/*
	pr_exec.c

	(description)

	Copyright (C) 1996-1997  Id Software, Inc.

	This program is free software; you can redistribute it and/or
	modify it under the terms of the GNU General Public License
	as published by the Free Software Foundation; either version 2
	of the License, or (at your option) any later version.

	This program is distributed in the hope that it will be useful,
	but WITHOUT ANY WARRANTY; without even the implied warranty of
	MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

	See the GNU General Public License for more details.

	You should have received a copy of the GNU General Public License
	along with this program; if not, write to:

		Free Software Foundation, Inc.
		59 Temple Place - Suite 330
		Boston, MA  02111-1307, USA

*/
#ifdef HAVE_CONFIG_H
# include "config.h"
#endif

#ifdef HAVE_STRING_H
# include <string.h>
#endif
#ifdef HAVE_STRINGS_H
# include <strings.h>
#endif
#include <signal.h>
#include <stdarg.h>

#include "QF/cvar.h"
#include "QF/dstring.h"
#include "QF/mathlib.h"
#include "QF/progs.h"
#include "QF/sys.h"
#include "QF/zone.h"

#include "QF/simd/vec2d.h"
#include "QF/simd/vec2f.h"
#include "QF/simd/vec2i.h"
#include "QF/simd/vec4d.h"
#include "QF/simd/vec4f.h"
#include "QF/simd/vec4i.h"
#include "compat.h"

const char *prdebug_names[] = {
	[prd_none] = "none",
	[prd_trace] = "trace",
	[prd_breakpoint] = "breakpoint",
	[prd_watchpoint] = "watchpoint",
	[prd_subenter] = "subenter",
	[prd_subexit] = "subexit",
	[prd_begin] = "begin",
	[prd_terminate] = "terminate",
	[prd_runerror] = "runerror",
	[prd_error] = "error",
};

/*
	PR_RunError

	Aborts the currently executing function
*/
VISIBLE void
PR_RunError (progs_t * pr, const char *error, ...)
{
	dstring_t  *string = dstring_new ();//FIXME leaks when debugging
	va_list     argptr;

	va_start (argptr, error);
	dvsprintf (string, error, argptr);
	va_end (argptr);

	if (pr->debug_handler) {
		pr->debug_handler (prd_runerror, string->str, pr->debug_data);
		// not expected to return, but if so, behave as if there was no handler
	}

	Sys_Printf ("%s\n", string->str);

	PR_DumpState (pr);

	// dump the stack so PR_Error can shutdown functions
	pr->pr_depth = 0;
	pr->localstack_used = 0;

	PR_Error (pr, "Program error: %s", string->str);
}

VISIBLE pr_stashed_params_t *
_PR_SaveParams (progs_t *pr, pr_stashed_params_t *params)
{
	int         i;
	int         size = pr->pr_param_size * sizeof (pr_type_t);

	params->param_ptrs[0] = pr->pr_params[0];
	params->param_ptrs[1] = pr->pr_params[1];
	pr->pr_params[0] = pr->pr_real_params[0];
	pr->pr_params[1] = pr->pr_real_params[1];
	for (i = 0; i < pr->pr_argc; i++) {
		memcpy (params->params + i * pr->pr_param_size,
				pr->pr_real_params[i], size);
		if (i < 2) { //XXX FIXME what the what?!?
			memcpy (pr->pr_real_params[i], params->param_ptrs[0], size);
		}
	}
	params->argc = pr->pr_argc;
	return params;
}

VISIBLE void
PR_RestoreParams (progs_t *pr, pr_stashed_params_t *params)
{
	int         i;
	int         size = pr->pr_param_size * sizeof (pr_type_t);

	pr->pr_params[0] = params->param_ptrs[0];
	pr->pr_params[1] = params->param_ptrs[1];
	pr->pr_argc = params->argc;
	for (i = 0; i < pr->pr_argc; i++) {
		memcpy (pr->pr_real_params[i],
				params->params + i * pr->pr_param_size, size);
	}
}

VISIBLE inline void
PR_PushFrame (progs_t *pr)
{
	prstack_t  *frame;

	if (pr->pr_depth == MAX_STACK_DEPTH)
		PR_RunError (pr, "stack overflow");

	frame = pr->pr_stack + pr->pr_depth++;

	frame->staddr = pr->pr_xstatement;
	frame->func   = pr->pr_xfunction;
	frame->tstr   = pr->pr_xtstr;

	pr->pr_xtstr = pr->pr_pushtstr;
	pr->pr_pushtstr = 0;
	pr->pr_xfunction = 0;
}

VISIBLE inline void
PR_PopFrame (progs_t *pr)
{
	prstack_t  *frame;

	if (pr->pr_depth <= 0)
		PR_Error (pr, "prog stack underflow");

	if (pr->pr_xtstr)
		PR_FreeTempStrings (pr);
	// normally, this won't happen, but if a builtin pushed a temp string
	// when calling a function and the callee was another builtin that
	// did not call a progs function, then the push strings will still be
	// valid because PR_EnterFunction was never called
	// however, not if a temp string survived: better to hold on to the push
	// strings a little longer than lose one erroneously
	if (!pr->pr_xtstr && pr->pr_pushtstr) {
		pr->pr_xtstr = pr->pr_pushtstr;
		pr->pr_pushtstr = 0;
		PR_FreeTempStrings (pr);
	}

	// up stack
	frame = pr->pr_stack + --pr->pr_depth;

	pr->pr_xfunction  = frame->func;
	pr->pr_xstatement = frame->staddr;
	pr->pr_xtstr      = frame->tstr;
}

static __attribute__((pure)) long
align_offset (long offset, dparmsize_t parmsize)
{
	int         mask = (1 << parmsize.alignment) - 1;
	return (offset + mask) & ~mask;
}

static void
copy_param (pr_type_t *dst, pr_type_t *src, size_t size)
{
	while (size--) {
		memcpy (dst++, src++, sizeof (pr_type_t));
	}
}

/** Setup the stackframe prior to calling a progs function. Saves all local
	data the called function will trample on and copies the parameters used
	by the function into the function's local data space.
	\param pr pointer to progs_t VM struct
	\param f pointer to the descriptor for the called function
	\note Passing a descriptor for a builtin function will result in
	undefined behavior.
*/
static void
PR_EnterFunction (progs_t *pr, bfunction_t *f)
{
	pr_int_t    i;
	pr_type_t  *dstParams[MAX_PARMS];
	pointer_t   paramofs = 0;

	if (pr->pr_trace && !pr->debug_handler) {
		Sys_Printf ("Entering function %s\n",
					PR_GetString (pr, f->descriptor->name));
	}

	PR_PushFrame (pr);

	if (f->numparms > 0) {
		paramofs = f->parm_start;
		for (i = 0; i < f->numparms; i++) {
			paramofs = align_offset (paramofs, f->parm_size[i]);
			dstParams[i] = pr->pr_globals + paramofs;
			paramofs += f->parm_size[i].size;
			if (pr->pr_params[i] != pr->pr_real_params[i]) {
				copy_param (pr->pr_real_params[i], pr->pr_params[i],
							f->parm_size[i].size);
				pr->pr_params[i] = pr->pr_real_params[i];
			}
		}
	} else if (f->numparms < 0) {
		paramofs = f->parm_start + 2;	// argc and argv
		for (i = 0; i < -f->numparms - 1; i++) {
			paramofs = align_offset (paramofs, f->parm_size[i]);
			dstParams[i] = pr->pr_globals + paramofs;
			paramofs += f->parm_size[i].size;
			if (pr->pr_params[i] != pr->pr_real_params[i]) {
				copy_param (pr->pr_real_params[i], pr->pr_params[i],
							f->parm_size[i].size);
				pr->pr_params[i] = pr->pr_real_params[i];
			}
		}
		dparmsize_t parmsize = { pr->pr_param_size, pr->pr_param_alignment };
		paramofs = align_offset (paramofs, parmsize );
		if (i < MAX_PARMS) {
			dstParams[i] = pr->pr_globals + paramofs;
		}
		for (; i < pr->pr_argc; i++) {
			if (pr->pr_params[i] != pr->pr_real_params[i]) {
				copy_param (pr->pr_real_params[i], pr->pr_params[i],
							parmsize.size);
				pr->pr_params[i] = pr->pr_real_params[i];
			}
		}
	}

	//Sys_Printf("%s:\n", PR_GetString(pr,f->name));
	pr->pr_xfunction = f;
	pr->pr_xstatement = f->first_statement - 1;      		// offset the st++

	// save off any locals that the new function steps on
	if (pr->localstack_used + f->locals > LOCALSTACK_SIZE)
		PR_RunError (pr, "PR_EnterFunction: locals stack overflow");

	memcpy (&pr->localstack[pr->localstack_used],
			&pr->pr_globals[f->parm_start],
			sizeof (pr_type_t) * f->locals);
	pr->localstack_used += f->locals;

	if (pr_deadbeef_locals->int_val)
		for (i = f->parm_start; i < f->parm_start + f->locals; i++)
			pr->pr_globals[i].integer_var = 0xdeadbeef;

	// copy parameters
	if (f->numparms >= 0) {
		for (i = 0; i < f->numparms; i++) {
			copy_param (dstParams[i], pr->pr_params[i], f->parm_size[i].size);
		}
	} else {
		int         copy_args;
		pr_type_t  *argc = &pr->pr_globals[f->parm_start + 0];
		pr_type_t  *argv = &pr->pr_globals[f->parm_start + 1];
		for (i = 0; i < -f->numparms - 1; i++) {
			copy_param (dstParams[i], pr->pr_params[i], f->parm_size[i].size);
		}
		copy_args = pr->pr_argc - i;
		argc->integer_var = copy_args;
		argv->integer_var = dstParams[i] - pr->pr_globals;
		if (i < MAX_PARMS) {
			memcpy (dstParams[i], pr->pr_params[i],
					(copy_args * pr->pr_param_size) * sizeof (pr_type_t));
		}
	}
}

static void
PR_LeaveFunction (progs_t *pr, int to_engine)
{
	bfunction_t *f = pr->pr_xfunction;

	PR_PopFrame (pr);

	if (pr->pr_trace && !pr->debug_handler) {
		Sys_Printf ("Leaving function %s\n",
					PR_GetString (pr, f->descriptor->name));
		if (to_engine) {
			Sys_Printf ("Returning to engine\n");
		} else {
			bfunction_t *rf = pr->pr_xfunction;
			if (rf) {
				Sys_Printf ("Returning to function %s\n",
							PR_GetString (pr, rf->descriptor->name));
			}
		}
	}

	// restore locals from the stack
	pr->localstack_used -= f->locals;
	if (pr->localstack_used < 0)
		PR_RunError (pr, "PR_LeaveFunction: locals stack underflow");

	memcpy (&pr->pr_globals[f->parm_start],
			&pr->localstack[pr->localstack_used],
			sizeof (pr_type_t) * f->locals);
}

VISIBLE void
PR_BoundsCheckSize (progs_t *pr, pointer_t addr, unsigned size)
{
	if (addr < (pointer_t) (pr->pr_return - pr->pr_globals))
		PR_RunError (pr, "null pointer access");
	if (addr >= pr->globals_size
		|| size > (unsigned) (pr->globals_size - addr))
		PR_RunError (pr, "invalid memory access: %d (0 to %d-%d)", addr,
					 pr->globals_size, size);
	if (pr_boundscheck->int_val >= 2
		&& PR_GetPointer (pr, addr + size) > (pr_type_t *) pr->zone) {
		void       *mem = (void *) PR_GetPointer (pr, addr);
		Z_CheckPointer (pr->zone, mem, size * sizeof (pr_type_t));
	}
}

VISIBLE void
PR_BoundsCheck (progs_t *pr, int addr, etype_t type)
{
	PR_BoundsCheckSize (pr, addr, pr_type_size[type]);
}

#define OPA(type) (*((pr_##type##_t *) (op_a)))
#define OPB(type) (*((pr_##type##_t *) (op_b)))
#define OPC(type) (*((pr_##type##_t *) (op_c)))

/*
	This gets around the problem of needing to test for -0.0 but denormals
	causing exceptions (or wrong results for what we need) on the alpha.
*/
#define FNZ(x) ((x) & ~0x80000000u)

static int
signal_hook (int sig, void *data)
{
	progs_t    *pr = (progs_t *) data;

	if (sig == SIGFPE && pr_faultchecks->int_val) {
		dstatement_t *st;
		pr_type_t  *op_a, *op_b, *op_c;

		st = pr->pr_statements + pr->pr_xstatement;
		op_a = pr->pr_globals + st->a;
		op_b = pr->pr_globals + st->b;
		op_c = pr->pr_globals + st->c;

		switch (st->op) {
			case OP_DIV_F_v6p:
				if ((OPA(int) & 0x80000000)
					^ (OPB(int) & 0x80000000))
					OPC(int) = 0xff7fffff;
				else
					OPC(int) = 0x7f7fffff;
				return 1;
			case OP_DIV_I_v6p:
				if (OPA(int) & 0x80000000)
					OPC(int) = -0x80000000;
				else
					OPC(int) = 0x7fffffff;
				return 1;
			case OP_MOD_I_v6p:
			case OP_MOD_F_v6p:
			case OP_REM_I_v6p:
			case OP_REM_F_v6p:
				OPC(int) = 0x00000000;
				return 1;
			default:
				break;
		}
	}
	PR_DumpState (pr);
	fflush (stdout);
	return 0;
}

static void
error_handler (void *data)
{
	progs_t    *pr = (progs_t *) data;
	PR_DumpState (pr);
	fflush (stdout);
}

