- add POPT_ARG_BITSET handling.

8 files changed, 1196 insertions, 23 deletions

diff --git a/CHANGES b/CHANGES
index a0025e2..1a0ae46 100644
--- a/CHANGES
+++ b/CHANGES
@@ -1,4 +1,5 @@
 1.15 -> 1.16:
+    - add POPT_ARG_BITSET handling.
     - add POPT_ARG_SHORT handling.
     - handle all callback traversals within a C switch (for extendability ease).
     - add popt.pc.
diff --git a/lookup3.c b/lookup3.c
new file mode 100644
index 0000000..0584c39
--- /dev/null
+++ b/lookup3.c
@@ -0,0 +1,969 @@
+/* -------------------------------------------------------------------- */
+/*
+ * lookup3.c, by Bob Jenkins, May 2006, Public Domain.
+ * 
+ * These are functions for producing 32-bit hashes for hash table lookup.
+ * jlu32w(), jlu32l(), jlu32lpair(), jlu32b(), _JLU3_MIX(), and _JLU3_FINAL() 
+ * are externally useful functions.  Routines to test the hash are included 
+ * if SELF_TEST is defined.  You can use this free for any purpose.  It's in
+ * the public domain.  It has no warranty.
+ * 
+ * You probably want to use jlu32l().  jlu32l() and jlu32b()
+ * hash byte arrays.  jlu32l() is is faster than jlu32b() on
+ * little-endian machines.  Intel and AMD are little-endian machines.
+ * On second thought, you probably want jlu32lpair(), which is identical to
+ * jlu32l() except it returns two 32-bit hashes for the price of one.  
+ * You could implement jlu32bpair() if you wanted but I haven't bothered here.
+ * 
+ * If you want to find a hash of, say, exactly 7 integers, do
+ *   a = i1;  b = i2;  c = i3;
+ *   _JLU3_MIX(a,b,c);
+ *   a += i4; b += i5; c += i6;
+ *   _JLU3_MIX(a,b,c);
+ *   a += i7;
+ *   _JLU3_FINAL(a,b,c);
+ * then use c as the hash value.  If you have a variable size array of
+ * 4-byte integers to hash, use jlu32w().  If you have a byte array (like
+ * a character string), use jlu32l().  If you have several byte arrays, or
+ * a mix of things, see the comments above jlu32l().  
+ * 
+ * Why is this so big?  I read 12 bytes at a time into 3 4-byte integers, 
+ * then mix those integers.  This is fast (you can do a lot more thorough
+ * mixing with 12*3 instructions on 3 integers than you can with 3 instructions
+ * on 1 byte), but shoehorning those bytes into integers efficiently is messy.
+*/
+/* -------------------------------------------------------------------- */
+
+#include <stdint.h>
+
+#if defined(_JLU3_SELFTEST)
+# define _JLU3_jlu32w		1
+# define _JLU3_jlu32l		1
+# define _JLU3_jlu32lpair	1
+# define _JLU3_jlu32b		1
+#endif
+
+/*@-redef@*/
+/*@unchecked@*/
+static const union _dbswap {
+    const uint32_t ui;
+    const unsigned char uc[4];
+} endian = { .ui = 0x11223344 };
+# define HASH_LITTLE_ENDIAN	(endian.uc[0] == (unsigned char) 0x44)
+# define HASH_BIG_ENDIAN	(endian.uc[0] == (unsigned char) 0x11)
+/*@=redef@*/
+
+#ifndef ROTL32
+# define ROTL32(x, s) (((x) << (s)) | ((x) >> (32 - (s))))
+#endif
+
+/* NOTE: The _size parameter should be in bytes. */
+#define	_JLU3_INIT(_h, _size)	(0xdeadbeef + ((uint32_t)(_size)) + (_h))
+
+/* -------------------------------------------------------------------- */
+/*
+ * _JLU3_MIX -- mix 3 32-bit values reversibly.
+ * 
+ * This is reversible, so any information in (a,b,c) before _JLU3_MIX() is
+ * still in (a,b,c) after _JLU3_MIX().
+ * 
+ * If four pairs of (a,b,c) inputs are run through _JLU3_MIX(), or through
+ * _JLU3_MIX() in reverse, there are at least 32 bits of the output that
+ * are sometimes the same for one pair and different for another pair.
+ * This was tested for:
+ * * pairs that differed by one bit, by two bits, in any combination
+ *   of top bits of (a,b,c), or in any combination of bottom bits of
+ *   (a,b,c).
+ * * "differ" is defined as +, -, ^, or ~^.  For + and -, I transformed
+ *   the output delta to a Gray code (a^(a>>1)) so a string of 1's (as
+ *   is commonly produced by subtraction) look like a single 1-bit
+ *   difference.
+ * * the base values were pseudorandom, all zero but one bit set, or 
+ *   all zero plus a counter that starts at zero.
+ * 
+ * Some k values for my "a-=c; a^=ROTL32(c,k); c+=b;" arrangement that
+ * satisfy this are
+ *     4  6  8 16 19  4
+ *     9 15  3 18 27 15
+ *    14  9  3  7 17  3
+ * Well, "9 15 3 18 27 15" didn't quite get 32 bits diffing
+ * for "differ" defined as + with a one-bit base and a two-bit delta.  I
+ * used http://burtleburtle.net/bob/hash/avalanche.html to choose 
+ * the operations, constants, and arrangements of the variables.
+ * 
+ * This does not achieve avalanche.  There are input bits of (a,b,c)
+ * that fail to affect some output bits of (a,b,c), especially of a.  The
+ * most thoroughly mixed value is c, but it doesn't really even achieve
+ * avalanche in c.
+ * 
+ * This allows some parallelism.  Read-after-writes are good at doubling
+ * the number of bits affected, so the goal of mixing pulls in the opposite
+ * direction as the goal of parallelism.  I did what I could.  Rotates
+ * seem to cost as much as shifts on every machine I could lay my hands
+ * on, and rotates are much kinder to the top and bottom bits, so I used
+ * rotates.
+ */
+/* -------------------------------------------------------------------- */
+#define _JLU3_MIX(a,b,c) \
+{ \
+  a -= c;  a ^= ROTL32(c, 4);  c += b; \
+  b -= a;  b ^= ROTL32(a, 6);  a += c; \
+  c -= b;  c ^= ROTL32(b, 8);  b += a; \
+  a -= c;  a ^= ROTL32(c,16);  c += b; \
+  b -= a;  b ^= ROTL32(a,19);  a += c; \
+  c -= b;  c ^= ROTL32(b, 4);  b += a; \
+}
+
+/* -------------------------------------------------------------------- */
+/**
+ * _JLU3_FINAL -- final mixing of 3 32-bit values (a,b,c) into c
+ * 
+ * Pairs of (a,b,c) values differing in only a few bits will usually
+ * produce values of c that look totally different.  This was tested for
+ * * pairs that differed by one bit, by two bits, in any combination
+ *   of top bits of (a,b,c), or in any combination of bottom bits of
+ *   (a,b,c).
+ * * "differ" is defined as +, -, ^, or ~^.  For + and -, I transformed
+ *   the output delta to a Gray code (a^(a>>1)) so a string of 1's (as
+ *   is commonly produced by subtraction) look like a single 1-bit
+ *   difference.
+ * * the base values were pseudorandom, all zero but one bit set, or 
+ *   all zero plus a counter that starts at zero.
+ * 
+ * These constants passed:
+ *  14 11 25 16 4 14 24
+ *  12 14 25 16 4 14 24
+ * and these came close:
+ *   4  8 15 26 3 22 24
+ *  10  8 15 26 3 22 24
+ *  11  8 15 26 3 22 24
+ */
+/* -------------------------------------------------------------------- */
+#define _JLU3_FINAL(a,b,c) \
+{ \
+  c ^= b; c -= ROTL32(b,14); \
+  a ^= c; a -= ROTL32(c,11); \
+  b ^= a; b -= ROTL32(a,25); \
