On 26/10/2012, at 4:04 AM, Tim Margheim wrote:
> On 10/24/2012 3:59 PM, john skaller wrote:
>> On 25/10/2012, at 2:35 AM, Tim Margheim wrote:
>>> Thanks. John's answer fixes the crashing loop. I still have the access
>>> violation in 64-bit mode when populating the tree, though.
>> Repost your code with the fix.
>> Do you know which call causes the access violation, i.e. exactly where?
> I didn't repost it because I hadn't changed that code--the violation was
> occurring in the first loop (which populates the tree), and I had only
> changed the third loop. It was a "Attempted to read or write protected
> memory" error, and it was occurring somewhere in JudySLIns.
>
> However, I fixed it! (I got a hint from a compiler warning I'd missed.) I
> tracked it down to this in JudyPrivate.h:
> #define cJU_ALLONES (~0UL)
> which I had to change to
> #define cJU_ALLONES (~0ULL)
>
> That fixed the access violation.
~(Word_t)0
would always work (assuming Word_t is defined).
Also plain old
(Word_t)(-1)
would also work on 2-compilement machines (which is all of them these days).
>
> It turns out that the Judy code uses some literal values with a UL suffix,
> e.g. "0x100UL". And in WIN64, that's only 32 bits--which can lead to
> shenanigans, since the code expects it to be 64 bits when in 64-bit mode. So
> I need to do "0x100ULL", or perhaps "(Word_t)0x100". Though a run-time
> typecast would add some instructions (right?),
Compiler dependent, but generally it shouldn't: constant folding in
the compiler should handle it.
--
john skaller
[email protected]
http://felix-lang.org
------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_sfd2d_oct
_______________________________________________
Judy-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/judy-devel