VISIBLE int
PR_CallFunction (progs_t *pr, func_t fnum)
{
	bfunction_t *f;

	if (!fnum)
		PR_RunError (pr, "NULL function");
	f = pr->function_table + fnum;
	if (f->first_statement < 0) {
		// negative statements are built in functions
		if (pr->pr_trace && !pr->debug_handler) {
			Sys_Printf ("Calling builtin %s @ %p\n",
						PR_GetString (pr, f->descriptor->name), f->func);
		}
		f->func (pr);
		return 0;
	} else {
		PR_EnterFunction (pr, f);
		return 1;
	}
}

static void
check_stack_pointer (progs_t *pr, pointer_t stack, int size)
{
	if (stack < pr->stack_bottom) {
		PR_RunError (pr, "Progs stack overflow");
	}
	if (stack > pr->globals_size - size) {
		PR_RunError (pr, "Progs stack underflow");
	}
}

static inline void
pr_memset (pr_type_t *dst, int val, int count)
{
	while (count-- > 0) {
		(*dst++).integer_var = val;
	}
}

static void
pr_exec_quakec (progs_t *pr, int exitdepth)
{
	int         profile, startprofile;
	int         fldofs;
	pr_uint_t   pointer;
	dstatement_t *st;
	pr_type_t  *ptr;
	pr_type_t   old_val = {0};

	// make a stack frame
	startprofile = profile = 0;

	st = pr->pr_statements + pr->pr_xstatement;

	if (pr->watch) {
		old_val = *pr->watch;
	}
	while (1) {
		pr_type_t  *op_a, *op_b, *op_c;

		st++;
		++pr->pr_xstatement;
		if (pr->pr_xstatement != st - pr->pr_statements)
			PR_RunError (pr, "internal error");
		if (++profile > 1000000 && !pr->no_exec_limit) {
			PR_RunError (pr, "runaway loop error");
		}

		op_a = pr->pr_globals + st->a;
		op_b = pr->pr_globals + st->b;
		op_c = pr->pr_globals + st->c;

		if (pr->pr_trace) {
			if (pr->debug_handler) {
				pr->debug_handler (prd_trace, 0, pr->debug_data);
			} else {
				PR_PrintStatement (pr, st, 1);
			}
		}

		if (st->op & OP_BREAK) {
			if (pr->debug_handler) {
				pr->debug_handler (prd_breakpoint, 0, pr->debug_data);
			} else {
				PR_RunError (pr, "breakpoint hit");
			}
		}

		pr_opcode_v6p_e op = st->op & ~OP_BREAK;
		switch (op) {
			case OP_ADD_D_v6p:
				OPC(double) = OPA(double) + OPB(double);
				break;
			case OP_ADD_F_v6p:
				OPC(float) = OPA(float) + OPB(float);
				break;
			case OP_ADD_V_v6p:
				VectorAdd (&OPA(float), &OPB(float), &OPC(float));
				break;
			case OP_ADD_Q_v6p:
				QuatAdd (&OPA(float), &OPB(float), &OPC(float));
				break;
			case OP_ADD_S_v6p:
				OPC(string) = PR_CatStrings (pr,
												PR_GetString (pr,
															  OPA(string)),
												PR_GetString (pr,
															  OPB(string)));
				break;
			case OP_SUB_D_v6p:
				OPC(double) = OPA(double) - OPB(double);
				break;
			case OP_SUB_F_v6p:
				OPC(float) = OPA(float) - OPB(float);
				break;
			case OP_SUB_V_v6p:
				VectorSubtract (&OPA(float), &OPB(float),
								&OPC(float));
				break;
			case OP_SUB_Q_v6p:
				QuatSubtract (&OPA(float), &OPB(float), &OPC(float));
				break;
			case OP_MUL_D_v6p:
				OPC(double) = OPA(double) * OPB(double);
				break;
			case OP_MUL_F_v6p:
				OPC(float) = OPA(float) * OPB(float);
				break;
			case OP_MUL_V_v6p:
				OPC(float) = DotProduct (&OPA(float), &OPB(float));
				break;
			case OP_MUL_DV_v6p:
				{
					// avoid issues with the likes of x = x.x * x;
					// makes for faster code, too
					double      scale = OPA(double);
					VectorScale (&OPB(float), scale, &OPC(float));
				}
				break;
			case OP_MUL_VD_v6p:
				{
					// avoid issues with the likes of x = x * x.x;
					// makes for faster code, too
					double      scale = OPB(double);
					VectorScale (&OPA(float), scale, &OPC(float));
				}
				break;
			case OP_MUL_FV_v6p:
				{
					// avoid issues with the likes of x = x.x * x;
					// makes for faster code, too
					float       scale = OPA(float);
					VectorScale (&OPB(float), scale, &OPC(float));
				}
				break;
			case OP_MUL_VF_v6p:
				{
					// avoid issues with the likes of x = x * x.x;
					// makes for faster code, too
					float       scale = OPB(float);
					VectorScale (&OPA(float), scale, &OPC(float));
				}
				break;
			case OP_MUL_Q_v6p:
				QuatMult (&OPA(float), &OPB(float), &OPC(float));
				break;
			case OP_MUL_QV_v6p:
				QuatMultVec (&OPA(float), &OPB(float), &OPC(float));
				break;
			case OP_MUL_DQ_v6p:
				{
					// avoid issues with the likes of x = x.s * x;
					// makes for faster code, too
					double      scale = OPA(double);
					QuatScale (&OPB(float), scale, &OPC(float));
				}
				break;
			case OP_MUL_QD_v6p:
				{
					// avoid issues with the likes of x = x * x.s;
					// makes for faster code, too
					double      scale = OPB(double);
					QuatScale (&OPA(float), scale, &OPC(float));
				}
				break;
			case OP_MUL_FQ_v6p:
				{
					// avoid issues with the likes of x = x.s * x;
					// makes for faster code, too
					float       scale = OPA(float);
					QuatScale (&OPB(float), scale, &OPC(float));
				}
				break;
			case OP_MUL_QF_v6p:
				{
					// avoid issues with the likes of x = x * x.s;
					// makes for faster code, too
					float       scale = OPB(float);
					QuatScale (&OPA(float), scale, &OPC(float));
				}
				break;
			case OP_CONJ_Q_v6p:
				QuatConj (&OPA(float), &OPC(float));
				break;
			case OP_DIV_D_v6p:
				OPC(double) = OPA(double) / OPB(double);
				break;
			case OP_DIV_F_v6p:
				OPC(float) = OPA(float) / OPB(float);
				break;
			case OP_BITAND_v6p:
				OPC(float) = (int) OPA(float) & (int) OPB(float);
				break;
			case OP_BITOR_v6p:
				OPC(float) = (int) OPA(float) | (int) OPB(float);
				break;
			case OP_BITXOR_F_v6p:
				OPC(float) = (int) OPA(float) ^ (int) OPB(float);
				break;
			case OP_BITNOT_F_v6p:
				OPC(float) = ~ (int) OPA(float);
				break;
			case OP_SHL_F_v6p:
				OPC(float) = (int) OPA(float) << (int) OPB(float);
				break;
			case OP_SHR_F_v6p:
				OPC(float) = (int) OPA(float) >> (int) OPB(float);
				break;
			case OP_SHL_I_v6p:
				OPC(int) = OPA(int) << OPB(int);
				break;
			case OP_SHR_I_v6p:
				OPC(int) = OPA(int) >> OPB(int);
				break;
			case OP_SHR_U_v6p:
				OPC(uint) = OPA(uint) >> OPB(int);
				break;
			case OP_GE_F_v6p:
				OPC(float) = OPA(float) >= OPB(float);
				break;
			case OP_LE_F_v6p:
				OPC(float) = OPA(float) <= OPB(float);
				break;
			case OP_GT_F_v6p:
				OPC(float) = OPA(float) > OPB(float);
				break;
			case OP_LT_F_v6p:
				OPC(float) = OPA(float) < OPB(float);
				break;
			case OP_AND_v6p:	// OPA and OPB have to be float for -0.0
				OPC(int) = FNZ (OPA(uint)) && FNZ (OPB(uint));
				break;
			case OP_OR_v6p:		// OPA and OPB have to be float for -0.0
				OPC(int) = FNZ (OPA(uint)) || FNZ (OPB(uint));
				break;
			case OP_NOT_F_v6p:
				OPC(int) = !FNZ (OPA(uint));
				break;
			case OP_NOT_V_v6p:
				OPC(int) = VectorIsZero (&OPA(float));
				break;
			case OP_NOT_Q_v6p:
				OPC(int) = QuatIsZero (&OPA(float));
				break;
			case OP_NOT_S_v6p:
				OPC(int) = !OPA(string) ||
					!*PR_GetString (pr, OPA(string));
				break;
			case OP_NOT_FN_v6p:
				OPC(int) = !OPA(uint);
				break;
			case OP_NOT_ENT_v6p:
				OPC(int) = !OPA(uint);
				break;
			case OP_EQ_F_v6p:
				OPC(int) = OPA(float) == OPB(float);
				break;
			case OP_EQ_V_v6p:
				OPC(int) = VectorCompare (&OPA(float),
												 &OPB(float));
				break;
			case OP_EQ_Q_v6p:
				OPC(int) = QuatCompare (&OPA(float), &OPB(float));
				break;
			case OP_EQ_E_v6p:
				OPC(int) = OPA(int) == OPB(int);
				break;
			case OP_EQ_FN_v6p:
				OPC(int) = OPA(uint) == OPB(uint);
				break;
			case OP_NE_F_v6p:
				OPC(int) = OPA(float) != OPB(float);
				break;
			case OP_NE_V_v6p:
				OPC(int) = !VectorCompare (&OPA(float),
												  &OPB(float));
				break;
			case OP_NE_Q_v6p:
				OPC(int) = !QuatCompare (&OPA(float), &OPB(float));
				break;
			case OP_LE_S_v6p:
			case OP_GE_S_v6p:
			case OP_LT_S_v6p:
			case OP_GT_S_v6p:
			case OP_NE_S_v6p:
			case OP_EQ_S_v6p:
				{
					int cmp = strcmp (PR_GetString (pr, OPA(string)),
									  PR_GetString (pr, OPB(string)));
					switch (st->op) {
						case OP_LE_S_v6p: cmp = (cmp <= 0); break;
						case OP_GE_S_v6p: cmp = (cmp >= 0); break;
						case OP_LT_S_v6p: cmp = (cmp < 0); break;
						case OP_GT_S_v6p: cmp = (cmp > 0); break;
						case OP_NE_S_v6p: break;
						case OP_EQ_S_v6p: cmp = !cmp; break;
						default:      break;
					}
					OPC(int) = cmp;
				}
				break;
			case OP_NE_E_v6p:
				OPC(int) = OPA(int) != OPB(int);
				break;
			case OP_NE_FN_v6p:
				OPC(int) = OPA(uint) != OPB(uint);
				break;

