Fixed a potential bug that could cause memory to be read out-of-bounds.

(Unfortunately, this does not fix the Kyra bug I'm looking for.)

In the most extreme case:

* DR and RR will point to &DR_TABLE[60], and AR will point to &AR_TABLE[60]
* SLOT->KSR will be 0
* CH->kcode will be 15

In that case, it will attempt to access AR[15], RR[15] and DR[15], i.e.
AR_TABLE[75] and DR_TABLE[75]. So these arrays need to be 76 elements, not 75.

We used to initialise element 75, but this was changed to 74 to match the size
of the arrays. Buf if my reasoning is correct, it was the arrays that were too
small.

svn-id: r21301

0 files changed, 0 insertions, 0 deletions