+  c ^= b; c -= ROTL32(b,16); \
+  a ^= c; a -= ROTL32(c,4);  \
+  b ^= a; b -= ROTL32(a,14); \
+  c ^= b; c -= ROTL32(b,24); \
+}
+
+#if defined(_JLU3_jlu32w)
+uint32_t jlu32w(uint32_t h, /*@null@*/ const uint32_t *k, size_t size)
+	/*@*/;
+/* -------------------------------------------------------------------- */
+/**
+ *  This works on all machines.  To be useful, it requires
+ *  -- that the key be an array of uint32_t's, and
+ *  -- that the size be the number of uint32_t's in the key
+ * 
+ *  The function jlu32w() is identical to jlu32l() on little-endian
+ *  machines, and identical to jlu32b() on big-endian machines,
+ *  except that the size has to be measured in uint32_ts rather than in
+ *  bytes.  jlu32l() is more complicated than jlu32w() only because
+ *  jlu32l() has to dance around fitting the key bytes into registers.
+ *
+ * @param h		the previous hash, or an arbitrary value
+ * @param *k		the key, an array of uint32_t values
+ * @param size		the size of the key, in uint32_ts
+ * @return		the lookup3 hash
+ */
+/* -------------------------------------------------------------------- */
+uint32_t jlu32w(uint32_t h, const uint32_t *k, size_t size)
+{
+    uint32_t a = _JLU3_INIT(h, (size * sizeof(*k)));
+    uint32_t b = a;
+    uint32_t c = a;
+
+    if (k == NULL)
+	goto exit;
+
+    /*----------------------------------------------- handle most of the key */
+    while (size > 3) {
+	a += k[0];
+	b += k[1];
+	c += k[2];
+	_JLU3_MIX(a,b,c);
+	size -= 3;
+	k += 3;
+    }
+
+    /*----------------------------------------- handle the last 3 uint32_t's */
+    switch (size) {
+    case 3 : c+=k[2];
+    case 2 : b+=k[1];
+    case 1 : a+=k[0];
+	_JLU3_FINAL(a,b,c);
+	/*@fallthrough@*/
+    case 0:
+	break;
+    }
+    /*---------------------------------------------------- report the result */
+exit:
+    return c;
+}
+#endif	/* defined(_JLU3_jlu32w) */
+
+#if defined(_JLU3_jlu32l)
+uint32_t jlu32l(uint32_t h, const void *key, size_t size)
+	/*@*/;
+/* -------------------------------------------------------------------- */
+/*
+ * jlu32l() -- hash a variable-length key into a 32-bit value
+ *   h       : can be any 4-byte value
+ *   k       : the key (the unaligned variable-length array of bytes)
+ *   size    : the size of the key, counting by bytes
+ * Returns a 32-bit value.  Every bit of the key affects every bit of
+ * the return value.  Two keys differing by one or two bits will have
+ * totally different hash values.
+ * 
+ * The best hash table sizes are powers of 2.  There is no need to do
+ * mod a prime (mod is sooo slow!).  If you need less than 32 bits,
+ * use a bitmask.  For example, if you need only 10 bits, do
+ *   h = (h & hashmask(10));
+ * In which case, the hash table should have hashsize(10) elements.
+ * 
+ * If you are hashing n strings (uint8_t **)k, do it like this:
+ *   for (i=0, h=0; i<n; ++i) h = jlu32l(h, k[i], len[i]);
+ * 
+ * By Bob Jenkins, 2006.  bob_jenkins@burtleburtle.net.  You may use this
+ * code any way you wish, private, educational, or commercial.  It's free.
+ * 
+ * Use for hash table lookup, or anything where one collision in 2^^32 is
+ * acceptable.  Do NOT use for cryptographic purposes.
+ *
+ * @param h		the previous hash, or an arbitrary value
+ * @param *k		the key, an array of uint8_t values
+ * @param size		the size of the key
+ * @return		the lookup3 hash
+ */
+/* -------------------------------------------------------------------- */
+uint32_t jlu32l(uint32_t h, const void *key, size_t size)
+{
+    union { const void *ptr; size_t i; } u;
+    uint32_t a = _JLU3_INIT(h, size);
+    uint32_t b = a;
+    uint32_t c = a;
+
+    if (key == NULL)
+	goto exit;
+
+    u.ptr = key;
+    if (HASH_LITTLE_ENDIAN && ((u.i & 0x3) == 0)) {
+	const uint32_t *k = (const uint32_t *)key;	/* read 32-bit chunks */
+#ifdef	VALGRIND
+	const uint8_t  *k8;
+#endif
+
+    /*------ all but last block: aligned reads and affect 32 bits of (a,b,c) */
+	while (size > 12) {
+	    a += k[0];
+	    b += k[1];
+	    c += k[2];
+	    _JLU3_MIX(a,b,c);
+	    size -= 12;
+	    k += 3;
+	}
+
+	/*------------------------- handle the last (probably partial) block */
+	/* 
+	 * "k[2]&0xffffff" actually reads beyond the end of the string, but
+	 * then masks off the part it's not allowed to read.  Because the
+	 * string is aligned, the masked-off tail is in the same word as the
+	 * rest of the string.  Every machine with memory protection I've seen
+	 * does it on word boundaries, so is OK with this.  But VALGRIND will
+	 * still catch it and complain.  The masking trick does make the hash
+	 * noticably faster for short strings (like English words).
+	 */
+#ifndef VALGRIND
+
+	switch (size) {
+	case 12:	c += k[2]; b+=k[1]; a+=k[0]; break;
+	case 11:	c += k[2]&0xffffff; b+=k[1]; a+=k[0]; break;
+	case 10:	c += k[2]&0xffff; b+=k[1]; a+=k[0]; break;
+	case  9:	c += k[2]&0xff; b+=k[1]; a+=k[0]; break;
+	case  8:	b += k[1]; a+=k[0]; break;
+	case  7:	b += k[1]&0xffffff; a+=k[0]; break;
+	case  6:	b += k[1]&0xffff; a+=k[0]; break;
+	case  5:	b += k[1]&0xff; a+=k[0]; break;
+	case  4:	a += k[0]; break;
+	case  3:	a += k[0]&0xffffff; break;
+	case  2:	a += k[0]&0xffff; break;
+	case  1:	a += k[0]&0xff; break;
+	case  0:	goto exit;
+	}
+
+#else /* make valgrind happy */
+
+	k8 = (const uint8_t *)k;
+	switch (size) {
+	case 12:	c += k[2]; b+=k[1]; a+=k[0]	break;
+	case 11:	c += ((uint32_t)k8[10])<<16;	/*@fallthrough@*/
+	case 10:	c += ((uint32_t)k8[9])<<8;	/*@fallthrough@*/
+	case  9:	c += k8[8];			/*@fallthrough@*/
+	case  8:	b += k[1]; a+=k[0];		break;
+	case  7:	b += ((uint32_t)k8[6])<<16;	/*@fallthrough@*/
+	case  6:	b += ((uint32_t)k8[5])<<8;	/*@fallthrough@*/
+	case  5:	b += k8[4];			/*@fallthrough@*/
+	case  4:	a += k[0];			break;
+	case  3:	a += ((uint32_t)k8[2])<<16;	/*@fallthrough@*/
+	case  2:	a += ((uint32_t)k8[1])<<8;	/*@fallthrough@*/
+	case  1:	a += k8[0];			break;
+	case  0:	goto exit;
+	}
+
+#endif /* !valgrind */
+
+    } else if (HASH_LITTLE_ENDIAN && ((u.i & 0x1) == 0)) {
+	const uint16_t *k = (const uint16_t *)key;	/* read 16-bit chunks */
+	const uint8_t  *k8;
+
+	/*----------- all but last block: aligned reads and different mixing */
+	while (size > 12) {
+	    a += k[0] + (((uint32_t)k[1])<<16);
+	    b += k[2] + (((uint32_t)k[3])<<16);
+	    c += k[4] + (((uint32_t)k[5])<<16);
+	    _JLU3_MIX(a,b,c);
+	    size -= 12;
+	    k += 6;
+	}
+
+	/*------------------------- handle the last (probably partial) block */
+	k8 = (const uint8_t *)k;
+	switch (size) {
+	case 12:
+	    c += k[4]+(((uint32_t)k[5])<<16);
+	    b += k[2]+(((uint32_t)k[3])<<16);
+	    a += k[0]+(((uint32_t)k[1])<<16);
+	    break;
+	case 11:
+	    c += ((uint32_t)k8[10])<<16;
+	    /*@fallthrough@*/
+	case 10:
+	    c += (uint32_t)k[4];
+	    b += k[2]+(((uint32_t)k[3])<<16);
+	    a += k[0]+(((uint32_t)k[1])<<16);
+	    break;
+	case  9:
+	    c += (uint32_t)k8[8];
+	    /*@fallthrough@*/
+	case  8:
+	    b += k[2]+(((uint32_t)k[3])<<16);
+	    a += k[0]+(((uint32_t)k[1])<<16);
+	    break;
+	case  7:
+	    b += ((uint32_t)k8[6])<<16;
+	    /*@fallthrough@*/
+	case  6:
+	    b += (uint32_t)k[2];
+	    a += k[0]+(((uint32_t)k[1])<<16);
+	    break;
+	case  5:
+	    b += (uint32_t)k8[4];
+	    /*@fallthrough@*/
+	case  4:
+	    a += k[0]+(((uint32_t)k[1])<<16);
+	    break;
+	case  3:
+	    a += ((uint32_t)k8[2])<<16;
+	    /*@fallthrough@*/
+	case  2:
+	    a += (uint32_t)k[0];
+	    break;
+	case  1:
+	    a += (uint32_t)k8[0];
+	    break;
+	case  0:
+	    goto exit;
+	}
+
+    } else {		/* need to read the key one byte at a time */
+	const uint8_t *k = (const uint8_t *)key;
+
+	/*----------- all but the last block: affect some 32 bits of (a,b,c) */
+	while (size > 12) {
+	    a += (uint32_t)k[0];
+	    a += ((uint32_t)k[1])<<8;
+	    a += ((uint32_t)k[2])<<16;
+	    a += ((uint32_t)k[3])<<24;
+	    b += (uint32_t)k[4];
+	    b += ((uint32_t)k[5])<<8;
+	    b += ((uint32_t)k[6])<<16;
+	    b += ((uint32_t)k[7])<<24;
+	    c += (uint32_t)k[8];
+	    c += ((uint32_t)k[9])<<8;
+	    c += ((uint32_t)k[10])<<16;
+	    c += ((uint32_t)k[11])<<24;
+	    _JLU3_MIX(a,b,c);
+	    size -= 12;
+	    k += 12;
+	}
+
+	/*---------------------------- last block: affect all 32 bits of (c) */
+	switch (size) {
+	case 12:	c += ((uint32_t)k[11])<<24;	/*@fallthrough@*/
+	case 11:	c += ((uint32_t)k[10])<<16;	/*@fallthrough@*/
+	case 10:	c += ((uint32_t)k[9])<<8;	/*@fallthrough@*/
+	case  9:	c += (uint32_t)k[8];		/*@fallthrough@*/
+	case  8:	b += ((uint32_t)k[7])<<24;	/*@fallthrough@*/
+	case  7:	b += ((uint32_t)k[6])<<16;	/*@fallthrough@*/
+	case  6:	b += ((uint32_t)k[5])<<8;	/*@fallthrough@*/
+	case  5:	b += (uint32_t)k[4];		/*@fallthrough@*/
+	case  4:	a += ((uint32_t)k[3])<<24;	/*@fallthrough@*/
+	case  3:	a += ((uint32_t)k[2])<<16;	/*@fallthrough@*/
+	case  2:	a += ((uint32_t)k[1])<<8;	/*@fallthrough@*/
+	case  1:	a += (uint32_t)k[0];
+	    break;
+	case  0:
+	    goto exit;
+	}
+    }
+
+    _JLU3_FINAL(a,b,c);
+
+exit:
+    return c;
+}
+#endif	/* defined(_JLU3_jlu32l) */
+
+#if defined(_JLU3_jlu32lpair)
+/**
+ * jlu32lpair: return 2 32-bit hash values.
+ *
+ * This is identical to jlu32l(), except it returns two 32-bit hash
+ * values instead of just one.  This is good enough for hash table
+ * lookup with 2^^64 buckets, or if you want a second hash if you're not
+ * happy with the first, or if you want a probably-unique 64-bit ID for
+ * the key.  *pc is better mixed than *pb, so use *pc first.  If you want
+ * a 64-bit value do something like "*pc + (((uint64_t)*pb)<<32)".
+ *
+ * @param h		the previous hash, or an arbitrary value
+ * @param *key		the key, an array of uint8_t values
+ * @param size		the size of the key in bytes
+ * @retval *pc,		IN: primary initval, OUT: primary hash
+ * *retval *pb		IN: secondary initval, OUT: secondary hash
+ */
+void jlu32lpair(const void *key, size_t size, uint32_t *pc, uint32_t *pb)
+{
+    union { const void *ptr; size_t i; } u;
+    uint32_t a = _JLU3_INIT(*pc, size);
+    uint32_t b = a;
+    uint32_t c = a;
+
+    if (key == NULL)
+	goto exit;
+
+    c += *pb;	/* Add the secondary hash. */
+
+    u.ptr = key;
+    if (HASH_LITTLE_ENDIAN && ((u.i & 0x3) == 0)) {
+	const uint32_t *k = (const uint32_t *)key;	/* read 32-bit chunks */
+#ifdef	VALGRIND
+	const uint8_t  *k8;
+#endif
+
+	/*-- all but last block: aligned reads and affect 32 bits of (a,b,c) */
+	while (size > (size_t)12) {
+	    a += k[0];
+	    b += k[1];
+	    c += k[2];
+	    _JLU3_MIX(a,b,c);
+	    size -= 12;
+	    k += 3;
+	}
+	/*------------------------- handle the last (probably partial) block */
+	/* 
+	 * "k[2]&0xffffff" actually reads beyond the end of the string, but
+	 * then masks off the part it's not allowed to read.  Because the
+	 * string is aligned, the masked-off tail is in the same word as the
+	 * rest of the string.  Every machine with memory protection I've seen
+	 * does it on word boundaries, so is OK with this.  But VALGRIND will
+	 * still catch it and complain.  The masking trick does make the hash
+	 * noticably faster for short strings (like English words).
+	 */
+#ifndef VALGRIND
+
+	switch (size) {
+	case 12:	c += k[2]; b+=k[1]; a+=k[0]; break;
+	case 11:	c += k[2]&0xffffff; b+=k[1]; a+=k[0]; break;
+	case 10:	c += k[2]&0xffff; b+=k[1]; a+=k[0]; break;
+	case  9:	c += k[2]&0xff; b+=k[1]; a+=k[0]; break;
+	case  8:	b += k[1]; a+=k[0]; break;
+	case  7:	b += k[1]&0xffffff; a+=k[0]; break;
+	case  6:	b += k[1]&0xffff; a+=k[0]; break;
+	case  5:	b += k[1]&0xff; a+=k[0]; break;
+	case  4:	a += k[0]; break;
+	case  3:	a += k[0]&0xffffff; break;
+	case  2:	a += k[0]&0xffff; break;
+	case  1:	a += k[0]&0xff; break;
+	case  0:	goto exit;
+	}
+
+#else /* make valgrind happy */
+
+	k8 = (const uint8_t *)k;
+	switch (size) {
+	case 12:	c += k[2]; b+=k[1]; a+=k[0];	break;
+	case 11:	c += ((uint32_t)k8[10])<<16;	/*@fallthrough@*/
+	case 10:	c += ((uint32_t)k8[9])<<8;	/*@fallthrough@*/
+	case  9:	c += k8[8];			/*@fallthrough@*/
+	case  8:	b += k[1]; a+=k[0];		break;
+	case  7:	b += ((uint32_t)k8[6])<<16;	/*@fallthrough@*/
+	case  6:	b += ((uint32_t)k8[5])<<8;	/*@fallthrough@*/
+	case  5:	b += k8[4];			/*@fallthrough@*/
+	case  4:	a += k[0];			break;
+	case  3:	a += ((uint32_t)k8[2])<<16;	/*@fallthrough@*/
+	case  2:	a += ((uint32_t)k8[1])<<8;	/*@fallthrough@*/
+	case  1:	a += k8[0];			break;
+	case  0:	goto exit;
+	}
+
+#endif /* !valgrind */
+
+    } else if (HASH_LITTLE_ENDIAN && ((u.i & 0x1) == 0)) {
+	const uint16_t *k = (const uint16_t *)key;	/* read 16-bit chunks */
+	const uint8_t  *k8;
+
+	/*----------- all but last block: aligned reads and different mixing */
+	while (size > (size_t)12) {
+	    a += k[0] + (((uint32_t)k[1])<<16);
+	    b += k[2] + (((uint32_t)k[3])<<16);
+	    c += k[4] + (((uint32_t)k[5])<<16);
+	    _JLU3_MIX(a,b,c);
+	    size -= 12;
+	    k += 6;
+	}
+
+	/*------------------------- handle the last (probably partial) block */
+	k8 = (const uint8_t *)k;
+	switch (size) {
+	case 12:
+	    c += k[4]+(((uint32_t)k[5])<<16);
+	    b += k[2]+(((uint32_t)k[3])<<16);
+	    a += k[0]+(((uint32_t)k[1])<<16);
+	    break;
+	case 11:
+	    c += ((uint32_t)k8[10])<<16;
+	    /*@fallthrough@*/
+	case 10:
+	    c += k[4];
+	    b += k[2]+(((uint32_t)k[3])<<16);
+	    a += k[0]+(((uint32_t)k[1])<<16);
+	    break;
+	case  9:
+	    c += k8[8];
+	    /*@fallthrough@*/