			// ==================
			case OP_STORE_F_v6p:
			case OP_STORE_ENT_v6p:
			case OP_STORE_FLD_v6p:			// integers
			case OP_STORE_S_v6p:
			case OP_STORE_FN_v6p:			// pointers
			case OP_STORE_I_v6p:
			case OP_STORE_P_v6p:
				OPB(int) = OPA(int);
				break;
			case OP_STORE_V_v6p:
				VectorCopy (&OPA(float), &OPB(float));
				break;
			case OP_STORE_Q_v6p:
				QuatCopy (&OPA(float), &OPB(float));
				break;
			case OP_STORE_D_v6p:
				OPB(double) = OPA(double);
				break;

			case OP_STOREP_F_v6p:
			case OP_STOREP_ENT_v6p:
			case OP_STOREP_FLD_v6p:		// integers
			case OP_STOREP_S_v6p:
			case OP_STOREP_FN_v6p:		// pointers
			case OP_STOREP_I_v6p:
			case OP_STOREP_P_v6p:
				pointer = OPB(int);
				if (pr_boundscheck->int_val) {
					PR_BoundsCheck (pr, pointer, ev_integer);
				}
				ptr = pr->pr_globals + pointer;
				ptr->integer_var = OPA(int);
				break;
			case OP_STOREP_V_v6p:
				pointer = OPB(int);
				if (pr_boundscheck->int_val) {
					PR_BoundsCheck (pr, pointer, ev_vector);
				}
				ptr = pr->pr_globals + pointer;
				VectorCopy (&OPA(float), &ptr->vector_var);
				break;
			case OP_STOREP_Q_v6p:
				pointer = OPB(int);
				if (pr_boundscheck->int_val) {
					PR_BoundsCheck (pr, pointer, ev_quat);
				}
				ptr = pr->pr_globals + pointer;
				QuatCopy (&OPA(float), &ptr->quat_var);
				break;
			case OP_STOREP_D_v6p:
				pointer = OPB(int);
				if (pr_boundscheck->int_val) {
					PR_BoundsCheck (pr, pointer, ev_double);
				}
				ptr = pr->pr_globals + pointer;
				*(double *) ptr = OPA(double);
				break;

			case OP_ADDRESS_v6p:
				if (pr_boundscheck->int_val) {
					if (OPA(uint) >= pr->pr_edict_area_size)
						PR_RunError (pr, "Progs attempted to address an out "
									 "of bounds edict");
					if (OPA(uint) == 0 && pr->null_bad)
						PR_RunError (pr, "assignment to world entity");
					if (OPB(uint) >= pr->progs->entityfields)
						PR_RunError (pr, "Progs attempted to address an "
									 "invalid field in an edict");
				}
				fldofs = OPA(uint) + OPB(int);
				OPC(int) = &pr->pr_edict_area[fldofs] - pr->pr_globals;
				break;
			case OP_ADDRESS_VOID_v6p:
			case OP_ADDRESS_F_v6p:
			case OP_ADDRESS_V_v6p:
			case OP_ADDRESS_Q_v6p:
			case OP_ADDRESS_S_v6p:
			case OP_ADDRESS_ENT_v6p:
			case OP_ADDRESS_FLD_v6p:
			case OP_ADDRESS_FN_v6p:
			case OP_ADDRESS_I_v6p:
			case OP_ADDRESS_P_v6p:
			case OP_ADDRESS_D_v6p:
				OPC(int) = st->a;
				break;

			case OP_LOAD_F_v6p:
			case OP_LOAD_FLD_v6p:
			case OP_LOAD_ENT_v6p:
			case OP_LOAD_S_v6p:
			case OP_LOAD_FN_v6p:
			case OP_LOAD_I_v6p:
			case OP_LOAD_P_v6p:
				if (pr_boundscheck->int_val) {
					if (OPA(uint) >= pr->pr_edict_area_size)
						PR_RunError (pr, "Progs attempted to read an out of "
									 "bounds edict number");
					if (OPB(uint) >= pr->progs->entityfields)
						PR_RunError (pr, "Progs attempted to read an invalid "
									 "field in an edict");
				}
				fldofs = OPA(uint) + OPB(int);
				OPC(int) = pr->pr_edict_area[fldofs].integer_var;
				break;
			case OP_LOAD_V_v6p:
				if (pr_boundscheck->int_val) {
					if (OPA(uint) >= pr->pr_edict_area_size)
						PR_RunError (pr, "Progs attempted to read an out of "
									 "bounds edict number");
					if (OPB(uint) + 2 >= pr->progs->entityfields)
						PR_RunError (pr, "Progs attempted to read an invalid "
									 "field in an edict");
				}
				fldofs = OPA(uint) + OPB(int);
				memcpy (op_c, &pr->pr_edict_area[fldofs], 3 * sizeof (*op_c));
				break;
			case OP_LOAD_Q_v6p:
				if (pr_boundscheck->int_val) {
					if (OPA(uint) >= pr->pr_edict_area_size)
						PR_RunError (pr, "Progs attempted to read an out of "
									 "bounds edict number");
					if (OPB(uint) + 3 >= pr->progs->entityfields)
						PR_RunError (pr, "Progs attempted to read an invalid "
									 "field in an edict");
				}
				fldofs = OPA(uint) + OPB(int);
				memcpy (op_c, &pr->pr_edict_area[fldofs], 4 * sizeof (*op_c));
				break;
			case OP_LOAD_D_v6p:
				if (pr_boundscheck->int_val) {
					if (OPA(uint) >= pr->pr_edict_area_size)
						PR_RunError (pr, "Progs attempted to read an out of "
									 "bounds edict number");
					if (OPB(uint) + 1 >= pr->progs->entityfields)
						PR_RunError (pr, "Progs attempted to read an invalid "
									 "field in an edict");
				}
				fldofs = OPA(uint) + OPB(int);
				memcpy (op_c, &pr->pr_edict_area[fldofs], sizeof (double));
				break;

			case OP_LOADB_F_v6p:
			case OP_LOADB_S_v6p:
			case OP_LOADB_ENT_v6p:
			case OP_LOADB_FLD_v6p:
			case OP_LOADB_FN_v6p:
			case OP_LOADB_I_v6p:
			case OP_LOADB_P_v6p:
				pointer = OPA(int) + OPB(int);
				if (pr_boundscheck->int_val) {
					PR_BoundsCheck (pr, pointer, ev_integer);
				}
				ptr = pr->pr_globals + pointer;
				OPC(int) = ptr->integer_var;
				break;
			case OP_LOADB_V_v6p:
				pointer = OPA(int) + OPB(int);
				if (pr_boundscheck->int_val) {
					PR_BoundsCheck (pr, pointer, ev_vector);
				}
				ptr = pr->pr_globals + pointer;
				VectorCopy (&ptr->vector_var, &OPC(float));
				break;
			case OP_LOADB_Q_v6p:
				pointer = OPA(int) + OPB(int);
				if (pr_boundscheck->int_val) {
					PR_BoundsCheck (pr, pointer, ev_quat);
				}
				ptr = pr->pr_globals + pointer;
				QuatCopy (&ptr->quat_var, &OPC(float));
				break;
			case OP_LOADB_D_v6p:
				pointer = OPA(int) + OPB(int);
				if (pr_boundscheck->int_val) {
					PR_BoundsCheck (pr, pointer, ev_double);
				}
				ptr = pr->pr_globals + pointer;
				OPC(double) = *(double *) ptr;
				break;

			case OP_LOADBI_F_v6p:
			case OP_LOADBI_S_v6p:
			case OP_LOADBI_ENT_v6p:
			case OP_LOADBI_FLD_v6p:
			case OP_LOADBI_FN_v6p:
			case OP_LOADBI_I_v6p:
			case OP_LOADBI_P_v6p:
				pointer = OPA(int) + (short) st->b;
				if (pr_boundscheck->int_val) {
					PR_BoundsCheck (pr, pointer, ev_integer);
				}
				ptr = pr->pr_globals + pointer;
				OPC(int) = ptr->integer_var;
				break;
			case OP_LOADBI_V_v6p:
				pointer = OPA(int) + (short) st->b;
				if (pr_boundscheck->int_val) {
					PR_BoundsCheck (pr, pointer, ev_vector);
				}
				ptr = pr->pr_globals + pointer;
				VectorCopy (&ptr->vector_var, &OPC(float));
				break;
			case OP_LOADBI_Q_v6p:
				pointer = OPA(int) + (short) st->b;
				if (pr_boundscheck->int_val) {
					PR_BoundsCheck (pr, pointer, ev_quat);
				}
				ptr = pr->pr_globals + pointer;
				QuatCopy (&ptr->quat_var, &OPC(float));
				break;
			case OP_LOADBI_D_v6p:
				pointer = OPA(int) + (short) st->b;
				if (pr_boundscheck->int_val) {
					PR_BoundsCheck (pr, pointer, ev_quat);
				}
				ptr = pr->pr_globals + pointer;
				OPC(double) = *(double *) ptr;
				break;

			case OP_LEA_v6p:
				pointer = OPA(int) + OPB(int);
				OPC(int) = pointer;
				break;

			case OP_LEAI_v6p:
				pointer = OPA(int) + (short) st->b;
				OPC(int) = pointer;
				break;

			case OP_STOREB_F_v6p:
			case OP_STOREB_S_v6p:
			case OP_STOREB_ENT_v6p:
			case OP_STOREB_FLD_v6p:
			case OP_STOREB_FN_v6p:
			case OP_STOREB_I_v6p:
			case OP_STOREB_P_v6p:
				pointer = OPB(int) + OPC(int);
				if (pr_boundscheck->int_val) {
					PR_BoundsCheck (pr, pointer, ev_integer);
				}
				ptr = pr->pr_globals + pointer;
				ptr->integer_var = OPA(int);
				break;
			case OP_STOREB_V_v6p:
				pointer = OPB(int) + OPC(int);
				if (pr_boundscheck->int_val) {
					PR_BoundsCheck (pr, pointer, ev_vector);
				}
				ptr = pr->pr_globals + pointer;
				VectorCopy (&OPA(float), &ptr->vector_var);
				break;
			case OP_STOREB_Q_v6p:
				pointer = OPB(int) + OPC(int);
				if (pr_boundscheck->int_val) {
					PR_BoundsCheck (pr, pointer, ev_quat);
				}
				ptr = pr->pr_globals + pointer;
				QuatCopy (&OPA(float), &ptr->quat_var);
				break;
			case OP_STOREB_D_v6p:
				pointer = OPB(int) + OPC(int);
				if (pr_boundscheck->int_val) {
					PR_BoundsCheck (pr, pointer, ev_quat);
				}
				ptr = pr->pr_globals + pointer;
				*(double *) ptr = OPA(double);
				break;

			case OP_STOREBI_F_v6p:
			case OP_STOREBI_S_v6p:
			case OP_STOREBI_ENT_v6p:
			case OP_STOREBI_FLD_v6p:
			case OP_STOREBI_FN_v6p:
			case OP_STOREBI_I_v6p:
			case OP_STOREBI_P_v6p:
				pointer = OPB(int) + (short) st->c;
				if (pr_boundscheck->int_val) {
					PR_BoundsCheck (pr, pointer, ev_integer);
				}
				ptr = pr->pr_globals + pointer;
				ptr->integer_var = OPA(int);
				break;
			case OP_STOREBI_V_v6p:
				pointer = OPB(int) + (short) st->c;
				if (pr_boundscheck->int_val) {
					PR_BoundsCheck (pr, pointer, ev_vector);
				}
				ptr = pr->pr_globals + pointer;
				VectorCopy (&OPA(float), &ptr->vector_var);
				break;
			case OP_STOREBI_Q_v6p:
				pointer = OPB(int) + (short) st->c;
				if (pr_boundscheck->int_val) {
					PR_BoundsCheck (pr, pointer, ev_quat);
				}
				ptr = pr->pr_globals + pointer;
				QuatCopy (&OPA(float), &ptr->quat_var);
				break;
			case OP_STOREBI_D_v6p:
				pointer = OPB(int) + (short) st->c;
				if (pr_boundscheck->int_val) {
					PR_BoundsCheck (pr, pointer, ev_quat);
				}
				ptr = pr->pr_globals + pointer;
				*(double *) ptr = OPA(double);
				break;

			case OP_PUSH_F_v6p:
			case OP_PUSH_FLD_v6p:
			case OP_PUSH_ENT_v6p:
			case OP_PUSH_S_v6p:
			case OP_PUSH_FN_v6p:
			case OP_PUSH_I_v6p:
			case OP_PUSH_P_v6p:
				{
					pointer_t   stack = *pr->globals.stack - 1;
					pr_type_t  *stk = pr->pr_globals + stack;
					if (pr_boundscheck->int_val) {
						check_stack_pointer (pr, stack, 1);
					}
					stk->integer_var = OPA(int);
					*pr->globals.stack = stack;
				}
				break;
			case OP_PUSH_V_v6p:
				{
					pointer_t   stack = *pr->globals.stack - 3;
					pr_type_t  *stk = pr->pr_globals + stack;
					if (pr_boundscheck->int_val) {
						check_stack_pointer (pr, stack, 3);
					}
					memcpy (stk, op_a, 3 * sizeof (*op_c));
					*pr->globals.stack = stack;
				}
				break;
			case OP_PUSH_Q_v6p:
				{
					pointer_t   stack = *pr->globals.stack - 4;
					pr_type_t  *stk = pr->pr_globals + stack;
					if (pr_boundscheck->int_val) {
						check_stack_pointer (pr, stack, 4);
					}
					memcpy (stk, op_a, 4 * sizeof (*op_c));
					*pr->globals.stack = stack;
				}
				break;

			case OP_PUSHB_F_v6p:
			case OP_PUSHB_S_v6p:
			case OP_PUSHB_ENT_v6p:
			case OP_PUSHB_FLD_v6p:
			case OP_PUSHB_FN_v6p:
			case OP_PUSHB_I_v6p:
			case OP_PUSHB_P_v6p:
				{
					pointer_t   stack = *pr->globals.stack - 1;
					pr_type_t  *stk = pr->pr_globals + stack;

					pointer = OPA(int) + OPB(int);
					ptr = pr->pr_globals + pointer;

					if (pr_boundscheck->int_val) {
						check_stack_pointer (pr, stack, 1);
						PR_BoundsCheck (pr, pointer, ev_integer);
					}

					stk->integer_var = ptr->integer_var;
					*pr->globals.stack = stack;
				}
				break;
			case OP_PUSHB_V_v6p:
				{
					pointer_t   stack = *pr->globals.stack - 3;
					pr_type_t  *stk = pr->pr_globals + stack;

					pointer = OPA(int) + OPB(int);
					ptr = pr->pr_globals + pointer;

					if (pr_boundscheck->int_val) {
						check_stack_pointer (pr, stack, 3);
						PR_BoundsCheck (pr, pointer, ev_integer);
					}

