From twisti@complang.tuwien.ac.at Wed May 21 15:21:00 2008 From: twisti@complang.tuwien.ac.at (Christian Thalinger) Date: Wed, 21 May 2008 15:21:00 -0000 Subject: build directory support Message-ID: <1211383268.24194.59.camel@imac523d.theobroma-systems.com> Hi! I've just commited a bunch of patches to add the missing Uses tags for various packages. This is a preparation for the real patch that adds build directory support. I decided to implement/fix this because I didn't want to copy a handfull of Mauve sources around to test it, but rather use one checkout on a NFS share. Currently it works very well (locally) and the Harness only compiles sources when there is no corresponding classfile or the sourcefile is newer than the classfile (inner classes are not checked). There's only one problem left: different files some tests depend on. Some testlets, like gnu/testlet/java/beans/SimpleBeanInfo/loadImage.java, require the file to be relative to the classfile location, so we need a way to copy these files to the build directory. On IRC I talked to Mark and Andrew and we concluded that a new tag would be the best solution. Like: // Files: testImage1.gif This tag would then copy the file from the source directory to the build directory, if it does not already exist. I'd like to commit my changes ASAP because I'm not sure I can find all broken testlets. - twisti From David.Herron@Sun.COM Tue Jun 17 22:37:00 2008 From: David.Herron@Sun.COM (David Herron) Date: Tue, 17 Jun 2008 22:37:00 -0000 Subject: "I'm Sun. Can I use the Mauve Project's testsuite?" Message-ID: <48583C7B.5090507@sun.com> Hi, I'm getting ready to request permission to officially use Mauve in-house on OpenJDK testing. Reviewing the license headers I found an instance of the Common Public License in a properties file. mauve//gnu/testlet/java/util/zip/InflaterInputStream/messages.properties I don't know yet if this will be a problem. The policy says I have to tell the lawyers about this other license ....soooooo... Hm, I also found a file under the Apache license and that the CORBA tests were imported from elsewhere and are under either LGPL or GPL. Then there is: mauve//gnu/testlet/java/util/regex/Pattern/pcrematches.java: /** Regex test suites from adapted from PCRE (http://www.pcre.org/license.txt). */ But that's just the BSD license. I really just wanted to send you an email with that question from your FAQ because it's such a cool question for me to ask. - David Herron From green@redhat.com Tue Jun 17 23:14:00 2008 From: green@redhat.com (Anthony Green) Date: Tue, 17 Jun 2008 23:14:00 -0000 Subject: "I'm Sun. Can I use the Mauve Project's testsuite?" In-Reply-To: <48583C7B.5090507@sun.com> References: <48583C7B.5090507@sun.com> Message-ID: <485844C2.1000206@redhat.com> David Herron wrote: > Hi, I'm getting ready to request permission to officially use Mauve > in-house on OpenJDK testing. Excellent! > > Reviewing the license headers I found an instance of the Common Public > License in a properties file. > > mauve//gnu/testlet/java/util/zip/InflaterInputStream/messages.properties > > I don't know yet if this will be a problem. The policy says I have to > tell the lawyers about this other license ....soooooo... Hm, I also > found a file under the Apache license and that the CORBA tests were > imported from elsewhere and are under either LGPL or GPL. > > Then there is: > mauve//gnu/testlet/java/util/regex/Pattern/pcrematches.java: /** > Regex test suites from adapted from PCRE > (http://www.pcre.org/license.txt). */ > > But that's just the BSD license. We've been flexible with the licensing because, among other things, we're not distributing binaries. > I really just wanted to send you an email with that question from your > FAQ because it's such a cool question for me to