+	case  8:
+	    b += k[2]+(((uint32_t)k[3])<<16);
+	    a += k[0]+(((uint32_t)k[1])<<16);
+	    break;
+	case  7:
+	    b += ((uint32_t)k8[6])<<16;
+	    /*@fallthrough@*/
+	case  6:
+	    b += k[2];
+	    a += k[0]+(((uint32_t)k[1])<<16);
+	    break;
+	case  5:
+	    b += k8[4];
+	    /*@fallthrough@*/
+	case  4:
+	    a += k[0]+(((uint32_t)k[1])<<16);
+	    break;
+	case  3:
+	    a += ((uint32_t)k8[2])<<16;
+	    /*@fallthrough@*/
+	case  2:
+	    a += k[0];
+	    break;
+	case  1:
+	    a += k8[0];
+	    break;
+	case  0:
+	    goto exit;
+	}
+
+    } else {		/* need to read the key one byte at a time */
+	const uint8_t *k = (const uint8_t *)key;
+
+	/*----------- all but the last block: affect some 32 bits of (a,b,c) */
+	while (size > (size_t)12) {
+	    a += k[0];
+	    a += ((uint32_t)k[1])<<8;
+	    a += ((uint32_t)k[2])<<16;
+	    a += ((uint32_t)k[3])<<24;
+	    b += k[4];
+	    b += ((uint32_t)k[5])<<8;
+	    b += ((uint32_t)k[6])<<16;
+	    b += ((uint32_t)k[7])<<24;
+	    c += k[8];
+	    c += ((uint32_t)k[9])<<8;
+	    c += ((uint32_t)k[10])<<16;
+	    c += ((uint32_t)k[11])<<24;
+	    _JLU3_MIX(a,b,c);
+	    size -= 12;
+	    k += 12;
+	}
+
+	/*---------------------------- last block: affect all 32 bits of (c) */
+	switch (size) {
+	case 12:	c += ((uint32_t)k[11])<<24;	/*@fallthrough@*/
+	case 11:	c += ((uint32_t)k[10])<<16;	/*@fallthrough@*/
+	case 10:	c += ((uint32_t)k[9])<<8;	/*@fallthrough@*/
+	case  9:	c += k[8];			/*@fallthrough@*/
+	case  8:	b += ((uint32_t)k[7])<<24;	/*@fallthrough@*/
+	case  7:	b += ((uint32_t)k[6])<<16;	/*@fallthrough@*/
+	case  6:	b += ((uint32_t)k[5])<<8;	/*@fallthrough@*/
+	case  5:	b += k[4];			/*@fallthrough@*/
+	case  4:	a += ((uint32_t)k[3])<<24;	/*@fallthrough@*/
+	case  3:	a += ((uint32_t)k[2])<<16;	/*@fallthrough@*/
+	case  2:	a += ((uint32_t)k[1])<<8;	/*@fallthrough@*/
+	case  1:	a += k[0];
+	    break;
+	case  0:
+	    goto exit;
+	}
+    }
+
+    _JLU3_FINAL(a,b,c);
+
+exit:
+    *pc = c;
+    *pb = b;
+    return;
+}
+#endif	/* defined(_JLU3_jlu32lpair) */
+
+#if defined(_JLU3_jlu32b)
+uint32_t jlu32b(uint32_t h, /*@null@*/ const void *key, size_t size)
+	/*@*/;
+/*
+ * jlu32b():
+ * This is the same as jlu32w() on big-endian machines.  It is different
+ * from jlu32l() on all machines.  jlu32b() takes advantage of
+ * big-endian byte ordering. 
+ *
+ * @param h		the previous hash, or an arbitrary value
+ * @param *k		the key, an array of uint8_t values
+ * @param size		the size of the key
+ * @return		the lookup3 hash
+ */
+uint32_t jlu32b(uint32_t h, const void *key, size_t size)
+{
+    union { const void *ptr; size_t i; } u;
+    uint32_t a = _JLU3_INIT(h, size);
+    uint32_t b = a;
+    uint32_t c = a;
+
+    if (key == NULL)
+	return h;
+
+    u.ptr = key;
+    if (HASH_BIG_ENDIAN && ((u.i & 0x3) == 0)) {
+	const uint32_t *k = (const uint32_t *)key;	/* read 32-bit chunks */
+#ifdef	VALGRIND
+	const uint8_t  *k8;
+#endif
+
+	/*-- all but last block: aligned reads and affect 32 bits of (a,b,c) */
+	while (size > 12) {
+	    a += k[0];
+	    b += k[1];
+	    c += k[2];
+	    _JLU3_MIX(a,b,c);
+	    size -= 12;
+	    k += 3;
+	}
+
+	/*------------------------- handle the last (probably partial) block */
+	/* 
+	 * "k[2]<<8" actually reads beyond the end of the string, but
+	 * then shifts out the part it's not allowed to read.  Because the
+	 * string is aligned, the illegal read is in the same word as the
+	 * rest of the string.  Every machine with memory protection I've seen
+	 * does it on word boundaries, so is OK with this.  But VALGRIND will
+	 * still catch it and complain.  The masking trick does make the hash
+	 * noticably faster for short strings (like English words).
+	 */
+#ifndef VALGRIND
+
+	switch (size) {
+	case 12:	c += k[2]; b+=k[1]; a+=k[0]; break;
+	case 11:	c += k[2]&0xffffff00; b+=k[1]; a+=k[0]; break;
+	case 10:	c += k[2]&0xffff0000; b+=k[1]; a+=k[0]; break;
+	case  9:	c += k[2]&0xff000000; b+=k[1]; a+=k[0]; break;
+	case  8:	b += k[1]; a+=k[0]; break;
+	case  7:	b += k[1]&0xffffff00; a+=k[0]; break;
+	case  6:	b += k[1]&0xffff0000; a+=k[0]; break;
+	case  5:	b += k[1]&0xff000000; a+=k[0]; break;
+	case  4:	a += k[0]; break;
+	case  3:	a += k[0]&0xffffff00; break;
+	case  2:	a += k[0]&0xffff0000; break;
+	case  1:	a += k[0]&0xff000000; break;
+	case  0:	goto exit;
+    }
+
+#else  /* make valgrind happy */
+
+	k8 = (const uint8_t *)k;
+	switch (size) {	/* all the case statements fall through */
+	case 12:	c += k[2]; b+=k[1]; a+=k[0];	break;
+	case 11:	c += ((uint32_t)k8[10])<<8;	/*@fallthrough@*/
+	case 10:	c += ((uint32_t)k8[9])<<16;	/*@fallthrough@*/
+	case  9:	c += ((uint32_t)k8[8])<<24;	/*@fallthrough@*/
+	case  8:	b += k[1]; a+=k[0];		break;
+	case  7:	b += ((uint32_t)k8[6])<<8;	/*@fallthrough@*/
+	case  6:	b += ((uint32_t)k8[5])<<16;	/*@fallthrough@*/
+	case  5:	b += ((uint32_t)k8[4])<<24;	/*@fallthrough@*/
+	case  4:	a += k[0];			break;
+	case  3:	a += ((uint32_t)k8[2])<<8;	/*@fallthrough@*/
+	case  2:	a += ((uint32_t)k8[1])<<16;	/*@fallthrough@*/
+	case  1:	a += ((uint32_t)k8[0])<<24;	break;
+	case  0:	goto exit;
+    }
+
+#endif /* !VALGRIND */
+
+    } else {                        /* need to read the key one byte at a time */
+	const uint8_t *k = (const uint8_t *)key;
+
+	/*----------- all but the last block: affect some 32 bits of (a,b,c) */
+	while (size > 12) {
+	    a += ((uint32_t)k[0])<<24;
+	    a += ((uint32_t)k[1])<<16;
+	    a += ((uint32_t)k[2])<<8;
+	    a += ((uint32_t)k[3]);
+	    b += ((uint32_t)k[4])<<24;
+	    b += ((uint32_t)k[5])<<16;
+	    b += ((uint32_t)k[6])<<8;
+	    b += ((uint32_t)k[7]);
+	    c += ((uint32_t)k[8])<<24;
+	    c += ((uint32_t)k[9])<<16;
+	    c += ((uint32_t)k[10])<<8;
+	    c += ((uint32_t)k[11]);
+	    _JLU3_MIX(a,b,c);
+	    size -= 12;
+	    k += 12;
+	}
+
+	/*---------------------------- last block: affect all 32 bits of (c) */
+	switch (size) {	/* all the case statements fall through */
+	case 12:	c += k[11];			/*@fallthrough@*/
+	case 11:	c += ((uint32_t)k[10])<<8;	/*@fallthrough@*/
+	case 10:	c += ((uint32_t)k[9])<<16;	/*@fallthrough@*/
+	case  9:	c += ((uint32_t)k[8])<<24;	/*@fallthrough@*/
+	case  8:	b += k[7];			/*@fallthrough@*/
+	case  7:	b += ((uint32_t)k[6])<<8;	/*@fallthrough@*/
+	case  6:	b += ((uint32_t)k[5])<<16;	/*@fallthrough@*/
+	case  5:	b += ((uint32_t)k[4])<<24;	/*@fallthrough@*/
+	case  4:	a += k[3];			/*@fallthrough@*/
+	case  3:	a += ((uint32_t)k[2])<<8;	/*@fallthrough@*/
+	case  2:	a += ((uint32_t)k[1])<<16;	/*@fallthrough@*/
+	case  1:	a += ((uint32_t)k[0])<<24;	/*@fallthrough@*/
+	    break;
+	case  0:
+	    goto exit;
+	}
+    }
+
+    _JLU3_FINAL(a,b,c);
+
+exit:
+    return c;
+}
