| Re: [aspectj-dev] AspectJ 1.1rc1 Package-and-Test dry-run report |
I tested the rc1 jar against the project code that I'm working on and have hit what looks like a priority 1 bug. I am working on narrowing it down now...
Exception thrown from AspectJ 1.1rc1
This might be logged as a bug already -- find current bugs at
http://bugs.eclipse.org/bugs/buglist.cgi?product=AspectJ&component=Compiler
Bugs for exceptions thrown have titles File:line from the top stack,
e.g., "SomeFile.java:243"
If you don't find the exception below in a bug, please add a new bug
at http://bugs.eclipse.org/bugs/enter_bug.cgi?product=AspectJ
To make the bug a priority, please include a test program
that can reproduce this exception.
0
0
java.lang.ArrayIndexOutOfBoundsException: 0
at org.aspectj.weaver.IntMap.get(IntMap.java:109)
at org.aspectj.weaver.patterns.CflowPointcut.concretize1(CflowPointcut.j
ava:170)
at org.aspectj.weaver.patterns.ReferencePointcut.concretize1(ReferencePo
intcut.java:270)
at org.aspectj.weaver.patterns.OrPointcut.concretize1(OrPointcut.java:91
)
at org.aspectj.weaver.patterns.AndPointcut.concretize1(AndPointcut.java:
88)
at org.aspectj.weaver.patterns.Pointcut.concretize(Pointcut.java:127)
at org.aspectj.weaver.patterns.Pointcut.concretize(Pointcut.java:120)
at org.aspectj.weaver.Advice.concretize(Advice.java:207)
at org.aspectj.weaver.CrosscuttingMembers.addShadowMunger(CrosscuttingMe
mbers.java:78)
at org.aspectj.weaver.CrosscuttingMembers.addShadowMungers(CrosscuttingM
embers.java:72)
at org.aspectj.weaver.ResolvedTypeX.collectCrosscuttingMembers(ResolvedT
ypeX.java:328)
at org.aspectj.weaver.CrosscuttingMembersSet.addOrReplaceAspect(Crosscut
tingMembersSet.java:54)
at org.aspectj.weaver.bcel.BcelWeaver.prepareForWeave(BcelWeaver.java:16
4)
at org.aspectj.weaver.bcel.BcelWeaver.weave(BcelWeaver.java:214)
at org.aspectj.ajdt.internal.core.builder.AjBuildManager.weaveAndGenerat
eClassFiles(AjBuildManager.java:394)
at org.aspectj.ajdt.internal.core.builder.AjBuildManager.batchBuild(AjBu
ildManager.java:149)
at org.aspectj.ajdt.ajc.AjdtCommand.runCommand(AjdtCommand.java:55)
at org.aspectj.tools.ajc.Main.run(Main.java:216)
at org.aspectj.tools.ajc.Main.runMain(Main.java:155)
at org.aspectj.tools.ajc.Main.main(Main.java:72)
1 fail|abort
Adrian Colyer <adrian_colyer@xxxxxxxxxx> wrote:
Below you'll find the full notes for the dry-run package and test of
aspectj-1.1rc1. This file and all the records it refers to are stored in
CVS under org.aspectj/releases/aspectj-1.1rc1. We'll update those files
tomorrow as we do the final packaging.
Key things to note:
- we had a few problems running test suites against a non DEVELOPMENT
version jar - which we were able to fix
- in addition to the one failure we saw previously in ajcTests.xml (which
was reproducible on a clean machine), we
saw one extra failure that only occurs under a 1.3 JDK (the test passes
with a 1.4 JDK).
- README-11.html appears to be manually updated, and needs updating for
1.1rc1 (still refers to 1.1beta5)
Please take the aspectj-1.1rc1.jar for a spin, and verify from our log
below that we have run all important tests
(we'll use this script as the basis for tomorrow's packaging).
We built an AJDT distribution off this level as well, which you can obtain
from the update site.
Almost there now!!
Regards,
Adrian.
AspectJ 1.1rc1 Package and Test Notes
=======================================
Initial package and test run: 12th March 2003, Adrian Colyer & George
Harley
Environment:
------------
Eclipse 2.1 RC1 on Windows
Sun JDK 1.3.1.07
Full CVS extract of org.aspectj/modules as of 12th Mar 2003, 9:00am GMT
Log:
----
* ran ant on build/build.xml using libraries in lib/ant/lib and
lib/junit/lib in place of
default eclipse libraries
* ran junitModules.xml tests
- 2 failures : taskdefs nofile and testIncremental (expected)
* ran ajcTests.xml with and without emacssym
- 2 failures (documented later)
This created a benchmark of test results before changing any files.
Then...