					VectorCopy (&ptr->vector_var, &stk->vector_var);
					*pr->globals.stack = stack;
				}
				break;
			case OP_PUSHB_Q_v6p:
				{
					pointer_t   stack = *pr->globals.stack - 4;
					pr_type_t  *stk = pr->pr_globals + stack;

					pointer = OPA(int) + OPB(int);
					ptr = pr->pr_globals + pointer;

					if (pr_boundscheck->int_val) {
						check_stack_pointer (pr, stack, 4);
						PR_BoundsCheck (pr, pointer, ev_quat);
					}

					QuatCopy (&ptr->quat_var, &stk->quat_var);
					*pr->globals.stack = stack;
				}
				break;

			case OP_PUSHBI_F_v6p:
			case OP_PUSHBI_S_v6p:
			case OP_PUSHBI_ENT_v6p:
			case OP_PUSHBI_FLD_v6p:
			case OP_PUSHBI_FN_v6p:
			case OP_PUSHBI_I_v6p:
			case OP_PUSHBI_P_v6p:
				{
					pointer_t   stack = *pr->globals.stack - 1;
					pr_type_t  *stk = pr->pr_globals + stack;

					pointer = OPA(int) + st->b;
					ptr = pr->pr_globals + pointer;

					if (pr_boundscheck->int_val) {
						check_stack_pointer (pr, stack, 1);
						PR_BoundsCheck (pr, pointer, ev_integer);
					}

					stk->integer_var = ptr->integer_var;
					*pr->globals.stack = stack;
				}
				break;
			case OP_PUSHBI_V_v6p:
				{
					pointer_t   stack = *pr->globals.stack - 3;
					pr_type_t  *stk = pr->pr_globals + stack;

					pointer = OPA(int) + st->b;
					ptr = pr->pr_globals + pointer;

					if (pr_boundscheck->int_val) {
						check_stack_pointer (pr, stack, 3);
						PR_BoundsCheck (pr, pointer, ev_integer);
					}

					VectorCopy (&ptr->vector_var, &stk->vector_var);
					*pr->globals.stack = stack;
				}
				break;
			case OP_PUSHBI_Q_v6p:
				{
					pointer_t   stack = *pr->globals.stack - 4;
					pr_type_t  *stk = pr->pr_globals + stack;

					pointer = OPA(int) + st->b;
					ptr = pr->pr_globals + pointer;

					if (pr_boundscheck->int_val) {
						check_stack_pointer (pr, stack, 4);
						PR_BoundsCheck (pr, pointer, ev_quat);
					}

					QuatCopy (&ptr->quat_var, &stk->quat_var);
					*pr->globals.stack = stack;
				}
				break;

			case OP_POP_F_v6p:
			case OP_POP_FLD_v6p:
			case OP_POP_ENT_v6p:
			case OP_POP_S_v6p:
			case OP_POP_FN_v6p:
			case OP_POP_I_v6p:
			case OP_POP_P_v6p:
				{
					pointer_t   stack = *pr->globals.stack;
					pr_type_t  *stk = pr->pr_globals + stack;
					if (pr_boundscheck->int_val) {
						check_stack_pointer (pr, stack, 1);
					}
					OPA(int) = stk->integer_var;
					*pr->globals.stack = stack + 1;
				}
				break;
			case OP_POP_V_v6p:
				{
					pointer_t   stack = *pr->globals.stack;
					pr_type_t  *stk = pr->pr_globals + stack;
					if (pr_boundscheck->int_val) {
						check_stack_pointer (pr, stack, 3);
					}
					memcpy (op_a, stk, 3 * sizeof (*op_c));
					*pr->globals.stack = stack + 3;
				}
				break;
			case OP_POP_Q_v6p:
				{
					pointer_t   stack = *pr->globals.stack;
					pr_type_t  *stk = pr->pr_globals + stack;
					if (pr_boundscheck->int_val) {
						check_stack_pointer (pr, stack, 4);
					}
					memcpy (op_a, stk, 4 * sizeof (*op_c));
					*pr->globals.stack = stack + 4;
				}
				break;

			case OP_POPB_F_v6p:
			case OP_POPB_S_v6p:
			case OP_POPB_ENT_v6p:
			case OP_POPB_FLD_v6p:
			case OP_POPB_FN_v6p:
			case OP_POPB_I_v6p:
			case OP_POPB_P_v6p:
				{
					pointer_t   stack = *pr->globals.stack;
					pr_type_t  *stk = pr->pr_globals + stack;

					pointer = OPA(int) + OPB(int);
					ptr = pr->pr_globals + pointer;

					if (pr_boundscheck->int_val) {
						check_stack_pointer (pr, stack, 1);
						PR_BoundsCheck (pr, pointer, ev_integer);
					}

					ptr->integer_var = stk->integer_var;
					*pr->globals.stack = stack + 1;
				}
				break;
			case OP_POPB_V_v6p:
				{
					pointer_t   stack = *pr->globals.stack;
					pr_type_t  *stk = pr->pr_globals + stack;

					pointer = OPA(int) + OPB(int);
					ptr = pr->pr_globals + pointer;

					if (pr_boundscheck->int_val) {
						check_stack_pointer (pr, stack, 3);
						PR_BoundsCheck (pr, pointer, ev_integer);
					}

					VectorCopy (&stk->vector_var, &ptr->vector_var);
					*pr->globals.stack = stack + 3;
				}
				break;
			case OP_POPB_Q_v6p:
				{
					pointer_t   stack = *pr->globals.stack;
					pr_type_t  *stk = pr->pr_globals + stack;

					pointer = OPA(int) + OPB(int);
					ptr = pr->pr_globals + pointer;

					if (pr_boundscheck->int_val) {
						check_stack_pointer (pr, stack, 4);
						PR_BoundsCheck (pr, pointer, ev_quat);
					}

					QuatCopy (&stk->quat_var, &ptr->quat_var);
					*pr->globals.stack = stack + 4;
				}
				break;

			case OP_POPBI_F_v6p:
			case OP_POPBI_S_v6p:
			case OP_POPBI_ENT_v6p:
			case OP_POPBI_FLD_v6p:
			case OP_POPBI_FN_v6p:
			case OP_POPBI_I_v6p:
			case OP_POPBI_P_v6p:
				{
					pointer_t   stack = *pr->globals.stack;
					pr_type_t  *stk = pr->pr_globals + stack;

					pointer = OPA(int) + st->b;
					ptr = pr->pr_globals + pointer;

					if (pr_boundscheck->int_val) {
						check_stack_pointer (pr, stack, 1);
						PR_BoundsCheck (pr, pointer, ev_integer);
					}

					ptr->integer_var = stk->integer_var;
					*pr->globals.stack = stack + 1;
				}
				break;
			case OP_POPBI_V_v6p:
				{
					pointer_t   stack = *pr->globals.stack;
					pr_type_t  *stk = pr->pr_globals + stack;

					pointer = OPA(int) + st->b;
					ptr = pr->pr_globals + pointer;

					if (pr_boundscheck->int_val) {
						check_stack_pointer (pr, stack, 3);
						PR_BoundsCheck (pr, pointer, ev_integer);
					}

					VectorCopy (&stk->vector_var, &ptr->vector_var);
					*pr->globals.stack = stack + 3;
				}
				break;
			case OP_POPBI_Q_v6p:
				{
					pointer_t   stack = *pr->globals.stack;
					pr_type_t  *stk = pr->pr_globals + stack;

					pointer = OPA(int) + st->b;
					ptr = pr->pr_globals + pointer;

					if (pr_boundscheck->int_val) {
						check_stack_pointer (pr, stack, 4);
						PR_BoundsCheck (pr, pointer, ev_quat);
					}

					QuatCopy (&stk->quat_var, &ptr->quat_var);
					*pr->globals.stack = stack + 4;
				}
				break;

			// ==================
			case OP_IFNOT_v6p:
				if (!OPA(int)) {
					pr->pr_xstatement += (short)st->b - 1;	// offset the st++
					st = pr->pr_statements + pr->pr_xstatement;
				}
				break;
			case OP_IF_v6p:
				if (OPA(int)) {
					pr->pr_xstatement += (short)st->b - 1;	// offset the st++
					st = pr->pr_statements + pr->pr_xstatement;
				}
				break;
			case OP_IFBE_v6p:
				if (OPA(int) <= 0) {
					pr->pr_xstatement += (short)st->b - 1;	// offset the st++
					st = pr->pr_statements + pr->pr_xstatement;
				}
				break;
			case OP_IFB_v6p:
				if (OPA(int) < 0) {
					pr->pr_xstatement += (short)st->b - 1;	// offset the st++
					st = pr->pr_statements + pr->pr_xstatement;
				}
				break;
			case OP_IFAE_v6p:
				if (OPA(int) >= 0) {
					pr->pr_xstatement += (short)st->b - 1;	// offset the st++
					st = pr->pr_statements + pr->pr_xstatement;
				}
				break;
			case OP_IFA_v6p:
				if (OPA(int) > 0) {
					pr->pr_xstatement += (short)st->b - 1;	// offset the st++
					st = pr->pr_statements + pr->pr_xstatement;
				}
				break;
			case OP_GOTO_v6p:
				pr->pr_xstatement += (short)st->a - 1;		// offset the st++
				st = pr->pr_statements + pr->pr_xstatement;
				break;
			case OP_JUMP_v6p:
				if (pr_boundscheck->int_val
					&& (OPA(uint) >= pr->progs->numstatements)) {
					PR_RunError (pr, "Invalid jump destination");
				}
				pr->pr_xstatement = OPA(uint) - 1;	// offset the st++
				st = pr->pr_statements + pr->pr_xstatement;
				break;
			case OP_JUMPB_v6p:
				pointer = st->a + OPB(int);
				if (pr_boundscheck->int_val) {
					PR_BoundsCheck (pr, pointer, ev_integer);
				}
				ptr = pr->pr_globals + pointer;
				pointer = ptr->integer_var;
				if (pr_boundscheck->int_val
					&& (pointer >= pr->progs->numstatements)) {
					PR_RunError (pr, "Invalid jump destination");
				}
				pr->pr_xstatement = pointer - 1;			// offset the st++
				st = pr->pr_statements + pr->pr_xstatement;
				break;

			case OP_RCALL2_v6p:
			case OP_RCALL3_v6p:
			case OP_RCALL4_v6p:
			case OP_RCALL5_v6p:
			case OP_RCALL6_v6p:
			case OP_RCALL7_v6p:
			case OP_RCALL8_v6p:
				pr->pr_params[1] = op_c;
				goto op_rcall;
			case OP_RCALL1_v6p:
				pr->pr_params[1] = pr->pr_real_params[1];
op_rcall:
				pr->pr_params[0] = op_b;
				pr->pr_argc = st->op - OP_RCALL1_v6p + 1;
				goto op_call;
			case OP_CALL0_v6p:
			case OP_CALL1_v6p:
			case OP_CALL2_v6p:
			case OP_CALL3_v6p:
			case OP_CALL4_v6p:
			case OP_CALL5_v6p:
			case OP_CALL6_v6p:
			case OP_CALL7_v6p:
			case OP_CALL8_v6p:
				PR_RESET_PARAMS (pr);
				pr->pr_argc = st->op - OP_CALL0_v6p;
op_call:
				pr->pr_xfunction->profile += profile - startprofile;
				startprofile = profile;
				PR_CallFunction (pr, OPA(uint));
				st = pr->pr_statements + pr->pr_xstatement;
				break;
			case OP_DONE_v6p:
			case OP_RETURN_v6p:
				if (!st->a)
					memset (&R_INT (pr), 0,
							pr->pr_param_size * sizeof (*op_a));
				else if (&R_INT (pr) != &OPA(int))
					memcpy (&R_INT (pr), op_a,
							pr->pr_param_size * sizeof (*op_a));
				// fallthrough
			case OP_RETURN_V_v6p:
				pr->pr_xfunction->profile += profile - startprofile;
				startprofile = profile;
				PR_LeaveFunction (pr, pr->pr_depth == exitdepth);
				st = pr->pr_statements + pr->pr_xstatement;
				if (pr->pr_depth == exitdepth) {
					if (pr->pr_trace && pr->pr_depth <= pr->pr_trace_depth)
						pr->pr_trace = false;
					// all done
					goto exit_program;
				}
				break;
			case OP_STATE_v6p:
				{
					int         self = *pr->globals.self;
					int         nextthink = pr->fields.nextthink + self;
					int         frame = pr->fields.frame + self;
					int         think = pr->fields.think + self;
					float       time = *pr->globals.time + 0.1;
					pr->pr_edict_area[nextthink].float_var = time;
					pr->pr_edict_area[frame].float_var = OPA(float);
					pr->pr_edict_area[think].func_var = OPB(uint);
				}
				break;
			case OP_STATE_F_v6p:
				{
					int         self = *pr->globals.self;
					int         nextthink = pr->fields.nextthink + self;
					int         frame = pr->fields.frame + self;
					int         think = pr->fields.think + self;
					float       time = *pr->globals.time + OPC(float);
					pr->pr_edict_area[nextthink].float_var = time;
					pr->pr_edict_area[frame].float_var = OPA(float);
					pr->pr_edict_area[think].func_var = OPB(uint);
				}
				break;
			case OP_ADD_I_v6p:
				OPC(int) = OPA(int) + OPB(int);
				break;
			case OP_SUB_I_v6p:
				OPC(int) = OPA(int) - OPB(int);
				break;
			case OP_MUL_I_v6p:
				OPC(int) = OPA(int) * OPB(int);
				break;
			case OP_DIV_I_v6p:
				OPC(int) = OPA(int) / OPB(int);
				break;
			case OP_MOD_I_v6p:
				{
					// implement true modulo for integers:
					//  5 mod  3 = 2
					// -5 mod  3 = 1
					//  5 mod -3 = -1
					// -5 mod -3 = -2
					int         a = OPA(int);
					int         b = OPB(int);
					int         c = a % b;
					// % is really remainder and so has the same sign rules
					// as division: -5 % 3 = -2, so need to add b (3 here)
					// if c's sign is incorrect, but only if c is non-zero
					int         mask = (a ^ b) >> 31;
					mask &= ~(!!c + 0) + 1;	// +0 to convert bool to int (gcc)
					OPC(int) = c + (mask & b);
				}
				break;
			case OP_REM_I_v6p:
				OPC(int) = OPA(int) % OPB(int);
				break;
			case OP_MOD_D_v6p:
				{
					double      a = OPA(double);
					double      b = OPB(double);
					// floating point modulo is so much easier :P
					OPC(double) = a - b * floor (a / b);
				}
				break;
			case OP_REM_D_v6p:
				{
					double      a = OPA(double);
					double      b = OPB(double);
					OPC(double) = a - b * trunc (a / b);
				}
				break;
			case OP_MOD_F_v6p:
				{
					float       a = OPA(float);
					float       b = OPB(float);
					OPC(float) = a - b * floorf (a / b);
				}
				break;
			case OP_REM_F_v6p:
				{
					float       a = OPA(float);
					float       b = OPB(float);
					OPC(float) = a - b * truncf (a / b);
				}
				break;
			case OP_CONV_IF_v6p:
				OPC(float) = OPA(int);
				break;
			case OP_CONV_FI_v6p:
				OPC(int) = OPA(float);
				break;
			case OP_BITAND_I_v6p:
				OPC(int) = OPA(int) & OPB(int);
				break;
			case OP_BITOR_I_v6p:
				OPC(int) = OPA(int) | OPB(int);
				break;
			case OP_BITXOR_I_v6p:
				OPC(int) = OPA(int) ^ OPB(int);
				break;
			case OP_BITNOT_I_v6p:
				OPC(int) = ~OPA(int);
				break;