ask. > Indeed! I've been waiting years to hear it! (I really need to update that page..) Anthony Green > - David Herron > From tromey@redhat.com Wed Jun 18 23:03:00 2008 From: tromey@redhat.com (Tom Tromey) Date: Wed, 18 Jun 2008 23:03:00 -0000 Subject: "I'm Sun. Can I use the Mauve Project's testsuite?" In-Reply-To: <48583C7B.5090507@sun.com> References: <48583C7B.5090507@sun.com> Message-ID: <18523.455.660483.827689@fleche.redhat.com> David> Reviewing the license headers I found an instance of the Common Public David> License in a properties file. David> mauve//gnu/testlet/java/util/zip/InflaterInputStream/messages.properties David> I don't know yet if this will be a problem. The policy says I have to David> tell the lawyers about this other license ....soooooo... Hm, I also David> found a file under the Apache license and that the CORBA tests were David> imported from elsewhere and are under either LGPL or GPL. Yeah, I guess it is a mish-mash. Personally I wouldn't be averse to just deleting selected tests if it makes life simpler. Supposedly I added that zip test. I don't remember doing it :) For the CORBA tests, email Audrius. He should be able to tell you exactly where they came from; or it may be in the Mauve list archives. David> I really just wanted to send you an email with that question from your David> FAQ because it's such a cool question for me to ask. :) The Mauve web site is hilariously un-updated. It still mentions getting CVS from Cyclic. Tom From David.Herron@Sun.COM Wed Jun 25 00:43:00 2008 From: David.Herron@Sun.COM (David Herron) Date: Wed, 25 Jun 2008 00:43:00 -0000 Subject: Am I doing something wrong? Message-ID: <4861947D.7070705@sun.com> Just to try out Mauve I'm running it on my OS X machine with Apple's Java. ./configure --with-vm=/Library/Java/Home/bin/java make java Harness -verbose -showpasses -compile -vm java gnu.testlet I figure since y'all haven't changed Harness.java in 2 yrs (to any significant change) that it must be working fine and that there's some kind of driver error on my part. I get a bunch of messages saying FAIL: One of the dependencies failed to compile Then it throws this:- Exception in thread "main" java.lang.StringIndexOutOfBoundsException: String index out of range: -48 at java.lang.AbstractStringBuilder.setLength(AbstractStringBuilder.java:146) at java.lang.StringBuffer.setLength(StringBuffer.java:154) at Harness.processFolder(Harness.java:1123) at Harness.processFolder(Harness.java:1153) at Harness.processFolder(Harness.java:1153) at Harness.processTest(Harness.java:899) at Harness.runAllTests(Harness.java:706) at Harness.main(Harness.java:184) ---> sb.setLength(sb.length() - fullPath.length() - 1); So that calculation is going negative, and inserting println's I see sb.length() is 0 ... I changed that to something like this sb.setLength( (sb.length() - fullPath.length() - 1) > 0 ? sb.length() - fullPath.length() - 1 : 0 ); And then this failed (line 1208) nextTest = nextTest.substring(0, nextTest.length() - 5); From gnu_andrew@member.fsf.org Wed Jun 25 00:52:00 2008 From: gnu_andrew@member.fsf.org (Andrew John Hughes) Date: Wed, 25 Jun 2008 00:52:00 -0000 Subject: Am I doing something wrong? In-Reply-To: <4861947D.7070705@sun.com> References: <4861947D.7070705@sun.com> Message-ID: <17c6771e0806241751q25d39598sf08857a9ed83a55e@mail.gmail.com> On 25/06/2008, David Herron wrote: > Just to try out Mauve I'm running it on my OS X machine with Apple's Java. > ./configure --with-vm=/Library/Java/Home/bin/java > make > java Harness -verbose -showpasses -compile -vm java gnu.testlet > > I figure since y'all haven't changed Harness.java in 2 yrs (to any > significant change) that it must be working fine and that there's some kind > of driver error on my part. > > I get a bunch of messages