Re: Ant 1.6.5 Classloading and JUnit Problem 1

2007-03-30 Thread Dominique Devienne
On 3/30/07, Christopher Gardner <[EMAIL PROTECTED]> wrote: Do you recommend I explicitly do this in my ant script: set CLASSPATH= Yep. ant.bat also supports a -nocp switch now (or something similar), which is highly recommended by the Ant team (AFAIK), and certainly myself. Just look at all th

Re: Ant 1.6.5 Classloading and JUnit Problem 1

2007-03-30 Thread Christopher Gardner
I didn't mean to imply that I don't set the classpath in the junit task. I definitely do that. I just don't change the environment variable anywhere, which is what you were talking about. Do you recommend I explicitly do this in my ant script: set CLASSPATH= (to null)? On 3/30/07, Dominique

Re: Ant 1.6.5 Classloading and JUnit Problem 1

2007-03-30 Thread Dominique Devienne
On 3/30/07, Christopher Gardner <[EMAIL PROTECTED]> wrote: I'm having cruise run the standard ant.bat in ANT_HOME/bin, but not explicitly resetting the CLASSPATH anywhere. On 3/30/07, Dominique Devienne <[EMAIL PROTECTED]> wrote: > On 3/30/07, Christopher Gardner <[EMAIL PROTECTED]> wrote: > > I

Re: Ant 1.6.5 Classloading and JUnit Problem 1

2007-03-30 Thread Christopher Gardner
I'm having cruise run the standard ant.bat in ANT_HOME/bin, but not explicitly resetting the CLASSPATH anywhere. On 3/30/07, Dominique Devienne <[EMAIL PROTECTED]> wrote: On 3/30/07, Christopher Gardner <[EMAIL PROTECTED]> wrote: > I just verified that I was using the antscript option in my crui

Re: Ant 1.6.5 Classloading and JUnit Problem 1

2007-03-30 Thread Dominique Devienne
On 3/30/07, Christopher Gardner <[EMAIL PROTECTED]> wrote: I just verified that I was using the antscript option in my cruise config.xml. Thanks for the idea, though. But do you reset the CLASSPATH env. var in that script? --DD -

Re: Ant 1.6.5 Classloading and JUnit Problem 1

2007-03-30 Thread Christopher Gardner
I just verified that I was using the antscript option in my cruise config.xml. Thanks for the idea, though. Chris On 3/30/07, Dominique Devienne <[EMAIL PROTECTED]> wrote: On 3/30/07, Christopher Gardner <[EMAIL PROTECTED]> wrote: > [...] The junit task doesn't override any forking or > class

Re: Ant 1.6.5 Classloading and JUnit Problem 1

2007-03-30 Thread Dominique Devienne
On 3/30/07, Christopher Gardner <[EMAIL PROTECTED]> wrote: [...] The junit task doesn't override any forking or classloading defaults. [...] Now, if I set the junit options to be fork="true" and forkmode="once", calling the master build.xml through the command line and eclipse works perfectly; ca

Ant 1.6.5 Classloading and JUnit Problem 2

2007-03-30 Thread Christopher Gardner
Same code base environment (projects a, b, c, and d) described in the prior posting "Ant 1.6.5 Classloading and JUnit Problem 1" (ant 1.6.5, Java 1.5, Windows, eclipse 3.2.2, CruiseControl). There is a third-party directory at the same level as the individual project directories. Ea

Ant 1.6.5 Classloading and JUnit Problem 1

2007-03-30 Thread Christopher Gardner
I'm having a couple of ant problems that appear to be related to classloading. The first problem may be more of a CruiseControl question, but I'll try this group as well. I'll describe the second problem in another posting. In a Windows environment with Java 1.5, I have a master build.xml file

ant/junit problem with FileNotFound

2006-04-26 Thread glenn opdycke-hansen
I have several ant scripts that invoke the next ant script via . The last ant script has a task that invokes Java code. There is a problem in the Java code that the File class cannot find the file with a relative path. I can resolve this by using the complete path to the file. This problem does no

Re: [tag] Re: The good ol' JUnit problem

2005-11-14 Thread Peter Reilly
Not quite, one can use the "classloader" task as implemented in bugzilla 28228. http://enitsys.sourceforge.net/ant-classloadertask/ Please note that this uses protected methods on the java classes and thus may not be future compible. However, I do think that optional ant tasks should not require t

RE: [tag] Re: The good ol' JUnit problem

2005-11-14 Thread Jon Skeet
> i didn't quite catch the resolution to this. is there a > solution to using a stock ant distribution , without > modifying your target environment (ANT_HOME/lib, ~/.ant/lib, > etc), and using the optional tasks such as junit, scp, ftp, > etc? if so, can someone give an example of how to do t

Re: The good ol' JUnit problem

2005-11-12 Thread Mark Lybarger
i didn't quite catch the resolution to this. is there a solution to using a stock ant distribution , without modifying your target environment (ANT_HOME/lib, ~/.ant/lib, etc), and using the optional tasks such as junit, scp, ftp, etc? if so, can someone give an example of how to do this? On 11/9