			case OP_GE_I_v6p:
			case OP_GE_P_v6p:
				OPC(int) = OPA(int) >= OPB(int);
				break;
			case OP_GE_U_v6p:
				OPC(int) = OPA(uint) >= OPB(uint);
				break;
			case OP_LE_I_v6p:
			case OP_LE_P_v6p:
				OPC(int) = OPA(int) <= OPB(int);
				break;
			case OP_LE_U_v6p:
				OPC(int) = OPA(uint) <= OPB(uint);
				break;
			case OP_GT_I_v6p:
			case OP_GT_P_v6p:
				OPC(int) = OPA(int) > OPB(int);
				break;
			case OP_GT_U_v6p:
				OPC(int) = OPA(uint) > OPB(uint);
				break;
			case OP_LT_I_v6p:
			case OP_LT_P_v6p:
				OPC(int) = OPA(int) < OPB(int);
				break;
			case OP_LT_U_v6p:
				OPC(int) = OPA(uint) < OPB(uint);
				break;

			case OP_AND_I_v6p:
				OPC(int) = OPA(int) && OPB(int);
				break;
			case OP_OR_I_v6p:
				OPC(int) = OPA(int) || OPB(int);
				break;
			case OP_NOT_I_v6p:
			case OP_NOT_P_v6p:
				OPC(int) = !OPA(int);
				break;
			case OP_EQ_I_v6p:
			case OP_EQ_P_v6p:
				OPC(int) = OPA(int) == OPB(int);
				break;
			case OP_NE_I_v6p:
			case OP_NE_P_v6p:
				OPC(int) = OPA(int) != OPB(int);
				break;

			case OP_MOVEI_v6p:
				memmove (op_c, op_a, st->b * 4);
				break;
			case OP_MOVEP_v6p:
				if (pr_boundscheck->int_val) {
					PR_BoundsCheckSize (pr, OPC(int), OPB(uint));
					PR_BoundsCheckSize (pr, OPA(int), OPB(uint));
				}
				memmove (pr->pr_globals + OPC(int),
						 pr->pr_globals + OPA(int),
						 OPB(uint) * 4);
				break;
			case OP_MOVEPI_v6p:
				if (pr_boundscheck->int_val) {
					PR_BoundsCheckSize (pr, OPC(int), st->b);
					PR_BoundsCheckSize (pr, OPA(int), st->b);
				}
				memmove (pr->pr_globals + OPC(int),
						 pr->pr_globals + OPA(int),
						 st->b * 4);
				break;
			case OP_MEMSETI_v6p:
				pr_memset (op_c, OPA(int), st->b);
				break;
			case OP_MEMSETP_v6p:
				if (pr_boundscheck->int_val) {
					PR_BoundsCheckSize (pr, OPC(pointer), OPB(int));
				}
				pr_memset (pr->pr_globals + OPC(pointer), OPA(int),
						   OPB(int));
				break;
			case OP_MEMSETPI_v6p:
				if (pr_boundscheck->int_val) {
					PR_BoundsCheckSize (pr, OPC(pointer), st->b);
				}
				pr_memset (pr->pr_globals + OPC(pointer), OPA(int),
						   st->b);
				break;
			case OP_GE_D_v6p:
				OPC(float) = OPA(double) >= OPB(double);
				break;
			case OP_LE_D_v6p:
				OPC(float) = OPA(double) <= OPB(double);
				break;
			case OP_GT_D_v6p:
				OPC(float) = OPA(double) > OPB(double);
				break;
			case OP_LT_D_v6p:
				OPC(float) = OPA(double) < OPB(double);
				break;
			case OP_NOT_D_v6p:
				OPC(int) = (op_a[0].integer_var
								   || (op_a[1].integer_var & ~0x80000000u));
				break;
			case OP_EQ_D_v6p:
				OPC(int) = OPA(double) == OPB(double);
				break;
			case OP_NE_D_v6p:
				OPC(int) = OPA(double) != OPB(double);
				break;
			case OP_CONV_ID_v6p:
				OPC(double) = OPA(int);
				break;
			case OP_CONV_DI_v6p:
				OPC(int) = OPA(double);
				break;
			case OP_CONV_FD_v6p:
				OPC(double) = OPA(float);
				break;
			case OP_CONV_DF_v6p:
				OPC(float) = OPA(double);
				break;

// LordHavoc: to be enabled when Progs version 7 (or whatever it will be numbered) is finalized
/*
			case OP_BOUNDCHECK_v6p:
				if (OPA(int) < 0 || OPA(int) >= st->b) {
					PR_RunError (pr, "Progs boundcheck failed at line number "
					"%d, value is < 0 or >= %d", st->b, st->c);
				}
				break;

*/
			default:
				PR_RunError (pr, "Bad opcode %i", st->op & ~OP_BREAK);
		}
		if (pr->watch && pr->watch->integer_var != old_val.integer_var) {
			if (!pr->wp_conditional
				|| pr->watch->integer_var == pr->wp_val.integer_var) {
				if (pr->debug_handler) {
					pr->debug_handler (prd_watchpoint, 0, pr->debug_data);
				} else {
					PR_RunError (pr, "watchpoint hit: %d -> %d",
								 old_val.integer_var, pr->watch->integer_var);
				}
			}
			old_val.integer_var = pr->watch->integer_var;
		}
	}
exit_program:
}

#define MM(type) (*((pr_##type##_t *) (mm)))
#define STK(type) (*((pr_##type##_t *) (stk)))

static pr_type_t *
pr_entity_mode (progs_t *pr, const dstatement_t *st, int shift)
{
	pr_type_t  *op_a = pr->pr_globals + st->a + PR_BASE (pr, st, A);
	pr_type_t  *op_b = pr->pr_globals + st->b + PR_BASE (pr, st, B);
	int         mm_ind = (st->op >> shift) & 3;
	pointer_t   edict_area = pr->pr_edict_area - pr->pr_globals;
	pointer_t   mm_offs = 0;

	switch (mm_ind) {
		case 0:
			// entity.field (equivalent to OP_LOAD_t_v6p)
			mm_offs = edict_area + OPA(uint) + OPB(uint);
			break;
		case 1:
			// simple pointer dereference: *a
			mm_offs = OPA(uint);
			break;
		case 2:
			// constant indexed pointer: *a + b (supports -ve offset)
			mm_offs = OPA(uint) + (short) st->b;
			break;
		case 3:
			// variable indexed pointer: *a + *b (supports -ve offset)
			mm_offs = OPA(uint) + OPB(int);
			break;
	}
	return pr->pr_globals + mm_offs;
}

static pr_type_t *
pr_address_mode (progs_t *pr, const dstatement_t *st, int shift)
{
	pr_type_t  *op_a = pr->pr_globals + st->a + PR_BASE (pr, st, A);
	pr_type_t  *op_b = pr->pr_globals + st->b + PR_BASE (pr, st, B);
	int         mm_ind = (st->op >> shift) & 3;
	pointer_t   mm_offs = 0;

	switch (mm_ind) {
		case 0:
			// regular global access
			mm_offs = op_a - pr->pr_globals;
			break;
		case 1:
			// simple pointer dereference: *a
			mm_offs = OPA(uint);
			break;
		case 2:
			// constant indexed pointer: *a + b (supports -ve offset)
			mm_offs = OPA(uint) + (short) st->b;
			break;
		case 3:
			// variable indexed pointer: *a + *b (supports -ve offset)
			mm_offs = OPA(uint) + OPB(int);
			break;
	}
	return pr->pr_globals + mm_offs;
}

static pr_pointer_t __attribute__((pure))
pr_jump_mode (progs_t *pr, const dstatement_t *st)
{
	pr_type_t  *op_a = pr->pr_globals + st->a + PR_BASE (pr, st, A);
	pr_type_t  *op_b = pr->pr_globals + st->b + PR_BASE (pr, st, B);
	int         jump_ind = st->op & 3;
	pointer_t   jump_offs = pr->pr_xstatement;

	switch (jump_ind) {
		case 0:
			// instruction relative offset
			jump_offs = jump_offs + (short) st->a;
			break;
		case 1:
			// simple pointer dereference: *a
			jump_offs = OPA(uint);
			break;
		case 2:
			// constant indexed pointer: *a + b (supports -ve offset)
			jump_offs = OPA(uint) + (short) st->b;
			break;
		case 3:
			// variable indexed pointer: *a + *b (supports -ve offset)
			jump_offs = OPA(uint) + OPB(int);
			break;
	}
	if (pr_boundscheck->int_val && jump_offs >= pr->progs->numstatements) {
		PR_RunError (pr, "out of bounds: %x", jump_offs);
	}
	return jump_offs - 1;	// for st++
}

static pr_pointer_t __attribute__((pure))
pr_with (progs_t *pr, const dstatement_t *st)
{
	pointer_t   edict_area = pr->pr_edict_area - pr->pr_globals;
	pr_type_t  *op_b = pr->pr_globals + PR_BASE (pr, st, B) + st->b;

	switch (st->a) {
		// fixed offset
		case 0:
			// hard-0 base
			return st->b;
		case 1:
			// relative to current base
			return PR_BASE (pr, st, B) + st->b;
		case 2:
			// relative to stack (-ve offset)
			return *pr->globals.stack + (pr_short_t) st->b;
		case 3:
			// relative to edict_area (+ve only)
			return edict_area + st->b;

		case 4:
			// hard-0 base
			return pr->pr_globals[st->b].pointer_var;
		case 5:
			return OPB(pointer);
		case 6:
			// relative to stack (-ve offset)
			return *pr->globals.stack + OPB(int);
		case 7:
			// relative to edict_area (+ve only)
			return edict_area + OPB(uint);
	}
	PR_RunError (pr, "Invalid with index: %u", st->a);
}

static pr_type_t *
pr_stack_push (progs_t *pr)
{
	// keep the stack 16-byte aligned
	pointer_t   stack = *pr->globals.stack - 4;
	pr_type_t  *stk = pr->pr_globals + stack;
	if (pr_boundscheck->int_val) {
		check_stack_pointer (pr, stack, 4);
	}
	*pr->globals.stack = stack;
	return stk;
}

static pr_type_t *
pr_stack_pop (progs_t *pr)
{
	pointer_t   stack = *pr->globals.stack;
	pr_type_t  *stk = pr->pr_globals + stack;
	if (pr_boundscheck->int_val) {
		check_stack_pointer (pr, stack, 4);
	}
	// keep the stack 16-byte aligned
	*pr->globals.stack = stack + 4;
	return stk;
}