+#endif	/* defined(_JLU3_jlu32b) */
+
+#if defined(_JLU3_SELFTEST)
+
+/* used for timings */
+static void driver1(void)
+	/*@*/
+{
+    uint8_t buf[256];
+    uint32_t i;
+    uint32_t h=0;
+    time_t a,z;
+
+    time(&a);
+    for (i=0; i<256; ++i) buf[i] = 'x';
+    for (i=0; i<1; ++i) {
+	h = jlu32l(h, &buf[0], sizeof(buf[0]));
+    }
+    time(&z);
+    if (z-a > 0) printf("time %d %.8x\n", (int)(z-a), h);
+}
+
+/* check that every input bit changes every output bit half the time */
+#define HASHSTATE 1
+#define HASHLEN   1
+#define MAXPAIR 60
+#define MAXLEN  70
+static void driver2(void)
+	/*@*/
+{
+    uint8_t qa[MAXLEN+1], qb[MAXLEN+2], *a = &qa[0], *b = &qb[1];
+    uint32_t c[HASHSTATE], d[HASHSTATE], i=0, j=0, k, l, m=0, z;
+    uint32_t e[HASHSTATE],f[HASHSTATE],g[HASHSTATE],h[HASHSTATE];
+    uint32_t x[HASHSTATE],y[HASHSTATE];
+    uint32_t hlen;
+
+    printf("No more than %d trials should ever be needed \n",MAXPAIR/2);
+    for (hlen=0; hlen < MAXLEN; ++hlen) {
+	z=0;
+	for (i=0; i<hlen; ++i) {	/*-------------- for each input byte, */
+	    for (j=0; j<8; ++j) {	/*--------------- for each input bit, */
+		for (m=1; m<8; ++m) {	/*--- for serveral possible initvals, */
+		    for (l=0; l<HASHSTATE; ++l)
+			e[l]=f[l]=g[l]=h[l]=x[l]=y[l]=~((uint32_t)0);
+
+		    /* check that every output bit is affected by that input bit */
+		    for (k=0; k<MAXPAIR; k+=2) { 
+			uint32_t finished=1;
+			/* keys have one bit different */
+			for (l=0; l<hlen+1; ++l) {a[l] = b[l] = (uint8_t)0;}
+			/* have a and b be two keys differing in only one bit */
+			a[i] ^= (k<<j);
+			a[i] ^= (k>>(8-j));
+			c[0] = jlu32l(m, a, hlen);
+			b[i] ^= ((k+1)<<j);
+			b[i] ^= ((k+1)>>(8-j));
+			d[0] = jlu32l(m, b, hlen);
+			/* check every bit is 1, 0, set, and not set at least once */
+			for (l=0; l<HASHSTATE; ++l) {
+			    e[l] &= (c[l]^d[l]);
+			    f[l] &= ~(c[l]^d[l]);
+			    g[l] &= c[l];
+			    h[l] &= ~c[l];
+			    x[l] &= d[l];
+			    y[l] &= ~d[l];
+			    if (e[l]|f[l]|g[l]|h[l]|x[l]|y[l]) finished=0;
+			}
+			if (finished) break;
+		    }
+		    if (k>z) z=k;
+		    if (k == MAXPAIR) {
+			printf("Some bit didn't change: ");
+			printf("%.8x %.8x %.8x %.8x %.8x %.8x  ",
+				e[0],f[0],g[0],h[0],x[0],y[0]);
+			printf("i %d j %d m %d len %d\n", i, j, m, hlen);
+		    }
+		    if (z == MAXPAIR) goto done;
+		}
+	    }
+	}
+   done:
+	if (z < MAXPAIR) {
+	    printf("Mix success  %2d bytes  %2d initvals  ",i,m);
+	    printf("required  %d  trials\n", z/2);
+	}
+    }
+    printf("\n");
+}
+
+/* Check for reading beyond the end of the buffer and alignment problems */
+static void driver3(void)
+	/*@*/
+{
+    uint8_t buf[MAXLEN+20], *b;
+    uint32_t len;
+    uint8_t q[] = "This is the time for all good men to come to the aid of their country...";
+    uint32_t h;
+    uint8_t qq[] = "xThis is the time for all good men to come to the aid of their country...";
+    uint32_t i;
+    uint8_t qqq[] = "xxThis is the time for all good men to come to the aid of their country...";
+    uint32_t j;
+    uint8_t qqqq[] = "xxxThis is the time for all good men to come to the aid of their country...";
+    uint32_t ref,x,y;
+    uint8_t *p;
+    uint32_t m = 13;
+
+    printf("Endianness.  These lines should all be the same (for values filled in):\n");
+    printf("%.8x                            %.8x                            %.8x\n",
+	jlu32w(m, (const uint32_t *)q, (sizeof(q)-1)/4),
+	jlu32w(m, (const uint32_t *)q, (sizeof(q)-5)/4),
+	jlu32w(m, (const uint32_t *)q, (sizeof(q)-9)/4));
+    p = q;
+    printf("%.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x\n",
+	jlu32l(m, p, sizeof(q)-1), jlu32l(m, p, sizeof(q)-2),
+	jlu32l(m, p, sizeof(q)-3), jlu32l(m, p, sizeof(q)-4),
+	jlu32l(m, p, sizeof(q)-5), jlu32l(m, p, sizeof(q)-6),
+	jlu32l(m, p, sizeof(q)-7), jlu32l(m, p, sizeof(q)-8),
+	jlu32l(m, p, sizeof(q)-9), jlu32l(m, p, sizeof(q)-10),
+	jlu32l(m, p, sizeof(q)-11), jlu32l(m, p, sizeof(q)-12));
+    p = &qq[1];
+    printf("%.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x\n",
+	jlu32l(m, p, sizeof(q)-1), jlu32l(m, p, sizeof(q)-2),
+	jlu32l(m, p, sizeof(q)-3), jlu32l(m, p, sizeof(q)-4),
+	jlu32l(m, p, sizeof(q)-5), jlu32l(m, p, sizeof(q)-6),
+	jlu32l(m, p, sizeof(q)-7), jlu32l(m, p, sizeof(q)-8),
+	jlu32l(m, p, sizeof(q)-9), jlu32l(m, p, sizeof(q)-10),
+	jlu32l(m, p, sizeof(q)-11), jlu32l(m, p, sizeof(q)-12));
+    p = &qqq[2];
+    printf("%.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x\n",
+	jlu32l(m, p, sizeof(q)-1), jlu32l(m, p, sizeof(q)-2),
+	jlu32l(m, p, sizeof(q)-3), jlu32l(m, p, sizeof(q)-4),
+	jlu32l(m, p, sizeof(q)-5), jlu32l(m, p, sizeof(q)-6),
+	jlu32l(m, p, sizeof(q)-7), jlu32l(m, p, sizeof(q)-8),
+	jlu32l(m, p, sizeof(q)-9), jlu32l(m, p, sizeof(q)-10),
+	jlu32l(m, p, sizeof(q)-11), jlu32l(m, p, sizeof(q)-12));
+    p = &qqqq[3];
+    printf("%.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x\n",
+	jlu32l(m, p, sizeof(q)-1), jlu32l(m, p, sizeof(q)-2),
+	jlu32l(m, p, sizeof(q)-3), jlu32l(m, p, sizeof(q)-4),
+	jlu32l(m, p, sizeof(q)-5), jlu32l(m, p, sizeof(q)-6),
+	jlu32l(m, p, sizeof(q)-7), jlu32l(m, p, sizeof(q)-8),
+	jlu32l(m, p, sizeof(q)-9), jlu32l(m, p, sizeof(q)-10),
+	jlu32l(m, p, sizeof(q)-11), jlu32l(m, p, sizeof(q)-12));
+    printf("\n");
+    for (h=0, b=buf+1; h<8; ++h, ++b) {
+	for (i=0; i<MAXLEN; ++i) {
+	    len = i;
+	    for (j=0; j<i; ++j)
+		*(b+j)=0;
+
+	    /* these should all be equal */
+	    m = 1;
+	    ref = jlu32l(m, b, len);
+	    *(b+i)=(uint8_t)~0;
+	    *(b-1)=(uint8_t)~0;
+	    x = jlu32l(m, b, len);
+	    y = jlu32l(m, b, len);
+	    if ((ref != x) || (ref != y)) 
+		printf("alignment error: %.8x %.8x %.8x %d %d\n",ref,x,y, h, i);
+	}
+    }
+}
+
+/* check for problems with nulls */
+static void driver4(void)
+	/*@*/
+{
+    uint8_t buf[1];
+    uint32_t h;
+    uint32_t i;
+    uint32_t state[HASHSTATE];
+
+    buf[0] = ~0;
+    for (i=0; i<HASHSTATE; ++i)
+	state[i] = 1;
+    printf("These should all be different\n");
+    h = 0;
+    for (i=0; i<8; ++i) {
+	h = jlu32l(h, buf, 0);
+	printf("%2ld  0-byte strings, hash is  %.8x\n", (long)i, h);
+    }
+}
+
+
+int main(int argc, char ** argv)
+{
+    driver1();	/* test that the key is hashed: used for timings */
+    driver2();	/* test that whole key is hashed thoroughly */
+    driver3();	/* test that nothing but the key is hashed */
+    driver4();	/* test hashing multiple buffers (all buffers are null) */
+    return 1;
+}
+
+#endif  /* _JLU3_SELFTEST */
diff --git a/popt.c b/popt.c
index b697a0d..450a214 100644
--- a/popt.c
+++ b/popt.c
@@ -462,8 +462,10 @@ const char * findProgramPath(/*@null@*/ const char * argv0)
     }
 
