[#536] SEVG in bignum.c:505... — Sean Chittenden <sean@...>
$ ruby -e 'p [].to_s.strip.to_i'
5 messages
2002/10/13
[#537] darwin shared library patch — Eric Melville <eric@...>
In Darwin, the preferred way to build shared libraries is with two level
9 messages
2002/10/13
[#539] Re: darwin shared library patch
— "Akinori MUSHA" <knu@...>
2002/10/13
Hi,
[#544] Re: darwin shared library patch
— Eric Melville <eric@...>
2002/10/16
> The patch seems to make ENV[]= coredump.
[#546] Re: darwin shared library patch
— "Akinori MUSHA" <knu@...>
2002/10/16
At Thu, 17 Oct 2002 01:41:49 +0900,
[#541] Patch for MacOS X dln.c — Luc B駘anger <belanglu@...>
I have a patch for the dynamic linker in MacOS X, which permit to load
7 messages
2002/10/13
Weak references
From:
Torsten Rer <torsten.rueger@...>
Date:
2002-10-08 08:35:08 UTC
List:
ruby-core #528
Moi,
thanks for all previous answers, it's now working quite fine.
But there is a problem I could solve with what is called a weak
reference in Java.
I store Object references on the C side without the gc knowing about
them. This is how it needs to be, because my cache shouldn't keep them
from being collected. (unless there is another way here ?)
But how can I ensure later that dereferencing my reference won't result
in a core dumped ? Eg, when the memory has been reallocated and my
pointer points somewhere into the middle of an object.
Is there a way to save a unique bitpattern, so when I go back I'd know
it's an object. BTW: It's enough for me to know that a pointer is an
object, I can find out if it's the right object by myself.
Can I get a call from ruby's gc when an object is collected ?
I found gc::is_pointer_to_heap , does that detemine if it's an object,
or just if it sort of could be and given ruby's own usage pattern,
surely is ?
Thanks
Torsten