static pr_ivec4_t
pr_swizzle_f (pr_ivec4_t vec, pr_ushort_t swiz)
{
	goto do_swizzle;
#define swizzle __builtin_shuffle
	swizzle_xxxx: vec = swizzle (vec, (pr_ivec4_t) { 0, 0, 0, 0 }); goto negate;
	swizzle_yxxx: vec = swizzle (vec, (pr_ivec4_t) { 1, 0, 0, 0 }); goto negate;
	swizzle_zxxx: vec = swizzle (vec, (pr_ivec4_t) { 2, 0, 0, 0 }); goto negate;
	swizzle_wxxx: vec = swizzle (vec, (pr_ivec4_t) { 3, 0, 0, 0 }); goto negate;
	swizzle_xyxx: vec = swizzle (vec, (pr_ivec4_t) { 0, 1, 0, 0 }); goto negate;
	swizzle_yyxx: vec = swizzle (vec, (pr_ivec4_t) { 1, 1, 0, 0 }); goto negate;
	swizzle_zyxx: vec = swizzle (vec, (pr_ivec4_t) { 2, 1, 0, 0 }); goto negate;
	swizzle_wyxx: vec = swizzle (vec, (pr_ivec4_t) { 3, 1, 0, 0 }); goto negate;
	swizzle_xzxx: vec = swizzle (vec, (pr_ivec4_t) { 0, 2, 0, 0 }); goto negate;
	swizzle_yzxx: vec = swizzle (vec, (pr_ivec4_t) { 1, 2, 0, 0 }); goto negate;
	swizzle_zzxx: vec = swizzle (vec, (pr_ivec4_t) { 2, 2, 0, 0 }); goto negate;
	swizzle_wzxx: vec = swizzle (vec, (pr_ivec4_t) { 3, 2, 0, 0 }); goto negate;
	swizzle_xwxx: vec = swizzle (vec, (pr_ivec4_t) { 0, 3, 0, 0 }); goto negate;
	swizzle_ywxx: vec = swizzle (vec, (pr_ivec4_t) { 1, 3, 0, 0 }); goto negate;
	swizzle_zwxx: vec = swizzle (vec, (pr_ivec4_t) { 2, 3, 0, 0 }); goto negate;
	swizzle_wwxx: vec = swizzle (vec, (pr_ivec4_t) { 3, 3, 0, 0 }); goto negate;
	swizzle_xxyx: vec = swizzle (vec, (pr_ivec4_t) { 0, 0, 1, 0 }); goto negate;
	swizzle_yxyx: vec = swizzle (vec, (pr_ivec4_t) { 1, 0, 1, 0 }); goto negate;
	swizzle_zxyx: vec = swizzle (vec, (pr_ivec4_t) { 2, 0, 1, 0 }); goto negate;
	swizzle_wxyx: vec = swizzle (vec, (pr_ivec4_t) { 3, 0, 1, 0 }); goto negate;
	swizzle_xyyx: vec = swizzle (vec, (pr_ivec4_t) { 0, 1, 1, 0 }); goto negate;
	swizzle_yyyx: vec = swizzle (vec, (pr_ivec4_t) { 1, 1, 1, 0 }); goto negate;
	swizzle_zyyx: vec = swizzle (vec, (pr_ivec4_t) { 2, 1, 1, 0 }); goto negate;
	swizzle_wyyx: vec = swizzle (vec, (pr_ivec4_t) { 3, 1, 1, 0 }); goto negate;
	swizzle_xzyx: vec = swizzle (vec, (pr_ivec4_t) { 0, 2, 1, 0 }); goto negate;
	swizzle_yzyx: vec = swizzle (vec, (pr_ivec4_t) { 1, 2, 1, 0 }); goto negate;
	swizzle_zzyx: vec = swizzle (vec, (pr_ivec4_t) { 2, 2, 1, 0 }); goto negate;
	swizzle_wzyx: vec = swizzle (vec, (pr_ivec4_t) { 3, 2, 1, 0 }); goto negate;
	swizzle_xwyx: vec = swizzle (vec, (pr_ivec4_t) { 0, 3, 1, 0 }); goto negate;
	swizzle_ywyx: vec = swizzle (vec, (pr_ivec4_t) { 1, 3, 1, 0 }); goto negate;
	swizzle_zwyx: vec = swizzle (vec, (pr_ivec4_t) { 2, 3, 1, 0 }); goto negate;
	swizzle_wwyx: vec = swizzle (vec, (pr_ivec4_t) { 3, 3, 1, 0 }); goto negate;
	swizzle_xxzx: vec = swizzle (vec, (pr_ivec4_t) { 0, 0, 2, 0 }); goto negate;
	swizzle_yxzx: vec = swizzle (vec, (pr_ivec4_t) { 1, 0, 2, 0 }); goto negate;
	swizzle_zxzx: vec = swizzle (vec, (pr_ivec4_t) { 2, 0, 2, 0 }); goto negate;
	swizzle_wxzx: vec = swizzle (vec, (pr_ivec4_t) { 3, 0, 2, 0 }); goto negate;
	swizzle_xyzx: vec = swizzle (vec, (pr_ivec4_t) { 0, 1, 2, 0 }); goto negate;
	swizzle_yyzx: vec = swizzle (vec, (pr_ivec4_t) { 1, 1, 2, 0 }); goto negate;
	swizzle_zyzx: vec = swizzle (vec, (pr_ivec4_t) { 2, 1, 2, 0 }); goto negate;
	swizzle_wyzx: vec = swizzle (vec, (pr_ivec4_t) { 3, 1, 2, 0 }); goto negate;
	swizzle_xzzx: vec = swizzle (vec, (pr_ivec4_t) { 0, 2, 2, 0 }); goto negate;
	swizzle_yzzx: vec = swizzle (vec, (pr_ivec4_t) { 1, 2, 2, 0 }); goto negate;
	swizzle_zzzx: vec = swizzle (vec, (pr_ivec4_t) { 2, 2, 2, 0 }); goto negate;
	swizzle_wzzx: vec = swizzle (vec, (pr_ivec4_t) { 3, 2, 2, 0 }); goto negate;
	swizzle_xwzx: vec = swizzle (vec, (pr_ivec4_t) { 0, 3, 2, 0 }); goto negate;
	swizzle_ywzx: vec = swizzle (vec, (pr_ivec4_t) { 1, 3, 2, 0 }); goto negate;
	swizzle_zwzx: vec = swizzle (vec, (pr_ivec4_t) { 2, 3, 2, 0 }); goto negate;
	swizzle_wwzx: vec = swizzle (vec, (pr_ivec4_t) { 3, 3, 2, 0 }); goto negate;
	swizzle_xxwx: vec = swizzle (vec, (pr_ivec4_t) { 0, 0, 3, 0 }); goto negate;
	swizzle_yxwx: vec = swizzle (vec, (pr_ivec4_t) { 1, 0, 3, 0 }); goto negate;
	swizzle_zxwx: vec = swizzle (vec, (pr_ivec4_t) { 2, 0, 3, 0 }); goto negate;
	swizzle_wxwx: vec = swizzle (vec, (pr_ivec4_t) { 3, 0, 3, 0 }); goto negate;
	swizzle_xywx: vec = swizzle (vec, (pr_ivec4_t) { 0, 1, 3, 0 }); goto negate;
	swizzle_yywx: vec = swizzle (vec, (pr_ivec4_t) { 1, 1, 3, 0 }); goto negate;
	swizzle_zywx: vec = swizzle (vec, (pr_ivec4_t) { 2, 1, 3, 0 }); goto negate;
	swizzle_wywx: vec = swizzle (vec, (pr_ivec4_t) { 3, 1, 3, 0 }); goto negate;
	swizzle_xzwx: vec = swizzle (vec, (pr_ivec4_t) { 0, 2, 3, 0 }); goto negate;
	swizzle_yzwx: vec = swizzle (vec, (pr_ivec4_t) { 1, 2, 3, 0 }); goto negate;
	swizzle_zzwx: vec = swizzle (vec, (pr_ivec4_t) { 2, 2, 3, 0 }); goto negate;
	swizzle_wzwx: vec = swizzle (vec, (pr_ivec4_t) { 3, 2, 3, 0 }); goto negate;
	swizzle_xwwx: vec = swizzle (vec, (pr_ivec4_t) { 0, 3, 3, 0 }); goto negate;
	swizzle_ywwx: vec = swizzle (vec, (pr_ivec4_t) { 1, 3, 3, 0 }); goto negate;
	swizzle_zwwx: vec = swizzle (vec, (pr_ivec4_t) { 2, 3, 3, 0 }); goto negate;
	swizzle_wwwx: vec = swizzle (vec, (pr_ivec4_t) { 3, 3, 3, 0 }); goto negate;
	swizzle_xxxy: vec = swizzle (vec, (pr_ivec4_t) { 0, 0, 0, 1 }); goto negate;
	swizzle_yxxy: vec = swizzle (vec, (pr_ivec4_t) { 1, 0, 0, 1 }); goto negate;
	swizzle_zxxy: vec = swizzle (vec, (pr_ivec4_t) { 2, 0, 0, 1 }); goto negate;
	swizzle_wxxy: vec = swizzle (vec, (pr_ivec4_t) { 3, 0, 0, 1 }); goto negate;
	swizzle_xyxy: vec = swizzle (vec, (pr_ivec4_t) { 0, 1, 0, 1 }); goto negate;
	swizzle_yyxy: vec = swizzle (vec, (pr_ivec4_t) { 1, 1, 0, 1 }); goto negate;
	swizzle_zyxy: vec = swizzle (vec, (pr_ivec4_t) { 2, 1, 0, 1 }); goto negate;
	swizzle_wyxy: vec = swizzle (vec, (pr_ivec4_t) { 3, 1, 0, 1 }); goto negate;
	swizzle_xzxy: vec = swizzle (vec, (pr_ivec4_t) { 0, 2, 0, 1 }); goto negate;
	swizzle_yzxy: vec = swizzle (vec, (pr_ivec4_t) { 1, 2, 0, 1 }); goto negate;
	swizzle_zzxy: vec = swizzle (vec, (pr_ivec4_t) { 2, 2, 0, 1 }); goto negate;
	swizzle_wzxy: vec = swizzle (vec, (pr_ivec4_t) { 3, 2, 0, 1 }); goto negate;
	swizzle_xwxy: vec = swizzle (vec, (pr_ivec4_t) { 0, 3, 0, 1 }); goto negate;
	swizzle_ywxy: vec = swizzle (vec, (pr_ivec4_t) { 1, 3, 0, 1 }); goto negate;
	swizzle_zwxy: vec = swizzle (vec, (pr_ivec4_t) { 2, 3, 0, 1 }); goto negate;
	swizzle_wwxy: vec = swizzle (vec, (pr_ivec4_t) { 3, 3, 0, 1 }); goto negate;
	swizzle_xxyy: vec = swizzle (vec, (pr_ivec4_t) { 0, 0, 1, 1 }); goto negate;
	swizzle_yxyy: vec = swizzle (vec, (pr_ivec4_t) { 1, 0, 1, 1 }); goto negate;
	swizzle_zxyy: vec = swizzle (vec, (pr_ivec4_t) { 2, 0, 1, 1 }); goto negate;
	swizzle_wxyy: vec = swizzle (vec, (pr_ivec4_t) { 3, 0, 1, 1 }); goto negate;
	swizzle_xyyy: vec = swizzle (vec, (pr_ivec4_t) { 0, 1, 1, 1 }); goto negate;
	swizzle_yyyy: vec = swizzle (vec, (pr_ivec4_t) { 1, 1, 1, 1 }); goto negate;
	swizzle_zyyy: vec = swizzle (vec, (pr_ivec4_t) { 2, 1, 1, 1 }); goto negate;
	swizzle_wyyy: vec = swizzle (vec, (pr_ivec4_t) { 3, 1, 1, 1 }); goto negate;
	swizzle_xzyy: vec = swizzle (vec, (pr_ivec4_t) { 0, 2, 1, 1 }); goto negate;
	swizzle_yzyy: vec = swizzle (vec, (pr_ivec4_t) { 1, 2, 1, 1 }); goto negate;
	swizzle_zzyy: vec = swizzle (vec, (pr_ivec4_t) { 2, 2, 1, 1 }); goto negate;
	swizzle_wzyy: vec = swizzle (vec, (pr_ivec4_t) { 3, 2, 1, 1 }); goto negate;
	swizzle_xwyy: vec = swizzle (vec, (pr_ivec4_t) { 0, 3, 1, 1 }); goto negate;
	swizzle_ywyy: vec = swizzle (vec, (pr_ivec4_t) { 1, 3, 1, 1 }); goto negate;
	swizzle_zwyy: vec = swizzle (vec, (pr_ivec4_t) { 2, 3, 1, 1 }); goto negate;
	swizzle_wwyy: vec = swizzle (vec, (pr_ivec4_t) { 3, 3, 1, 1 }); goto negate;
	swizzle_xxzy: vec = swizzle (vec, (pr_ivec4_t) { 0, 0, 2, 1 }); goto negate;