* edited buildProperties.xml in build project
- called release 1.1rc1, long name "1.1 release candidate 1"
* edited org.aspectj.ajdt.ajc/messages.properties in org.aspectj.ajdt.core
project
- version = AspectJ Compiler 1.1rc1
- misc.usage = AspectJ Compiler 1.1 release candidate 1\n ..."
* from build/build,xml :
- ant clean
- ant
- ant clean
- ant
(run twice to ensure picked up any changes in build module - was this
strictly necessary?)
- saved build output to build1.1rc1.log
- saved built jar as aspectj-1.1rc1.jar
* ran junitModules.xml
- multiple failures due to checking for aspectjrt.jar file and
version no in manifest
(this test is only activated when Version != DEVELOPMENT)
- extended junitModules.xml target run-module-junit-tests to include
extra element in junit task:
this sets the JVM property "aspectjrt.path" to the value of the Ant
property "${aspectjrt.path}"
- extended NullIdeProperties.getClasspath() in
ajde/testsrc/org.aspectj.ajde to add value
of system property "aspectjrt.path" onto end of classpath. (This has
no effect when property is
not defined.
- rerun junitModules.xml specifying property aspectjrt.path
=../aj-build/dist/lib/aspectjrt.jar
(use properties tab in eclipse AntRunner, or -D option from ant
command-line
- now failing "testVersion" in BridgeModuleTests - was comparing time
from Version.time_text
against formatted version of time in Version.time. BUT, test used
simple date formatted (default
locale and timezone - GMT for us) and pre-formatted string used PST
timezone. Fixed testcase to
use PST timezone when formatting.
- junitModules.xml now runs "clean", test results saved in
junitModules.report.zip
: the 2 expected failures remain
(TaskdefsModuleTests:testNoFile and
TestingDriversModuleTests.testIncrementalSuite)
- saved report output in jnuitModules.report.zip
* ran harness -release -logFail ajcTests.xml , saved results to
ajcTests-result-1.1rc1.txt
- **2 failures**
(1) source locations within expressions [anticipated, as discussed
on aspectj-dev]
(2) assert and pertarget crashes compiler - this test passes on a 1.4
JDK, but fails under 1.3.1
as used for our testing (see Environment section)
* ran harness -release -logFail -emacssym ajcTests.xml , saved results to
ajcTests-result-1.1rc1-emacssym.txt
- same 2 failures as above.
* ran harness -release -logFail ajcTestsFailing.xl, saved results to
ajcTestsFailing-result-1.1rc1.txt
- all 5 fail (expected?? - but deviation from release notes in
readme-build-and-test-aspectj.html
* ran harness -release -logFail -emacssym ajcTestsFailing.xl, saved results
to ajcTestsFailing-result-1.1rc1-emacssym.txt
- all 5 fail (expected?? - but deviation from release notes in
readme-build-and-test-aspectj.html
Installed from aspectj-1.1rc1.jar
----------------------------------
* splash screen shows "Version 1.1 release candidate 1 built on Wednesday
Mar 12, 2003"
- see Installer-Splash-Screen.bmp
* installed OK
* checked README-AspectJ.html -> version OK
* checked README-11.html -> **NOT OK** - shows 1.1beta5 as version
* checked license - OK
* ajc -version => AspectJ Compiler 1.1rc1
* ajc => "AspectJ Compiler 1.1 release candidate 1..."
* from doc/examples run "ant" - spacewar builds and runs ok
* ran ant targets as follows:
- bean OK
- intro OK
- intro-clone OK
- intro-compare - OK
- intro-hash OK
- observer - OK
- spacewar - OK
- spacewar-demo **NOT OK** fails with NoClassDef found (see
spacewar-demo.log)
target to run was "spaceware.game" instead of
spacewar.game - now fixed.
- telecom - OK
- telecom-billing - OK
- telecom-timing -OK
- tjp - OK
- tracing-1 - OK
- tracing-2 - OK
- tracing-3 - OK
- tracing-bc - OK
- tracing-none - OK
* ran compiled spacewar on JVM 1.1.8 - OK
* verfified past failures:
- debug info included in .class files (used class file inspector on
aspectjtools.jar)
- ant-tasks.html has correct date and version info
- manual dates changed in docs. **but NOT README-11.html**
- runtime and tools version info consistent (tested by spacewar)
* bugs closed with this release but no automated tests (yet)
- 33474 bad source locations for declare error/warning : verfied from
AJDT - OK
* other release objectives:
- built new release of AJDT - OK
-- Adrian
Adrian_Colyer@xxxxxxxxxx
_______________________________________________
aspectj-dev mailing list
aspectj-dev@xxxxxxxxxxx
http://dev.eclipse.org/mailman/listinfo/aspectj-dev