Re: sigsegv?

[Andrew Haley <aph at redhat dot com>]

Tom Tromey writes:
 > >>>>> "Bryce" == Bryce McKinlay <bryce@mckinlay.net.nz> writes:
 > 
 > Bryce> Was your SEGV inside Method.invoke()? I've noticed strange things can
 > Bryce> happen in there, perhaps if the IP goes to some weird-but-valid
 > Bryce> address before the segv actually occurs.
 > 
 > Yeah, sorry.  Stack trace appended.
 > However, it doesn't look like we're in a particularly weird place:
 > 
 >       objClass = JV_CLASS (obj);
 > 
 > This is in Method::invoke.
 > 
 > Tom
 > 
 > (gdb) bt
 > #0  java::lang::reflect::Method::invoke(java::lang::Object*, JArray<java::lang::Object*>*) (this=0x821efe0, obj=0x0, args=0x0)
 >     at ../../../gcc/libjava/java/lang/reflect/natMethod.cc:164
 > #1  0x080b0336 in gnu.testlet.java.lang.reflect.Method.invoke.try_invoke(gnu.testlet.TestHarness, java.lang.reflect.Method, java.lang.Object, java.lang.Object[], java.lang.Object) (this=0x81f6fc0, harness=0x8215f28, method=0x0, obj=0x0, 
 >     args=0x0, expect=0x821ced0)
 >     at ../mauve/gnu/testlet/java/lang/reflect/Method/invoke.java:71

Uh yes, but the really interesting thing is what happens after this.
Does it abort, or what?  If it aborts, I suspect a bug in
MD_FALLBACK_FRAME_STATE_FOR, or maybe g++ is generating invalid unwind
data.

If _POSIX_VERSION is not defined on your system you won't get to call
unblock_signal().

Andrew.