	swizzle_yxzy: vec = swizzle (vec, (pr_ivec4_t) { 1, 0, 2, 1 }); goto negate;
	swizzle_zxzy: vec = swizzle (vec, (pr_ivec4_t) { 2, 0, 2, 1 }); goto negate;
	swizzle_wxzy: vec = swizzle (vec, (pr_ivec4_t) { 3, 0, 2, 1 }); goto negate;
	swizzle_xyzy: vec = swizzle (vec, (pr_ivec4_t) { 0, 1, 2, 1 }); goto negate;
	swizzle_yyzy: vec = swizzle (vec, (pr_ivec4_t) { 1, 1, 2, 1 }); goto negate;
	swizzle_zyzy: vec = swizzle (vec, (pr_ivec4_t) { 2, 1, 2, 1 }); goto negate;
	swizzle_wyzy: vec = swizzle (vec, (pr_ivec4_t) { 3, 1, 2, 1 }); goto negate;
	swizzle_xzzy: vec = swizzle (vec, (pr_ivec4_t) { 0, 2, 2, 1 }); goto negate;
	swizzle_yzzy: vec = swizzle (vec, (pr_ivec4_t) { 1, 2, 2, 1 }); goto negate;
	swizzle_zzzy: vec = swizzle (vec, (pr_ivec4_t) { 2, 2, 2, 1 }); goto negate;
	swizzle_wzzy: vec = swizzle (vec, (pr_ivec4_t) { 3, 2, 2, 1 }); goto negate;
	swizzle_xwzy: vec = swizzle (vec, (pr_ivec4_t) { 0, 3, 2, 1 }); goto negate;
	swizzle_ywzy: vec = swizzle (vec, (pr_ivec4_t) { 1, 3, 2, 1 }); goto negate;
	swizzle_zwzy: vec = swizzle (vec, (pr_ivec4_t) { 2, 3, 2, 1 }); goto negate;
	swizzle_wwzy: vec = swizzle (vec, (pr_ivec4_t) { 3, 3, 2, 1 }); goto negate;
	swizzle_xxwy: vec = swizzle (vec, (pr_ivec4_t) { 0, 0, 3, 1 }); goto negate;
	swizzle_yxwy: vec = swizzle (vec, (pr_ivec4_t) { 1, 0, 3, 1 }); goto negate;
	swizzle_zxwy: vec = swizzle (vec, (pr_ivec4_t) { 2, 0, 3, 1 }); goto negate;
	swizzle_wxwy: vec = swizzle (vec, (pr_ivec4_t) { 3, 0, 3, 1 }); goto negate;
	swizzle_xywy: vec = swizzle (vec, (pr_ivec4_t) { 0, 1, 3, 1 }); goto negate;
	swizzle_yywy: vec = swizzle (vec, (pr_ivec4_t) { 1, 1, 3, 1 }); goto negate;
	swizzle_zywy: vec = swizzle (vec, (pr_ivec4_t) { 2, 1, 3, 1 }); goto negate;
	swizzle_wywy: vec = swizzle (vec, (pr_ivec4_t) { 3, 1, 3, 1 }); goto negate;
	swizzle_xzwy: vec = swizzle (vec, (pr_ivec4_t) { 0, 2, 3, 1 }); goto negate;
	swizzle_yzwy: vec = swizzle (vec, (pr_ivec4_t) { 1, 2, 3, 1 }); goto negate;
	swizzle_zzwy: vec = swizzle (vec, (pr_ivec4_t) { 2, 2, 3, 1 }); goto negate;
	swizzle_wzwy: vec = swizzle (vec, (pr_ivec4_t) { 3, 2, 3, 1 }); goto negate;
	swizzle_xwwy: vec = swizzle (vec, (pr_ivec4_t) { 0, 3, 3, 1 }); goto negate;
	swizzle_ywwy: vec = swizzle (vec, (pr_ivec4_t) { 1, 3, 3, 1 }); goto negate;
	swizzle_zwwy: vec = swizzle (vec, (pr_ivec4_t) { 2, 3, 3, 1 }); goto negate;
	swizzle_wwwy: vec = swizzle (vec, (pr_ivec4_t) { 3, 3, 3, 1 }); goto negate;
	swizzle_xxxz: vec = swizzle (vec, (pr_ivec4_t) { 0, 0, 0, 2 }); goto negate;
	swizzle_yxxz: vec = swizzle (vec, (pr_ivec4_t) { 1, 0, 0, 2 }); goto negate;
	swizzle_zxxz: vec = swizzle (vec, (pr_ivec4_t) { 2, 0, 0, 2 }); goto negate;
	swizzle_wxxz: vec = swizzle (vec, (pr_ivec4_t) { 3, 0, 0, 2 }); goto negate;
	swizzle_xyxz: vec = swizzle (vec, (pr_ivec4_t) { 0, 1, 0, 2 }); goto negate;
	swizzle_yyxz: vec = swizzle (vec, (pr_ivec4_t) { 1, 1, 0, 2 }); goto negate;
	swizzle_zyxz: vec = swizzle (vec, (pr_ivec4_t) { 2, 1, 0, 2 }); goto negate;
	swizzle_wyxz: vec = swizzle (vec, (pr_ivec4_t) { 3, 1, 0, 2 }); goto negate;
	swizzle_xzxz: vec = swizzle (vec, (pr_ivec4_t) { 0, 2, 0, 2 }); goto negate;
	swizzle_yzxz: vec = swizzle (vec, (pr_ivec4_t) { 1, 2, 0, 2 }); goto negate;
	swizzle_zzxz: vec = swizzle (vec, (pr_ivec4_t) { 2, 2, 0, 2 }); goto negate;
	swizzle_wzxz: vec = swizzle (vec, (pr_ivec4_t) { 3, 2, 0, 2 }); goto negate;
	swizzle_xwxz: vec = swizzle (vec, (pr_ivec4_t) { 0, 3, 0, 2 }); goto negate;
	swizzle_ywxz: vec = swizzle (vec, (pr_ivec4_t) { 1, 3, 0, 2 }); goto negate;
	swizzle_zwxz: vec = swizzle (vec, (pr_ivec4_t) { 2, 3, 0, 2 }); goto negate;
	swizzle_wwxz: vec = swizzle (vec, (pr_ivec4_t) { 3, 3, 0, 2 }); goto negate;
	swizzle_xxyz: vec = swizzle (vec, (pr_ivec4_t) { 0, 0, 1, 2 }); goto negate;
	swizzle_yxyz: vec = swizzle (vec, (pr_ivec4_t) { 1, 0, 1, 2 }); goto negate;
	swizzle_zxyz: vec = swizzle (vec, (pr_ivec4_t) { 2, 0, 1, 2 }); goto negate;
	swizzle_wxyz: vec = swizzle (vec, (pr_ivec4_t) { 3, 0, 1, 2 }); goto negate;
	swizzle_xyyz: vec = swizzle (vec, (pr_ivec4_t) { 0, 1, 1, 2 }); goto negate;
	swizzle_yyyz: vec = swizzle (vec, (pr_ivec4_t) { 1, 1, 1, 2 }); goto negate;
	swizzle_zyyz: vec = swizzle (vec, (pr_ivec4_t) { 2, 1, 1, 2 }); goto negate;
	swizzle_wyyz: vec = swizzle (vec, (pr_ivec4_t) { 3, 1, 1, 2 }); goto negate;
	swizzle_xzyz: vec = swizzle (vec, (pr_ivec4_t) { 0, 2, 1, 2 }); goto negate;
	swizzle_yzyz: vec = swizzle (vec, (pr_ivec4_t) { 1, 2, 1, 2 }); goto negate;
	swizzle_zzyz: vec = swizzle (vec, (pr_ivec4_t) { 2, 2, 1, 2 }); goto negate;
	swizzle_wzyz: vec = swizzle (vec, (pr_ivec4_t) { 3, 2, 1, 2 }); goto negate;
	swizzle_xwyz: vec = swizzle (vec, (pr_ivec4_t) { 0, 3, 1, 2 }); goto negate;
	swizzle_ywyz: vec = swizzle (vec, (pr_ivec4_t) { 1, 3, 1, 2 }); goto negate;
	swizzle_zwyz: vec = swizzle (vec, (pr_ivec4_t) { 2, 3, 1, 2 }); goto negate;
	swizzle_wwyz: vec = swizzle (vec, (pr_ivec4_t) { 3, 3, 1, 2 }); goto negate;
	swizzle_xxzz: vec = swizzle (vec, (pr_ivec4_t) { 0, 0, 2, 2 }); goto negate;
	swizzle_yxzz: vec = swizzle (vec, (pr_ivec4_t) { 1, 0, 2, 2 }); goto negate;
	swizzle_zxzz: vec = swizzle (vec, (pr_ivec4_t) { 2, 0, 2, 2 }); goto negate;
	swizzle_wxzz: vec = swizzle (vec, (pr_ivec4_t) { 3, 0, 2, 2 }); goto negate;
	swizzle_xyzz: vec = swizzle (vec, (pr_ivec4_t) { 0, 1, 2, 2 }); goto negate;
	swizzle_yyzz: vec = swizzle (vec, (pr_ivec4_t) { 1, 1, 2, 2 }); goto negate;
	swizzle_zyzz: vec = swizzle (vec, (pr_ivec4_t) { 2, 1, 2, 2 }); goto negate;
	swizzle_wyzz: vec = swizzle (vec, (pr_ivec4_t) { 3, 1, 2, 2 }); goto negate;
	swizzle_xzzz: vec = swizzle (vec, (pr_ivec4_t) { 0, 2, 2, 2 }); goto negate;
	swizzle_yzzz: vec = swizzle (vec, (pr_ivec4_t) { 1, 2, 2, 2 }); goto negate;
	swizzle_zzzz: vec = swizzle (vec, (pr_ivec4_t) { 2, 2, 2, 2 }); goto negate;
	swizzle_wzzz: vec = swizzle (vec, (pr_ivec4_t) { 3, 2, 2, 2 }); goto negate;
	swizzle_xwzz: vec = swizzle (vec, (pr_ivec4_t) { 0, 3, 2, 2 }); goto negate;
	swizzle_ywzz: vec = swizzle (vec, (pr_ivec4_t) { 1, 3, 2, 2 }); goto negate;
	swizzle_zwzz: vec = swizzle (vec, (pr_ivec4_t) { 2, 3, 2, 2 }); goto negate;
	swizzle_wwzz: vec = swizzle (vec, (pr_ivec4_t) { 3, 3, 2, 2 }); goto negate;
	swizzle_xxwz: vec = swizzle (vec, (pr_ivec4_t) { 0, 0, 3, 2 }); goto negate;
	swizzle_yxwz: vec = swizzle (vec, (pr_ivec4_t) { 1, 0, 3, 2 }); goto negate;
	swizzle_zxwz: vec = swizzle (vec, (pr_ivec4_t) { 2, 0, 3, 2 }); goto negate;
	swizzle_wxwz: vec = swizzle (vec, (pr_ivec4_t) { 3, 0, 3, 2 }); goto negate;
	swizzle_xywz: vec = swizzle (vec, (pr_ivec4_t) { 0, 1, 3, 2 }); goto negate;
	swizzle_yywz: vec = swizzle (vec, (pr_ivec4_t) { 1, 1, 3, 2 }); goto negate;
	swizzle_zywz: vec = swizzle (vec, (pr_ivec4_t) { 2, 1, 3, 2 }); goto negate;
	swizzle_wywz: vec = swizzle (vec, (pr_ivec4_t) { 3, 1, 3, 2 }); goto negate;
	swizzle_xzwz: vec = swizzle (vec, (pr_ivec4_t) { 0, 2, 3, 2 }); goto negate;
	swizzle_yzwz: vec = swizzle (vec, (pr_ivec4_t) { 1, 2, 3, 2 }); goto negate;
	swizzle_zzwz: vec = swizzle (vec, (pr_ivec4_t) { 2, 2, 3, 2 }); goto negate;
	swizzle_wzwz: vec = swizzle (vec, (pr_ivec4_t) { 3, 2, 3, 2 }); goto negate;
	swizzle_xwwz: vec = swizzle (vec, (pr_ivec4_t) { 0, 3, 3, 2 }); goto negate;
	swizzle_ywwz: vec = swizzle (vec, (pr_ivec4_t) { 1, 3, 3, 2 }); goto negate;
	swizzle_zwwz: vec = swizzle (vec, (pr_ivec4_t) { 2, 3, 3, 2 }); goto negate;
	swizzle_wwwz: vec = swizzle (vec, (pr_ivec4_t) { 3, 3, 3, 2 }); goto negate;
	swizzle_xxxw: vec = swizzle (vec, (pr_ivec4_t) { 0, 0, 0, 3 }); goto negate;
	swizzle_yxxw: vec = swizzle (vec, (pr_ivec4_t) { 1, 0, 0, 3 }); goto negate;
	swizzle_zxxw: vec = swizzle (vec, (pr_ivec4_t) { 2, 0, 0, 3 }); goto negate;
	swizzle_wxxw: vec = swizzle (vec, (pr_ivec4_t) { 3, 0, 0, 3 }); goto negate;
