freenode/#clasp - IRC Chatlog
Search
21:40:46
drmeister
(loop repeat 10000000 do (let ((x (core:next-from-fixnum most-positive-fixnum))) (loop repeat 20 do (core:next-mul x x))))
21:51:46
drmeister
It depends on the DEBUG_GUARD_EXHAUSTIVE_VALIDATE being turned on. But that might slow down compilation quite a lot.
22:00:38
drmeister
I pushed a few changes, one to add (gctools:validate-object xxx) and the other to suppress that overload that linux hates.
22:01:11
Bike
" ../../src/gctools/memoryManagement.cc:342 Invalid object with header @ 0x125ea4b40 message: header stamps are invalid"
22:47:08
drmeister
I should write a function like ROOM that walks all of the memory and checks every object
22:52:37
Bike
"../../src/gctools/memoryManagement.cc:342 Invalid object with header @ 0x124bc2ab0 message: bad tail content"
22:54:21
drmeister
For some reason when I recompiled cclasp and bclasp images won't load on linux - so I have to rebuild cclasp
23:39:31
drmeister
So I get... ./../src/gctools/memoryManagement.cc:342 Invalid object with header @ 0x14425820 message: bad tail content
23:45:20
drmeister
From https://github.com/clasp-developers/clasp/blob/bignum/include/clasp/gctools/memoryManagement.h#L487
23:46:01
drmeister
I replicated info in the header incase something gets clobbered I can still reconstruct the info for the object.
23:48:13
drmeister
The _tail_start is 0x60 and the _tail_size is 0x60. The tail_size can vary from object to object and allocation to allocation in a random way to try and avoid bugs with registration and alignment.
23:50:41
drmeister
The header is at 0x14425820 and the client starts at 0x14425860 (vtable-ptr, badge, number of limbs, limbs(2))
23:51:47
drmeister
The tail SHOULD start at 0x14425820+0x60 = 0x14425880 - and it should contain 60 0xcc
23:52:30
drmeister
But the first 8 bytes of the tail contain 0x03ffffffffffffff - so we aren't calculating the size of the object properly when we are allocating it.
23:56:35
Bike
uh.... so what's the fix here? do you mean the wrong size is being passed to TheNextBignum_O::create?
23:59:25
drmeister
I'm not sure. But it looks like a mismatch somewhere between what the allocator things the layout of the object should be and what you are writing into it.
23:59:46
drmeister
I'll have to dig around some more. Right now my girls are dragging me out the door to get some groceries.