     /* If no executable was found in PATH, return NULL. */
+/*@-compdef@*/
     if (!(s && *s) && t != NULL)
 	t = _free(t);
+/*@=compdef@*/
 /*@-modobserver -observertrans -usedef @*/
     path = _free(path);
 /*@=modobserver =observertrans =usedef @*/
@@ -750,6 +752,111 @@ static void poptStripArg(/*@special@*/ poptContext con, int which)
 /*@=compdef =sizeoftype =usedef @*/
 }
 
+/*@unchecked@*/
+unsigned int _poptBitsN = _POPT_BITS_N;
+/*@unchecked@*/
+unsigned int _poptBitsM = _POPT_BITS_M;
+/*@unchecked@*/
+unsigned int _poptBitsK = _POPT_BITS_K;
+
+/*@-sizeoftype@*/
+int poptBitsAdd(poptBits bits, const char * s)
+{
+    size_t ns = (s ? strlen(s) : 0);
+    uint32_t h0 = 0;
+    uint32_t h1 = 0;
+
+    if (bits == NULL || ns == 0)
+	return POPT_ERROR_NULLARG;
+
+    /* XXX parse CSV? */
+    jlu32lpair(s, ns, &h0, &h1);
+
+    for (ns = 0; ns < (size_t)_poptBitsK; ns++) {
+        uint32_t h = h0 + ns * h1;
+        uint32_t ix = (h % _poptBitsM);
+        PBM_SET(ix, bits);
+    }
+    return 0;
+}
+
+int poptBitsChk(poptBits bits, const char * s)
+{
+    size_t ns = (s ? strlen(s) : 0);
+    uint32_t h0 = 0;
+    uint32_t h1 = 0;
+    int rc = 1;
+
+    if (bits == NULL || ns == 0)
+	return POPT_ERROR_NULLARG;
+
+    /* XXX parse CSV? */
+    jlu32lpair(s, ns, &h0, &h1);
+
+    for (ns = 0; ns < (size_t)_poptBitsK; ns++) {
+        uint32_t h = h0 + ns * h1;
+        uint32_t ix = (h % _poptBitsM);
+        if (PBM_ISSET(ix, bits))
+            continue;
+        rc = 0;
+        break;
+    }
+    return rc;
+}
+
+int poptBitsClr(poptBits bits)
+{
+    static size_t nbw = (__PBM_NBITS/8);
+    size_t nw = (__PBM_IX(_poptBitsM-1) + 1);
+
+    if (bits == NULL)
+	return POPT_ERROR_NULLARG;
+    memset(bits, 0, nw * nbw);
+    return 0;
+}
+
+int poptBitsDel(poptBits bits, const char * s)
+{
+    size_t ns = (s ? strlen(s) : 0);
+    uint32_t h0 = 0;
+    uint32_t h1 = 0;
+
+    if (bits == NULL || ns == 0)
+	return POPT_ERROR_NULLARG;
+
+    /* XXX parse CSV? */
+    jlu32lpair(s, ns, &h0, &h1);
+
+    for (ns = 0; ns < (size_t)_poptBitsK; ns++) {
+        uint32_t h = h0 + ns * h1;
+        uint32_t ix = (h % _poptBitsM);
+        PBM_CLR(ix, bits);
+    }
+    return 0;
+}
+
+int poptSaveBits(poptBits * bitsp,
+		/*@unused@*/ UNUSED(unsigned int argInfo), const char * s)
+{
+    if (bitsp == NULL)
+	return POPT_ERROR_NULLARG;
+
+    /* XXX handle negated initialization. */
+    if (*bitsp == NULL) {
+	if (_poptBitsN == 0) {
+	    _poptBitsN = _POPT_BITS_N;
+	    _poptBitsM = _POPT_BITS_M;
+	}
+	if (_poptBitsM == 0U) _poptBitsM = (3 * _poptBitsN) / 2;
+	if (_poptBitsK == 0U || _poptBitsK > 32U) _poptBitsK = _POPT_BITS_K;
+	*bitsp = PBM_ALLOC(_poptBitsM-1);
+    }
+/*@-nullstate@*/
+    return poptBitsAdd(*bitsp, s);
+/*@=nullstate@*/
+}
+/*@=sizeoftype@*/
+
 int poptSaveString(const char *** argvp,
 		/*@unused@*/ UNUSED(unsigned int argInfo), const char * val)
 {
@@ -952,6 +1059,31 @@ static unsigned int poptArgInfo(poptContext con, const struct poptOption * opt)
 }
 
