stack trace filtering for formatters. It will filter out
those unneeded frame from Ant and JUnit so the stack will
be much more readable.
nb: Looks like the JUnit task need serious refactoring since we have
been adding feature incrementally. It starts to be a real mess,
I can barely read my own code :-(
git-svn-id: https://svn.apache.org/repos/asf/ant/core/trunk@270099 13f79535-47bb-0310-9956-ffa450edef68
This is long for a test (=harder to tell which assert failed), but if you are going to sleep between touches then you should make the most of the test.
git-svn-id: https://svn.apache.org/repos/asf/ant/core/trunk@270096 13f79535-47bb-0310-9956-ffa450edef68
I also refactored the task a little bit to support directly primitive types such as
int or boolean and fixed a couple of mistakes about error handling in the original code.
It should be a good starting point for further refactoring.
pr: 5138, 5187
by: stevec@ignitesports.com (Steve Cohen)
git-svn-id: https://svn.apache.org/repos/asf/ant/core/trunk@270091 13f79535-47bb-0310-9956-ffa450edef68
factory so messages will get back to Ant's logging system.
Submitted by: David Leuschner <leuschner-list@gmx.net>
git-svn-id: https://svn.apache.org/repos/asf/ant/core/trunk@270089 13f79535-47bb-0310-9956-ffa450edef68
Made sure that operations internal to Project could not generate these warnings by adding a new setPropertyInternal method that didn't do any checks.
Also made the getUserProperties and getProperties methods return copies of maps rather than direct references.
Updated WHATSNEW to warn about these issues.
Added some unit tests to verify that property immutability is enforced - where we actually do enforce it ;)
Submitted by: "Erik Hatcher" <jakarta-ant@ehatchersolutions.com>
git-svn-id: https://svn.apache.org/repos/asf/ant/core/trunk@270078 13f79535-47bb-0310-9956-ffa450edef68
Didn't worry too much about WaitFor's backward compatibility issues as it is hardly a week old.
WaitFor has ben modified this way so that it is consistent with Tstamp and takes in enumerated attributes for unit instead of allowing arbitrary letters/ words suffixed to the time.
git-svn-id: https://svn.apache.org/repos/asf/ant/core/trunk@270074 13f79535-47bb-0310-9956-ffa450edef68
Submitted by: alan@coldsmoke.com (Alan Arvesen), greg_fenton@yahoo.com (Greg Fenton)
Reviewed by: Steve Loughran
Fixed the 4NT test as suggested, tested it by running the batch file without echoes. 4NT 4.00A actually works properly without the fix, because it understands NT CMD shell variables better. This fix should support earlier versions.
git-svn-id: https://svn.apache.org/repos/asf/ant/core/trunk@270057 13f79535-47bb-0310-9956-ffa450edef68
timeout occurs.
Submitted by: Denis Hennessy <denis@network365.com>
As a side effect, <condition> gains two new conditions. This also
means we now have a way to check whether we are on a networked system
and run/omit testcases that require network to be established based on
such a condition.
git-svn-id: https://svn.apache.org/repos/asf/ant/core/trunk@270050 13f79535-47bb-0310-9956-ffa450edef68
Change baseclass paramater to rootclass and make it use dot notation for
the classname rather than making it a file.
Also allow multiple nested <root> elements to define more than one root
class
Example:
<classfileset dir="build/classes" id="classes">
<root classname="org.apache.tools.ant.Main"/>
<root classname="org.apache.tools.ant.taskdefs.optional.ejb.EjbJar"/>
</classfileset>
git-svn-id: https://svn.apache.org/repos/asf/ant/core/trunk@270045 13f79535-47bb-0310-9956-ffa450edef68
resolveFile magic unless a filepaths has been specified.
Note that I've rewritten the checkFile() method in the filepath case
to match the comment in the sources (at least I think so).
Actually, I think we are generating too many passed checks, especially
with simple name specified == parent of parent dir + name but this is
a different issue.
git-svn-id: https://svn.apache.org/repos/asf/ant/core/trunk@270041 13f79535-47bb-0310-9956-ffa450edef68