[PATCH 7/7] jit: fix signal_new_exception_with_cause()

Signed-off-by: Tomek Grabiec --- jit/exception.c |3 +-- 1 files changed, 1 insertions(+), 2 deletions(-) diff --git a/jit/exception.c b/jit/exception.c index b1deb30..c5a7160 100644 --- a/jit/exception.c +++ b/jit/exception.c @@ -122,8 +122,7 @@ int signal_new_exception_with_cause(struct v

[PATCH 5/7] jit: do not duplicate dup_*() arguments before pushing when they are pure.

Signed-off-by: Tomek Grabiec --- jit/ostack-bc.c| 18 test/arch-x86/Makefile |2 + test/jit/bytecode-to-ir-stub.c | 10 test/jit/ostack-bc-test.c | 100 +++- 4 files changed, 48 insertions(+), 82 deletions(-)

[PATCH 6/7] regression: introduce jvm/DupTest.j

Signed-off-by: Tomek Grabiec --- Makefile |1 + regression/jvm/DupTest.j | 294 ++ regression/run-suite.sh |1 + 3 files changed, 296 insertions(+), 0 deletions(-) create mode 100644 regression/jvm/DupTest.j diff --git a/Mak

[PATCH 4/7] jit: make get_pure_expr() global

Signed-off-by: Tomek Grabiec --- include/jit/expression.h |1 + jit/expression.c | 14 ++ jit/object-bc.c | 14 -- 3 files changed, 15 insertions(+), 14 deletions(-) diff --git a/include/jit/expression.h b/include/jit/expression.h index be53649..

[PATCH 3/7] vm: make print_trace() use plain backtrace() when compiled with unit tests

When print_trace() is called from unit tests it cannot use stack walker functionallity because it's stubbed. There is no need to use it anyway because we should not encounter JIT frames there. Solution for this is to use standard backtrace() function for stack trace printing in unit tests. Signed-

[PATCH 1/7] x86: print mixed native and java backtrace.

Backtrace is now printed using the VM stack walker. One advantage of this is that it can recognize JIT, trampoline, VM native and JNI frames. JIT frames have different format than native frames - return address is misplaced so the stamdard backtrace() doesn't work with jit frames involved. Example

[PATCH 2/7] vm: do not print line number in backtrace when undefined

Signed-off-by: Tomek Grabiec --- vm/stack-trace.c |3 +++ 1 files changed, 3 insertions(+), 0 deletions(-) diff --git a/vm/stack-trace.c b/vm/stack-trace.c index 1fe185d..11fc9a7 100644 --- a/vm/stack-trace.c +++ b/vm/stack-trace.c @@ -455,6 +455,9 @@ void print_java_stack_trace_elem(struct

[PATCH 2/2] vm: use Throwable.stackTraceString() for exception printing.

As java/lang/String.substring() is now working so is java/lang/Throwable.stackTraceString(). It allows us to get the full exception descripton with just one method call. Signed-off-by: Tomek Grabiec --- include/vm/preload.h |8 +-- test/vm/preload-stub.c |8 +-- vm/preload.c

[PATCH 1/2] jit: fix signal_new_exception_with_cause()

Some exception classes have dedicated constructors for setting exception's cause. For such classes we shouldn't set the cause with initCause() bacuuse the cause might be stored somewhere else than Throwable.cause. See for example java/lang/ExceptionInInitializerError class. Signed-off-by: Tomek Gr

Re: [PATCH 4/4] vm: use Throwable.stackTraceString() for exception printing.

On Sun, 2009-07-26 at 13:22 +0200, Tomek Grabiec wrote: > Please hold on with merging this one because I noticed that for some > reason "Caused by:" parts are not printed after this patch, which > seems to be a bug in JIT. OK, I reverted the commit from master now. --

Re: [PATCH 4/4] vm: use Throwable.stackTraceString() for exception printing.

2009/7/26 Tomek Grabiec : > As java/lang/String.substring() is now working so is > java/lang/Throwable.stackTraceString(). It allows us to get the full > exception descripton with just one method call. > > Signed-off-by: Tomek Grabiec > --- >  include/vm/preload.h   |    8 +-- >  test/vm/preload-s

[PATCH 4/4] vm: use Throwable.stackTraceString() for exception printing.

As java/lang/String.substring() is now working so is java/lang/Throwable.stackTraceString(). It allows us to get the full exception descripton with just one method call. Signed-off-by: Tomek Grabiec --- include/vm/preload.h |8 +-- test/vm/preload-stub.c |8 +-- vm/preload.c

[PATCH 3/4] vm: do not convert special characters to <%d> in string_to_cstr()

This leads to a buglet that new line character '\n' is converted to <10> string, so when converting multiple lined strings to C strings we loose line breaks. Same applies for other control characters. Signed-off-by: Tomek Grabiec --- vm/object.c |2 +- 1 files changed, 1 insertions(+), 1 del

[PATCH 2/4] vm: add support for -Dproperty=value options

This can be used to set/override system properties. Signed-off-by: Tomek Grabiec --- vm/jato.c | 141 1 files changed, 122 insertions(+), 19 deletions(-) diff --git a/vm/jato.c b/vm/jato.c index b7f7e3d..f75ca00 100644 --- a/vm/jato.

[PATCH 1/4] vm: do not print error when dlopen() fails in vm_jni_load_object()

Before library is loaded classpath tries to load it from multiple paths. We shouldn't print eny errors if some of them fail. Signed-off-by: Tomek Grabiec --- vm/jni.c |4 +--- 1 files changed, 1 insertions(+), 3 deletions(-) diff --git a/vm/jni.c b/vm/jni.c index 8c80afe..9af45cb 100644 ---

Re: [PATCH 01/22] x86-64: make native_call() a stub

The series has been applied, thanks! -- ___ Jatovm-devel mailing list Jatovm-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jatovm-devel