Package: xmille
Version: 2.0-13
Severity: important
I've had loads of crashes with xmille lately on i386, but not on
amd64. Finally it annoyed enough to take a look. The problem was
that onecard was doing Numseen[safety(bat)] with bat == C_INIT,
which safety did not handle.
This is easily fixed by making safety treat C_INIT the same way
as if STOP had been played. I've checked the other callers of
safety and none of them will use the C_INIT value so this is a
noop for them.
I've also found onecard's logic to be broken where
1) It tries to check S_RIGHT_WAY in Numseen, when Numseen is
indexed by card number, so it was checking C_100 instead.
2) It checks Numseen != 0, where it should really be Numseen == 0
since we're trying to determine whether the human can play a
safety followed immediately by another card to finish the game.
Attached patch should fix all these issues:
diff --git a/comp.c b/comp.c
index 3d98ee4..dbc1f27 100644
--- a/comp.c
+++ b/comp.c
@@ -327,8 +327,8 @@ reg PLAY *pp; {
spd = pp->speed;
card = -1;
if (pp->can_go || ((isrepair(bat) || bat == C_STOP
- || spd == C_LIMIT) && Numseen[S_RIGHT_WAY] != 0)
- || Numseen[safety(bat)] != 0)
+ || spd == C_LIMIT) && Numseen[C_RIGHT_WAY] == 0)
+ || Numseen[safety(bat)] == 0)
switch (End - pp->mileage) {
case 200:
if (pp->nummiles[C_200] == 2)
diff --git a/types.c b/types.c
index 13a8f97..d1149ab 100644
--- a/types.c
+++ b/types.c
@@ -31,6 +31,7 @@ reg CARD card; {
case C_RIGHT_WAY:
case C_LIMIT:
case C_END_LIMIT:
+ case C_INIT:
return C_RIGHT_WAY;
}
/* NOTREACHED */
-- System Information
Debian Release: lenny/sid
Kernel Version: Linux gondolin 2.6.17-rc4 #1 SMP PREEMPT Wed May 17 17:28:00
EST 2006 i686 GNU/Linux
Versions of the packages xmille depends on:
ii libc6 2.7-16
GNU C Library: Shared libraries
ii libx11-6 2:1.1.5-2
X11 client-side library
ii libxext6 2:1.0.4-1
X11 miscellaneous extension library
--
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]
