| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| ... | |
| |
|
|
| |
default mode (inherited behaviour from JDT).
|
| |
|
|
| |
"declare soft" inside nested aspect does not work
|
| | |
|
| |
|
|
| |
convert Throwable obj to string and ajc aborts
|
| |
|
|
| |
- using extra source locations for recent test case (and retitling)
|
| |
|
|
|
|
|
| |
XLint warning for call PCD's using subtype of defining type
also added extraSourceLocations to IMessage+ for
message with multiple source lines
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
| |
IncompatibleClassChangeError at runtime when compiling with -1.4 option
|
| | |
|
| |
|
|
|
|
|
| |
java.lang.VerifyError in generated code
Bug was that a portion of ReferencePointcut.concretize was not behaving
functionally but was mutating state.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Language regression, or possible language improvement?
The problem was caused by moving name binding in pointcut declarations to
happen before declare parents are evaluated. Because of this, the
compiler doesn't know that ContainerDescriptor isa Key when resolving
the ContainerLoader.containerLoads reference.
The change in ordering was made to fix a bug reported in declare error
and declare soft whose pcds where being evaluated before name binding
had happened in the pointcut declarations. Unfortunately, declare error
and declare soft are concretized at the same time as declare parents
(and all other declares ;-), so this move also led to the regression
noted above.
|
| |
|
|
|
|
|
|
|
|
|
| |
Bugzilla Bug 42740
declare error fails on pointcuts composed from multiple classes
Bugzilla Bug 42746
within() pcd is confused for certain declare softs
and probably:
Bugzilla Bug 42739
Compiler crash in ajc head (post 1.1.1 rc1)
|
| |
|
|
| |
Regression: Cannot compile complex weaving anymore
|
| |
|
|
|
|
|
|
|
|
|
| |
This bug is a reversion from 1.1.0 and so was a high-priority to get fixed
before the 1.1.1rc2 release tommorrow. The bug was caused by a
conflict between the implementation of perthis and a fix made in 1.1.1rc1
to the code for signature matching to handle inter-type declarations
correctly. The fix was the addition of two null pointer checks.
I hope that the original reporter will file a bug report in bugzilla tomorrow
for tracking.
|
| |
|
|
| |
throw derivative pointcuts not advised
|
| |
|
|
| |
percflow aspects compiled from jars share one instance for all entry points
|
| | |
|
| |
|
|
| |
Weaving failure when using injars
|
| |
|
|
|
|
|
|
|
|
| |
binary aspect-declared methods conflict, lost their exception clauses
larger fix to address more issues with aspect-declared methods in bytecode form
also addressed declared exception issue more generally
this fix should be more stable than a smaller hack that just fixed this specific bug
report would have been. added a few tests to match increased scope.
|
| |
|
|
|
|
| |
binary libraries
Corresponding non-binary passing tests added to ajcTests.xml
|
| |
|
|
| |
some should fail) -- confirming on the list.
|
| |
|
|
| |
call pointcut with interface type fails if method declared on interface
|
| |
|
|
|
|
| |
super-qualified pointcut reference cause weaver stack trace
and Bugzilla Bug 40814
no error when defining interface pointcuts
|
| |
|
|
|
|
| |
ClassFormatError on external subtype-qualified ref to supertype pointcut
This was an important bug that was caused by the static fields used in the implementation of cflow being placed on the wrong class. This broke the rules used for name mangling and could occasionally result in name collisions as shown here.
|
| |
|
|
| |
aspects
|
| |
|
|
| |
Class Literals as non final fields
|
| |
|
|
|
| |
convenience method that was removed. Updated tests to enable the
test for the horrible switch bug that this fixes.
|
| |
|
|
|
|
| |
Default method impl for interface causes internal exception.
test submitted by George Harley
|
| |
|
|
| |
Compiler crashes in jar and cflow
|
| |
|
|
|
|
|
|
| |
ajc stack trace on declaring hashcode() method in aspect
added extra error-test for using super inside an inter-type declaration on
an interface with multiple parents -- the correct parent is either
hard or impossible to determine in that case
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
based on patch contributed by Andy Clement
Generalizes the patch with a method org.aspectj.weaver.bcel.Utility.copyInstruction
that works-around the bug in Select.copy(). Changed all calls to
Instruction.copy() to use this new method, would be nice to add the
rule:
* declare error:
* call(* Instruction.copy()) && within(org.aspectj.weaver)
* && !withincode(* Utility.copyInstruction(Instruction)):
* "use Utility.copyInstruction to work-around bug in Select.copy()";
|
| |
|
|
|
| |
[Bug 39458] New: Compiler crash in ajc 1.1
[Bug 39460] New: Missing import crashes compiler
|
| |
|
|
| |
VerifyError, Inconsistent stack height with try/switch/if combination
|
| |
|
|
|
|
| |
can not resolve this member warning
This test currently passes, and therefore must be missing the bug; however,
this is the closest I could come to the bug as reported.
|
| |
|
|
| |
running with -ajdeCompiler
|
| |
|
|
|
| |
insertion of lots of advice code can make branch offset for if too large
test entered as a knownLimitation for 1.1.0
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
| |
that won't work with that wrapper.
|
| |
|
|
| |
Weaving rt.jar results in stack overflow
|
