Re: [Pnet-developers] New interface dispatch mechanism

[Gopal V]

If memory serves me right, Rhys Weatherley wrote:
> I have committed a new implementation for "call_interface".  It is based on 
> the Interface Method Table (IMT) mechanism used by IBM's RVM engine for Java. 
>  
Looks like I'm having some problems with that ... running pnet's 
test_verify , at the ldstr test 5 . A segfault which seems to orginate
from support/thread.c.

Starting program: /tmp/dotgnu/pnet/tests/test_verify -f "Constant Instructions"
[New Thread 1024 (LWP 12537)]

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 1024 (LWP 12537)]
ILThreadGetObject (thread=0x81647e0) at thread.c:336
336             return thread->objectArg;
(gdb) bt
#0  ILThreadGetObject (thread=0x81647e0) at thread.c:336
#1  0x0806904e in BuildIMT (info=0x400248bc, classPrivate=0x81eee70)
    at layout.c:465
#2  0x080698c8 in LayoutClass (info=0x400248bc, layout=0xbffff550)
    at layout.c:993
#3  0x08069270 in LayoutClass (info=0x4002654c, layout=0xbffff550)
    at layout.c:617
#4  0x08069a13 in _ILLayoutClass (info=0x4002654c) at layout.c:1003
#5  0x08060bd9 in GetMethodToken (method=0x400191d8, pc=0x40018259 "o\001", 
    callSiteSig=0xbffff624) at verify_call.c:54
#6  0x08064b11 in _ILVerify (coder=0x813e148, start=0xbffff7a8, 
    method=0x400191d8, code=0xbffff7b0, unsafeAllowed=1, thread=0x0)
    at verify.c:1220
#7  0x0804b30f in verify () at eval.c:41

The thread.c mentioned is support/thread.c (I looked about in engine/thread.c
for some time :)

Gopal
-- 
The difference between insanity and genius is measured by success