrib
- manual entry is not there at all
- no unit tests yet
- code is Java 1.5-centric

However, I thought it might be useful to give an
preview, so that if I'm doing anything grossly stupid
or orthogonal to ant-contrib practices, it could be
caught now rather than wait until everything was tied
up with ribbons and bows. Please let me know if you
spot something not mentioned previously that needs
correcting.

Thanks!

Discussion

  • Matt Inger

    Matt Inger - 2005-02-03

    Logged In: YES
    user_id=83032

    Sorry for the lateness, but i would need a manual page and
    unit tests as well. Can you provide these if possible? When
    you say "java 1.5 centric" does this mean it requires jdk 5.0
    to run?

     
  • Christopher Heiny

    Logged In: YES
    user_id=117164

    Thought I'd posted a note to the dev list.

    I'm stuck in Asia at a customer site, and haven't been
    able to give this any attention for several weeks. If you
    can wait till early next week, I'll try to wrap up the
    doc/javadoc/etc then.

    1.5-centric does indeed mean there are a few "Java 5"isms
    in the code that require jdk 5.0 to compile. I'm not sure
    if it's required to run (in theory the byte codes should
    work anywhere, right?). It won't be hard to remove those,
    however.

    BTW, would it be better to rename this as <keepgoing>, to
    keep it in sync with the keepgoing attribute in <for>?
    While <relentless> does have a certain style to it,
    <keepgoing> is probably more meaningful to the casual
    user.

     
  • Christopher Heiny

    Logged In: YES
    user_id=117164

    I'm back from Asia (finally!) and have completed the manual
    page, javadoc, and test cases. I've attached a tarball with
    a patch against CVS top of tree, and a bunch of new files.

    Not having heard back on the previous comment, I didn't
    change the name from <relentless> to <keepgoing>. If
    <keepgoing> is thought to be a better name, I'll switch it
    and resubmit.

     
  • Christopher Heiny

    New files and patch

     
  • Matt Inger

    Matt Inger - 2006-08-09
    • status: open --> closed-fixed
     
  • Matt Inger

    Matt Inger - 2006-08-09

    Logged In: YES
    user_id=83032

    this has been added.

     

Log in to post a comment.