 /**
+ * Parse an integer expression.
+ * @retval *llp		integer expression value
+ * @param argInfo	integer expression type
+ * @param val		integer expression string
+ * @return		0 on success, otherwise POPT_* error.
+ */
+static int poptParseInteger(long long * llp,
+		/*@unused@*/ UNUSED(unsigned int argInfo),
+		/*@null@*/ const char * val)
+	/*@modifies *llp @*/
+{
+    if (val) {
+	char *end = NULL;
+	*llp = strtoll(val, &end, 0);
+
+	/* XXX parse scaling suffixes here. */
+
+	if (!(end && *end == '\0'))
+	    return POPT_ERROR_BADNUMBER;
+    } else
+	*llp = 0;
+    return 0;
+}
+
+/**
  * Save the option argument through the (*opt->arg) pointer.
  * @param con		context
  * @param opt           option
@@ -965,6 +1097,10 @@ static int poptSaveArg(poptContext con, const struct poptOption * opt)
     int rc = 0;		/* assume success */
 
     switch (poptArgType(opt)) {
+    case POPT_ARG_BITSET:
+	/* XXX memory leak, application is responsible for free. */
+	rc = poptSaveBits(arg.ptr, opt->argInfo, con->os->nextArg);
+	/*@switchbreak@*/ break;
     case POPT_ARG_ARGV:
 	/* XXX memory leak, application is responsible for free. */
 	rc = poptSaveString(arg.ptr, opt->argInfo, con->os->nextArg);
@@ -979,16 +1115,10 @@ static int poptSaveArg(poptContext con, const struct poptOption * opt)
     case POPT_ARG_LONG:
     case POPT_ARG_LONGLONG:
     {	unsigned int argInfo = poptArgInfo(con, opt);
-	char *end = NULL;
 	long long aNUM = 0;
 
-	if (con->os->nextArg) {
-	    aNUM = strtoll(con->os->nextArg, &end, 0);
-	    if (!(end && *end == '\0')) {
-		rc = POPT_ERROR_BADNUMBER;
-		break;
-	    }
-	}
+	if ((rc = poptParseInteger(&aNUM, argInfo, con->os->nextArg)) != 0)
+	    break;
 
 	switch (poptArgType(opt)) {
 	case POPT_ARG_LONGLONG:
diff --git a/popt.h b/popt.h
index 278357f..aa094df 100644
--- a/popt.h
+++ b/popt.h
@@ -38,6 +38,7 @@
 #define POPT_ARG_MAINCALL	16U+11U	/*!< EXPERIMENTAL: return (*arg) (argc, argv) */
 #define	POPT_ARG_ARGV		12U	/*!< dupe'd arg appended to realloc'd argv array. */
 #define	POPT_ARG_SHORT		13U	/*!< arg ==> short */
+#define	POPT_ARG_BITSET		16U+14U	/*!< arg ==> bit set */
 
 #define POPT_ARG_MASK		0x000000FFU
 #define POPT_GROUP_MASK		0x0000FF00U
@@ -610,7 +611,7 @@ int poptStrippedArgv(poptContext con, int argc, char ** argv)
  */
 /*@unused@*/
 int poptSaveString(/*@null@*/ const char *** argvp, unsigned int argInfo,
-		const char * val)
+		/*@null@*/const char * val)
 	/*@modifies *argvp @*/;
 
 /**
@@ -678,7 +679,56 @@ int poptSaveInt(/*@null@*/ int * arg, unsigned int argInfo, long aLong)
 	/*@requires maxSet(arg) >= 0 /\ maxRead(arg) == 0 @*/;
 /*@=incondefs@*/
 
+/* The bit set typedef. */
+/*@-exporttype@*/
+typedef struct poptBits_s {
+    unsigned int bits[1];
+} * poptBits;
+/*@=exporttype@*/
+
+#define _POPT_BITS_N    1024U    /* estimated population */
+#define _POPT_BITS_M    ((3U * _POPT_BITS_N) / 2U)
+#define _POPT_BITS_K    16U      /* no. of linear hash combinations */
+
+/*@-exportlocal -exportvar@*/
+/*@unchecked@*/
+extern unsigned int _poptBitsN;
+/*@unchecked@*/
+extern  unsigned int _poptBitsM;
+/*@unchecked@*/
+extern  unsigned int _poptBitsK;
+/*@=exportlocal =exportvar@*/
+
+/*@-exportlocal@*/
+int poptBitsAdd(/*@null@*/poptBits bits, /*@null@*/const char * s)
+	/*@modifies bits @*/;
+/*@-fcnuse@*/
+int poptBitsChk(/*@null@*/poptBits bits, /*@null@*/const char * s)
+	/*@*/;
+int poptBitsClr(/*@null@*/poptBits bits)
+	/*@modifies bits @*/;
+int poptBitsDel(/*@null@*/poptBits bits, /*@null@*/const char * s)
+	/*@modifies bits @*/;
+/*@=fcnuse@*/
+/*@=exportlocal@*/
+
+/**
+ * Save a string into a bit set (experimental).
+ * @retval *bits	bit set (lazily malloc'd if NULL)
+ * @param argInfo	logical operation (see POPT_ARGFLAG_*)
+ * @param s		string to add to bit set
+ * @return		0 on success, POPT_ERROR_NULLARG/POPT_ERROR_BADOPERATION
+ */
+/*@-incondefs@*/
+/*@unused@*/
+int poptSaveBits(/*@null@*/ poptBits * bitsp, unsigned int argInfo,
+		/*@null@*/ const char * s)
+	/*@globals _poptBitsN, _poptBitsM, _poptBitsK, internalState @*/
+	/*@modifies *bitsp, _poptBitsN, _poptBitsM, _poptBitsK, internalState @*/;
+/*@=incondefs@*/
+
 /*@=type@*/
+
 #ifdef  __cplusplus
 }
 #endif
diff --git a/poptint.c b/poptint.c
index 52ab0ae..d369d1f 100644
--- a/poptint.c
+++ b/poptint.c
@@ -2,6 +2,10 @@
 #include <stdarg.h>
 #include "poptint.h"
 