saying FAIL: One of the dependencies failed to > compile > > Then it throws this:- > Exception in thread "main" > java.lang.StringIndexOutOfBoundsException: String index out > of range: -48 > at > java.lang.AbstractStringBuilder.setLength(AbstractStringBuilder.java:146) > at > java.lang.StringBuffer.setLength(StringBuffer.java:154) > at Harness.processFolder(Harness.java:1123) > at Harness.processFolder(Harness.java:1153) > at Harness.processFolder(Harness.java:1153) > at Harness.processTest(Harness.java:899) > at Harness.runAllTests(Harness.java:706) > at Harness.main(Harness.java:184) > > > ---> sb.setLength(sb.length() - fullPath.length() - 1); > > So that calculation is going negative, and inserting println's I see > sb.length() is 0 ... > > I changed that to something like this > > sb.setLength( > (sb.length() - fullPath.length() - 1) > 0 > ? sb.length() - fullPath.length() - 1 > : 0 > ); > > > And then this failed (line 1208) > nextTest = nextTest.substring(0, nextTest.length() - 5); > > > FWIW, we run Mauve in a very similar manner on builder.classpath.org: exec > Log.Std 2>&1 ./configure --with-ecj-jar=/usr/share/java/ecj.jar --with-vm=jamvm && make && xvfb-run jamvm Harness gnu.testlet -vm jamvm -showpasses -timeout 180000 (this is testing jamvm instead of java, and uses xvfb-run to avoid needing an X display). I did run Mauve a while ago on OS X and I don't remember having any problems. Things may have changed though. I believe twisti has done a lot of recent work on the harness, but I don't think it's committed yet. -- Andrew :-) Support Free Java! Contribute to GNU Classpath and the OpenJDK http://www.gnu.org/software/classpath http://openjdk.java.net PGP Key: 94EFD9D8 (http://subkeys.pgp.net) Fingerprint: F8EF F1EA 401E 2E60 15FA 7927 142C 2591 94EF D9D8 From twisti@complang.tuwien.ac.at Wed Jun 25 07:35:00 2008 From: twisti@complang.tuwien.ac.at (Christian Thalinger) Date: Wed, 25 Jun 2008 07:35:00 -0000 Subject: Am I doing something wrong? In-Reply-To: <17c6771e0806241751q25d39598sf08857a9ed83a55e@mail.gmail.com> References: <4861947D.7070705@sun.com> <17c6771e0806241751q25d39598sf08857a9ed83a55e@mail.gmail.com> Message-ID: <1214379323.4472.3.camel@cthalinger.lan> On Wed, 2008-06-25 at 01:51 +0100, Andrew John Hughes wrote: > > ---> sb.setLength(sb.length() - fullPath.length() - 1); > > > > So that calculation is going negative, and inserting println's I see > > sb.length() is 0 ... > > > > I changed that to something like this > > > > sb.setLength( > > (sb.length() - fullPath.length() - 1) > 0 > > ? sb.length() - fullPath.length() - 1 > > : 0 > > ); > > > > > > And then this failed (line 1208) > > nextTest = nextTest.substring(0, nextTest.length() - 5); Actually I changed this code very much. > I believe twisti has done a lot of recent work on the harness, but I > don't think it's committed yet. And today is a good day to finally commit it :-) - twisti From twisti@complang.tuwien.ac.at Wed Jun 25 07:35:00 2008 From: twisti@complang.tuwien.ac.at (Christian Thalinger) Date: Wed, 25 Jun 2008 07:35:00 -0000 Subject: Am I doing something wrong? In-Reply-To: <4861947D.7070705@sun.com> References: <4861947D.7070705@sun.com> Message-ID: <1214379274.4472.1.camel@cthalinger.lan> On Tue, 2008-06-24 at 17:42 -0700, David Herron wrote: > Just to try out Mauve I'm running it on my OS X machine with Apple's Java. > > ./configure --with-vm=/Library/Java/Home/bin/java > make > java Harness -verbose -showpasses -compile -vm java gnu.testlet One thing that's wrong, but I'm not sure this is causing the exception, you have to pass an argument to -compile option: -compile [yes|no]: specify whether or not to compile the tests before running them. This When I try your commandline, it simply test all testlets except the ones I specify. - twisti