Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
RE: [aspectj-dev] 1.1rc2 **Preview Build** pre-install results on JDK 1.4

Here's an evaluation of the failures you saw:

The JDK13 failure has to do with differences between how the IBM JDK and the SUN JDK interpret the input to Class.forName for array types.  Your ability to find this bug is one of the nice benefits of a diverse development team.  I believe that I've fixed this in the tree (and I added a more obvious test case in case this shows up in the future); however, I haven't checked the fix on the IBM JDK because I don't want to fill out all the forms that seem to be required to download it.  Let me know if the fix works for you when you next do a test run.

<details>The definition of Class.forName doesn't specify what string form should be used for the names of array types.  The obvious java syntax of "java.lang.String[]" doesn't work in any implementations.  Many implementations will accept the bytecode form of the name, "[Ljava/lang/String;", but not all.  The eclipse compiler uses a hybrid form for resolving String[].class literals, i.e. "Ljava.lang.String;".  This appears to be the most portable representation, and I've changed the aspectj reflection code to use this.</details>

You aren't noticing any JDK14 failures, and I've fixed the nit-picky source location error that only shows up on SUN's 14 VM.

The emacssym failures are caused by the code just mailed in from Mik and committed by Wes to support incremental compilation in ajde.  One of them needs to resolve these bugs somehow.

The failing tests are failing as expected.  I'll take a pass at fixing them tonight, but if they prove hard to crack they're acceptable knownLimitations for the 1.1 release.

-Jim

> -----Original Message-----
> From: George Harley1 [mailto:GHARLEY@xxxxxxxxxx]
> Sent: Wednesday, April 30, 2003 2:08 PM
> To: aspectj-dev@xxxxxxxxxxx
> Subject: [aspectj-dev] 1.1rc2 **Preview Build** pre-install results on JDK
> 1.4
> 
> Hi,
> 
> Using a Sun 1.4.0 JDK I have run the JUnit tests and
> org.aspectj.testing.Harness tests against the initial/preview rc2 build
> that got published earlier. The results are now in the
> org.aspectj/releases/aspectj-1.1rc2 folder in CVS. The changes are ....
> 
> junitModules.report.JDK14.zip
> ......................................................... new file.
> Results
> better than on 1.3 run conducted earlier.
> ajcTests-result-1.1rc2-JDK14.txt
> ...................................................... updated file.
> Results of Harness with ajcTests.xml
> ajcTests-result-1.1rc2-emacssym-JDK14.txt ...............................
> updated file. Results of Harness with ajcTests.xml and -emacssym switch.
> ajcTestsFailing-result-1.1rc2-JDK14.txt
> ..........................................new file. Results of Harness
> with
> ajcTestsFailing.xml
> ajcTestsFailing-result-1.1rc2-emacssym-JDK14.txt ...................new
> file. Results of Harness with ajcTestsFailing.xml and -emacssym switch
> 
> Best regards,
> George
> ________________________________________
> George C. Harley
> IBM e-Business Integrated Technologies
> MP 146, Hursley Park, Winchester, Hants, UK
> External : +44 (0)1962 819549 / Internal : 7-249549
> Fax : +44 (0)1962 818999
> Email : gharley@xxxxxxxxxx
> Web : http://ibm.com/java
> 
> 
> _______________________________________________
> aspectj-dev mailing list
> aspectj-dev@xxxxxxxxxxx
> http://dev.eclipse.org/mailman/listinfo/aspectj-dev


Back to the top