+/* Any pair of 32 bit hashes can be used. lookup3.c generates pairs, will do. */
+#define _JLU3_jlu32lpair        1
+#include "lookup3.c"
+
 /*@-varuse +charint +ignoresigns @*/
 /*@unchecked@*/ /*@observer@*/
 static const unsigned char utf8_skip_data[256] = {
diff --git a/poptint.h b/poptint.h
index 885f56b..95674d3 100644
--- a/poptint.h
+++ b/poptint.h
@@ -9,6 +9,8 @@
 #ifndef H_POPTINT
 #define H_POPTINT
 
+#include <stdint.h>
+
 /**
  * Wrapper to free(3), hides const compilation noise, permit NULL, return NULL.
  * @param p		memory to free
@@ -42,6 +44,10 @@ typedef struct {
 #define PBM_CLR(d, s)   (__PBM_BITS (s)[__PBM_IX (d)] &= ~__PBM_MASK (d))
 #define PBM_ISSET(d, s) ((__PBM_BITS (s)[__PBM_IX (d)] & __PBM_MASK (d)) != 0)
 
+void jlu32lpair(/*@null@*/ const void *key, size_t size,
+                uint32_t *pc, uint32_t *pb)
+        /*@modifies *pc, *pb@*/;
+
 /** \ingroup popt
  * Typedef's for string and array of strings.
  */
diff --git a/test1.c b/test1.c
index c9effd5..35fd376 100644
--- a/test1.c
+++ b/test1.c
@@ -38,9 +38,9 @@ static unsigned int aFlag = 0x8aceU;
 static unsigned int bFlag = 0x8aceU;
 
 /*@unchecked@*/
-static short aShort = 4523;
+static short aShort = (short)4523;
 /*@unchecked@*/
-static short bShort = 4523;
+static short bShort = (short)4523;
 /*@unchecked@*/
 static int aInt = 271828;
 /*@unchecked@*/
@@ -61,8 +61,11 @@ static float bFloat = 3.1415926535;
 static double aDouble = 9.86960440108935861883;
 /*@unchecked@*/
 static double bDouble = 9.86960440108935861883;
+
 /*@unchecked@*/ /*@only@*/ /*@null@*/
 static const char ** aArgv = NULL;
+/*@unchecked@*/ /*@only@*/ /*@null@*/
+static void * aBits = NULL;
 
 /*@unchecked@*/ /*@null@*/
 static char * oStr = (char *) -1;
@@ -151,7 +154,9 @@ static struct poptOption options[] = {
 	"POPT_ARGFLAG_RANDOM: experimental", NULL },
 
    { "argv", '\0', POPT_ARG_ARGV, &aArgv, 0,
-	"POPT_ARG_ARGV: append arg to array (can be used multiple times)",NULL},
+	"POPT_ARG_ARGV: append string to argv array (can be used multiple times)","STRING"},
+   { "bits", '\0', POPT_ARG_BITSET|POPT_ARGFLAG_DOC_HIDDEN, &aBits, 0,
+	"POPT_ARG_BITSET: add string to bit set (can be used multiple times)","STRING"},
 
   { "bitset", '\0', POPT_BIT_SET | POPT_ARGFLAG_TOGGLE | POPT_ARGFLAG_SHOW_DEFAULT, &aFlag, 0x7777,
 	"POPT_BIT_SET: |= 0x7777", 0},
@@ -179,10 +184,10 @@ static struct poptOption options[] = {
 static void resetVars(void)
 	/*@globals arg1, arg2, arg3, inc, shortopt,
 		aVal, aFlag, aShort, aInt, aLong, aLongLong, aFloat, aDouble,
-		aArgv, oStr, singleDash, pass2 @*/
+		aArgv, aBits, oStr, singleDash, pass2 @*/
 	/*@modifies arg1, arg2, arg3, inc, shortopt,
 		aVal, aFlag, aShort, aInt, aLong, aLongLong, aFloat, aDouble,
-		aArgv, oStr, singleDash, pass2 @*/
+		aArgv, aBits, oStr, singleDash, pass2 @*/
 {
     arg1 = 0;
     arg2 = "(none)";
@@ -211,6 +216,10 @@ static void resetVars(void)
 	free(aArgv);
 	aArgv = NULL;
     }
+    if (aBits) {
+	free(aBits);
+	aBits = NULL;
+    }
 
     oStr = (char *) -1;
 
@@ -287,7 +296,7 @@ int main(int argc, const char ** argv)
     if (aFlag != bFlag)
 	fprintf(stdout, " aFlag: 0x%x", aFlag);
     if (aShort != bShort)
-	fprintf(stdout, " aShort: %d", aShort);
+	fprintf(stdout, " aShort: %hd", aShort);
     if (aInt != bInt)
 	fprintf(stdout, " aInt: %d", aInt);
     if (aLong != bLong)
@@ -307,6 +316,9 @@ int main(int argc, const char ** argv)
 	while ((arg = *av++) != NULL)
 	    fprintf(stdout, " %s", arg);
     }
+    if (aBits) {
+	fprintf(stdout, " aBits: non-null");
+    }
 /*@-nullpass@*/
     if (oStr != (char *)-1)
 	fprintf(stdout, " oStr: %s", (oStr ? oStr : "(none)"));
diff --git a/testit.sh b/testit.sh
index 1f833f9..857f6a4 100755
--- a/testit.sh
+++ b/testit.sh
@@ -112,17 +112,18 @@ run test1 "test1 - 54" "arg1: 0 arg2: (none) aFlag: 0x28c" --bitclr
 run test1 "test1 - 55" "arg1: 0 arg2: (none) aFlag: 0x8888" --nobitset
 run test1 "test1 - 56" "arg1: 0 arg2: (none) aFlag: 0xface" --nobitclr
 
-run test1 "test1 - 55" "\
+run test1 "test1 - 57" "arg1: 0 arg2: (none) aBits: non-null" --bits foo --bits bar
+
+run test1 "test1 - 58" "\
 Usage: lt-test1 [-I?] [-c|--cb2=STRING] [--arg1] [-2|--arg2=ARG]
         [-3|--arg3=ANARG] [-onedash] [--optional=STRING] [--val]
         [-i|--int=INT] [-s|--short=SHORT] [-l|--long=LONG]
         [-L|--longlong=LONGLONG] [-f|--float=FLOAT] [-d|--double=DOUBLE]
         [--randint=INT] [--randshort=SHORT] [--randlong=LONG]
-        [--randlonglong=LONGLONG] [--argv] [--bitset] [--bitclr] [--bitxor]
-        [--nstr=STRING] [--lstr=STRING] [-I|--inc] [-c|--cb=STRING]
-        [--longopt] [-?|--help] [--usage] [--simple=ARG]" --usage
-
-run test1 "test1 - 56" "\
+        [--randlonglong=LONGLONG] [--argv=STRING] [--bitset] [--bitclr]
+        [--bitxor] [--nstr=STRING] [--lstr=STRING] [-I|--inc]
+        [-c|--cb=STRING] [--longopt] [-?|--help] [--usage] [--simple=ARG]" --usage
+run test1 "test1 - 59" "\
 Usage: lt-test1 [OPTION...]
       --arg1                      First argument with a really long
                                   description. After all, we have to test
@@ -144,8 +145,8 @@ Usage: lt-test1 [OPTION...]
       --randshort=SHORT           POPT_ARGFLAG_RANDOM: experimental
       --randlong=LONG             POPT_ARGFLAG_RANDOM: experimental
       --randlonglong=LONGLONG     POPT_ARGFLAG_RANDOM: experimental
-      --argv                      POPT_ARG_ARGV: append arg to array (can be
-                                  used multiple times)
+      --argv STRING               POPT_ARG_ARGV: append string to argv array
+                                  (can be used multiple times)
       --[no]bitset                POPT_BIT_SET: |= 0x7777
       --[no]bitclr                POPT_BIT_CLR: &= ~0xf842
       --bitxor                    POPT_ARGFLAG_XOR: ^= (0x8ace^0xfeed)