Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
[jdt-core-dev] Bug cluster concerning JrtPackageFragmentRoot vs. ClassFile

Hi Team,

We have an ugly cluster of bugs:

http://bugs.eclipse.org/521995 - [9] module:java.base is not on its project's build path
http://bugs.eclipse.org/528058 - [9][model] inconsistency in handle identifiers for JrtPackageFragmentRoot
http://bugs.eclipse.org/519538 - [9] module not on project's build path exception on any type from JDK
http://bugs.eclipse.org/530308 - [9] check usage of ClassFileReader.classFileName

The first two are released for M5/M6, resp., the latter two unresolved.
For those that are already marked as resolved we had back port requests,
but I shied back declaring that we are not on firm ground.

Today I did some more research and believe I resolved the main riddle.

Now, for 4.7.3 this is *very* late, but since I believe this to be a
serious issue, I took the time to write a little essay into bug 530308.
If and only if that essay (together with referenced sources & bugs)
convinces two more engineers then I would in earnest propose the fix
for back porting to 4.7.3. If there's any doubt about the fix, then
please drop a comment in bug 530308, and I will hold my horses :)

Tests in master showed a few failures which I hold to be unrelated.
Gerrit tests in maintenance 4.7 don't use Java 9 :(
Locally RunAllJava9Tests did not show any unexpected failures.


I THINK we _could_ make some clients (in particular code recommenders)
happy, if we fix this for 4.7.3.

Any one?
Stephan


Back to the top