RE: The good ol' JUnit problem

2005-11-09 Thread Jon Skeet
> --- Jon Skeet <[EMAIL PROTECTED]> wrote: > [SNIP] > > I've made the change (as well as fixing up some HTML > > problems) and > > mailed the file to Steve for check-in. It looks like I > don't have SVN > > access after all - either that, or I've got a different username or > > password to befor

Re: The good ol' JUnit problem

2005-11-08 Thread Stefan Bodewig
Hi Jon, long time no see. On Tue, 8 Nov 2005, Jon Skeet <[EMAIL PROTECTED]> wrote: > Assuming I actually *do* still have write-access, shall I add these > two properties to the manual in SVN? No, you don't. This is the current list of people with write access to Ant's svn. ant=rubys,conor,bod

RE: The good ol' JUnit problem

2005-11-08 Thread Matt Benson
--- Jon Skeet <[EMAIL PROTECTED]> wrote: [SNIP] > I've made the change (as well as fixing up some HTML > problems) and > mailed the file to Steve for check-in. It looks like > I don't have SVN > access after all - either that, or I've got a > different username or > password to before. > To commi

RE: The good ol' JUnit problem

2005-11-08 Thread Jon Skeet
> FYI ${ant.home} is documented on > using.html#built-in-props > > ${ant.library.dir} should probably be documented alongside it > and share its caveat. Yup - I noticed that when I dived into SVN. (It's not in the online manual, which is what I was looking at before.) I've made the change (

Re: The good ol' JUnit problem

2005-11-08 Thread Matt Benson
--- Steve Loughran <[EMAIL PROTECTED]> wrote: > Jon Skeet wrote: > > Steve wrote: > > > >>${ant.home}/lib > >>${user.home}/.ant/lib > > > > > > Having just checked the source, it's even easier > than that (for the > > first one): ${ant.library.dir} > > > > Both ant.home and ant.library.dir w

RE: The good ol' JUnit problem

2005-11-08 Thread Jon Skeet
Steve wrote (replying to me): > > I think if Ant provided a property for its "home" location > ${ant.home}/lib > ${user.home}/.ant/lib You star. Could I suggest that ant.home should be added to the manual? (I did look on the website first, honest!) I suppose I could just edit it in CVS myself

Re: The good ol' JUnit problem

2005-11-08 Thread Steve Loughran
Jon Skeet wrote: Steve wrote: ${ant.home}/lib ${user.home}/.ant/lib Having just checked the source, it's even easier than that (for the first one): ${ant.library.dir} Both ant.home and ant.library.dir work from Eclipse, too, which I had worried about a little bit... Assuming I actually *do

RE: The good ol' JUnit problem

2005-11-08 Thread Jon Skeet
Steve wrote: > ${ant.home}/lib > ${user.home}/.ant/lib Having just checked the source, it's even easier than that (for the first one): ${ant.library.dir} Both ant.home and ant.library.dir work from Eclipse, too, which I had worried about a little bit... Assuming I actually *do* still have write-

Re: The good ol' JUnit problem

2005-11-08 Thread Steve Loughran
Jon Skeet wrote: This is a similar issue with the xalan classes in the 1.4 jdk. they're exposed directly by sun (and other jvm's) as org.apache.*. basically they just imported xalan and possibly other apache classes into the jvm. this gives your entire application (jvm) one chance to override

RE: The good ol' JUnit problem

2005-11-08 Thread Jon Skeet
> This is a similar issue with the xalan classes in the 1.4 jdk. > they're exposed directly by sun (and other jvm's) as org.apache.*. > basically they just imported xalan and possibly other apache > classes into the jvm. this gives your entire application > (jvm) one chance to override the ver

Re: The good ol' JUnit problem

2005-11-04 Thread Mark Lybarger
Jon, This is a similar issue with the xalan classes in the 1.4 jdk. they're exposed directly by sun (and other jvm's) as org.apache.*. basically they just imported xalan and possibly other apache classes into the jvm. this gives your entire application (jvm) one chance to override the version u

The good ol' JUnit problem

2005-10-28 Thread Jon Skeet
Currently, our build uses a modified distribution of Ant - all the optional libraries we need are bundled into the lib directory. We use a similarly modified version of Eclipse (i.e. the Ant part has the same changes made). Needless to say, this is somewhat suboptimal. I want to move to a system w

JUnit problem with ANT 1.6

2004-01-13 Thread Moses Guillen
This problem is similar to John Hohlen (1/6/2004) "Can't Run JUnits With ANT 1.6" I check Antoine's reply, however no luck. Here are my findings: I am in the process of migrating build scripts from ANT 1.5.4 to 1.6. The migration has gone relatively smooth until, I execute a target that has JUn

Re: junit problem

2003-07-15 Thread ½É¿ì°ï
Dear Han, My suggestion is: -- Original Message -- From : Jimmy Han([EMAIL PROTECTED]) To : [EMAIL PROTECTED] Sent : Wednesday, Jul 16, 2003 07:59 AM Subject : junit problem > I know this has been asked for manu times