	swizzle_xyxw: vec = swizzle (vec, (pr_ivec4_t) { 0, 1, 0, 3 }); goto negate;
	swizzle_yyxw: vec = swizzle (vec, (pr_ivec4_t) { 1, 1, 0, 3 }); goto negate;
	swizzle_zyxw: vec = swizzle (vec, (pr_ivec4_t) { 2, 1, 0, 3 }); goto negate;
	swizzle_wyxw: vec = swizzle (vec, (pr_ivec4_t) { 3, 1, 0, 3 }); goto negate;
	swizzle_xzxw: vec = swizzle (vec, (pr_ivec4_t) { 0, 2, 0, 3 }); goto negate;
	swizzle_yzxw: vec = swizzle (vec, (pr_ivec4_t) { 1, 2, 0, 3 }); goto negate;
	swizzle_zzxw: vec = swizzle (vec, (pr_ivec4_t) { 2, 2, 0, 3 }); goto negate;
	swizzle_wzxw: vec = swizzle (vec, (pr_ivec4_t) { 3, 2, 0, 3 }); goto negate;
	swizzle_xwxw: vec = swizzle (vec, (pr_ivec4_t) { 0, 3, 0, 3 }); goto negate;
	swizzle_ywxw: vec = swizzle (vec, (pr_ivec4_t) { 1, 3, 0, 3 }); goto negate;
	swizzle_zwxw: vec = swizzle (vec, (pr_ivec4_t) { 2, 3, 0, 3 }); goto negate;
	swizzle_wwxw: vec = swizzle (vec, (pr_ivec4_t) { 3, 3, 0, 3 }); goto negate;
	swizzle_xxyw: vec = swizzle (vec, (pr_ivec4_t) { 0, 0, 1, 3 }); goto negate;
	swizzle_yxyw: vec = swizzle (vec, (pr_ivec4_t) { 1, 0, 1, 3 }); goto negate;
	swizzle_zxyw: vec = swizzle (vec, (pr_ivec4_t) { 2, 0, 1, 3 }); goto negate;
	swizzle_wxyw: vec = swizzle (vec, (pr_ivec4_t) { 3, 0, 1, 3 }); goto negate;
	swizzle_xyyw: vec = swizzle (vec, (pr_ivec4_t) { 0, 1, 1, 3 }); goto negate;
	swizzle_yyyw: vec = swizzle (vec, (pr_ivec4_t) { 1, 1, 1, 3 }); goto negate;
	swizzle_zyyw: vec = swizzle (vec, (pr_ivec4_t) { 2, 1, 1, 3 }); goto negate;
	swizzle_wyyw: vec = swizzle (vec, (pr_ivec4_t) { 3, 1, 1, 3 }); goto negate;
	swizzle_xzyw: vec = swizzle (vec, (pr_ivec4_t) { 0, 2, 1, 3 }); goto negate;
	swizzle_yzyw: vec = swizzle (vec, (pr_ivec4_t) { 1, 2, 1, 3 }); goto negate;
	swizzle_zzyw: vec = swizzle (vec, (pr_ivec4_t) { 2, 2, 1, 3 }); goto negate;
	swizzle_wzyw: vec = swizzle (vec, (pr_ivec4_t) { 3, 2, 1, 3 }); goto negate;
	swizzle_xwyw: vec = swizzle (vec, (pr_ivec4_t) { 0, 3, 1, 3 }); goto negate;
	swizzle_ywyw: vec = swizzle (vec, (pr_ivec4_t) { 1, 3, 1, 3 }); goto negate;
	swizzle_zwyw: vec = swizzle (vec, (pr_ivec4_t) { 2, 3, 1, 3 }); goto negate;
	swizzle_wwyw: vec = swizzle (vec, (pr_ivec4_t) { 3, 3, 1, 3 }); goto negate;
	swizzle_xxzw: vec = swizzle (vec, (pr_ivec4_t) { 0, 0, 2, 3 }); goto negate;
	swizzle_yxzw: vec = swizzle (vec, (pr_ivec4_t) { 1, 0, 2, 3 }); goto negate;
	swizzle_zxzw: vec = swizzle (vec, (pr_ivec4_t) { 2, 0, 2, 3 }); goto negate;
	swizzle_wxzw: vec = swizzle (vec, (pr_ivec4_t) { 3, 0, 2, 3 }); goto negate;
	swizzle_xyzw: vec = swizzle (vec, (pr_ivec4_t) { 0, 1, 2, 3 }); goto negate;
	swizzle_yyzw: vec = swizzle (vec, (pr_ivec4_t) { 1, 1, 2, 3 }); goto negate;
	swizzle_zyzw: vec = swizzle (vec, (pr_ivec4_t) { 2, 1, 2, 3 }); goto negate;
	swizzle_wyzw: vec = swizzle (vec, (pr_ivec4_t) { 3, 1, 2, 3 }); goto negate;
	swizzle_xzzw: vec = swizzle (vec, (pr_ivec4_t) { 0, 2, 2, 3 }); goto negate;
	swizzle_yzzw: vec = swizzle (vec, (pr_ivec4_t) { 1, 2, 2, 3 }); goto negate;
	swizzle_zzzw: vec = swizzle (vec, (pr_ivec4_t) { 2, 2, 2, 3 }); goto negate;
	swizzle_wzzw: vec = swizzle (vec, (pr_ivec4_t) { 3, 2, 2, 3 }); goto negate;
	swizzle_xwzw: vec = swizzle (vec, (pr_ivec4_t) { 0, 3, 2, 3 }); goto negate;
	swizzle_ywzw: vec = swizzle (vec, (pr_ivec4_t) { 1, 3, 2, 3 }); goto negate;
	swizzle_zwzw: vec = swizzle (vec, (pr_ivec4_t) { 2, 3, 2, 3 }); goto negate;
	swizzle_wwzw: vec = swizzle (vec, (pr_ivec4_t) { 3, 3, 2, 3 }); goto negate;
	swizzle_xxww: vec = swizzle (vec, (pr_ivec4_t) { 0, 0, 3, 3 }); goto negate;
	swizzle_yxww: vec = swizzle (vec, (pr_ivec4_t) { 1, 0, 3, 3 }); goto negate;
	swizzle_zxww: vec = swizzle (vec, (pr_ivec4_t) { 2, 0, 3, 3 }); goto negate;
	swizzle_wxww: vec = swizzle (vec, (pr_ivec4_t) { 3, 0, 3, 3 }); goto negate;
	swizzle_xyww: vec = swizzle (vec, (pr_ivec4_t) { 0, 1, 3, 3 }); goto negate;
	swizzle_yyww: vec = swizzle (vec, (pr_ivec4_t) { 1, 1, 3, 3 }); goto negate;
	swizzle_zyww: vec = swizzle (vec, (pr_ivec4_t) { 2, 1, 3, 3 }); goto negate;
	swizzle_wyww: vec = swizzle (vec, (pr_ivec4_t) { 3, 1, 3, 3 }); goto negate;
	swizzle_xzww: vec = swizzle (vec, (pr_ivec4_t) { 0, 2, 3, 3 }); goto negate;
	swizzle_yzww: vec = swizzle (vec, (pr_ivec4_t) { 1, 2, 3, 3 }); goto negate;
	swizzle_zzww: vec = swizzle (vec, (pr_ivec4_t) { 2, 2, 3, 3 }); goto negate;
	swizzle_wzww: vec = swizzle (vec, (pr_ivec4_t) { 3, 2, 3, 3 }); goto negate;
	swizzle_xwww: vec = swizzle (vec, (pr_ivec4_t) { 0, 3, 3, 3 }); goto negate;
	swizzle_ywww: vec = swizzle (vec, (pr_ivec4_t) { 1, 3, 3, 3 }); goto negate;
	swizzle_zwww: vec = swizzle (vec, (pr_ivec4_t) { 2, 3, 3, 3 }); goto negate;
	swizzle_wwww: vec = swizzle (vec, (pr_ivec4_t) { 3, 3, 3, 3 }); goto negate;
	static void *swizzle_table[256] = {
		&&swizzle_xxxx, &&swizzle_yxxx, &&swizzle_zxxx, &&swizzle_wxxx,
		&&swizzle_xyxx, &&swizzle_yyxx, &&swizzle_zyxx, &&swizzle_wyxx,
		&&swizzle_xzxx, &&swizzle_yzxx, &&swizzle_zzxx, &&swizzle_wzxx,
		&&swizzle_xwxx, &&swizzle_ywxx, &&swizzle_zwxx, &&swizzle_wwxx,
		&&swizzle_xxyx, &&swizzle_yxyx, &&swizzle_zxyx, &&swizzle_wxyx,
		&&swizzle_xyyx, &&swizzle_yyyx, &&swizzle_zyyx, &&swizzle_wyyx,
		&&swizzle_xzyx, &&swizzle_yzyx, &&swizzle_zzyx, &&swizzle_wzyx,
		&&swizzle_xwyx, &&swizzle_ywyx, &&swizzle_zwyx, &&swizzle_wwyx,
		&&swizzle_xxzx, &&swizzle_yxzx, &&swizzle_zxzx, &&swizzle_wxzx,
		&&swizzle_xyzx, &&swizzle_yyzx, &&swizzle_zyzx, &&swizzle_wyzx,
		&&swizzle_xzzx, &&swizzle_yzzx, &&swizzle_zzzx, &&swizzle_wzzx,
		&&swizzle_xwzx, &&swizzle_ywzx, &&swizzle_zwzx, &&swizzle_wwzx,
		&&swizzle_xxwx, &&swizzle_yxwx, &&swizzle_zxwx, &&swizzle_wxwx,
		&&swizzle_xywx, &&swizzle_yywx, &&swizzle_zywx, &&swizzle_wywx,
		&&swizzle_xzwx, &&swizzle_yzwx, &&swizzle_zzwx, &&swizzle_wzwx,
		&&swizzle_xwwx, &&swizzle_ywwx, &&swizzle_zwwx, &&swizzle_wwwx,
		&&swizzle_xxxy, &&swizzle_yxxy, &&swizzle_zxxy, &&swizzle_wxxy,
		&&swizzle_xyxy, &&swizzle_yyxy, &&swizzle_zyxy, &&swizzle_wyxy,
		&&swizzle_xzxy, &&swizzle_yzxy, &&swizzle_zzxy, &&swizzle_wzxy,
		&&swizzle_xwxy, &&swizzle_ywxy, &&swizzle_zwxy, &&swizzle_wwxy,
		&&swizzle_xxyy, &&swizzle_yxyy, &&swizzle_zxyy, &&swizzle_wxyy,
		&&swizzle_xyyy, &&swizzle_yyyy, &&swizzle_zyyy, &&swizzle_wyyy,
		&&swizzle_xzyy, &&swizzle_yzyy, &&swizzle_zzyy, &&swizzle_wzyy,
		&&swizzle_xwyy, &&swizzle_ywyy, &&swizzle_zwyy, &&swizzle_wwyy,
		&&swizzle_xxzy, &&swizzle_yxzy, &&swizzle_zxzy, &&swizzle_wxzy,
		&&swizzle_xyzy, &&swizzle_yyzy, &&swizzle_zyzy, &&swizzle_wyzy,
		&&swizzle_xzzy, &&swizzle_yzzy, &&swizzle_zzzy, &&swizzle_wzzy,
		&&swizzle_xwzy, &&swizzle_ywzy, &&swizzle_zwzy, &&swizzle_wwzy,
		&&swizzle_xxwy, &&swizzle_yxwy, &&swizzle_zxwy, &&swizzle_wxwy,
		&&swizzle_xywy, &&swizzle_yywy, &&swizzle_zywy, &&swizzle_wywy,
		&&swizzle_xzwy, &&swizzle_yzwy, &&swizzle_zzwy, &&swizzle_wzwy,
		&&swizzle_xwwy, &&swizzle_ywwy, &&swizzle_zwwy, &&swizzle_wwwy,
		&&swizzle_xxxz, &&swizzle_yxxz, &&swizzle_zxxz, &&swizzle_wxxz,
		&&swizzle_xyxz, &&swizzle_yyxz, &&swizzle_zyxz, &&swizzle_wyxz,
		&&swizzle_xzxz, &&swizzle_yzxz, &&swizzle_zzxz, &&swizzle_wzxz,
		&&swizzle_xwxz, &&swizzle_ywxz, &&swizzle_zwxz, &&swizzle_wwxz,
		&&swizzle_xxyz, &&swizzle_yxyz, &&swizzle_zxyz, &&swizzle_wxyz,
		&&swizzle_xyyz, &&swizzle_yyyz, &&swizzle_zyyz, &&swizzle_wyyz,
		&&swizzle_xzyz, &&swizzle_yzyz, &&swizzle_zzyz, &&swizzle_wzyz,
		&&swizzle_xwyz, &&swizzle_ywyz, &&swizzle_zwyz, &&swizzle_wwyz,
		&&swizzle_xxzz, &&swizzle_yxzz, &&swizzle_zxzz, &&swizzle_wxzz,
		&&swizzle_xyzz, &&swizzle_yyzz, &&swizzle_zyzz, &&swizzle_wyzz,
		&&swizzle_xzzz, &&swizzle_yzzz, &&swizzle_zzzz, &&swizzle_wzzz,
		&&swizzle_xwzz, &&swizzle_ywzz, &&swizzle_zwzz, &&swizzle_wwzz,
		&&swizzle_xxwz, &&swizzle_yxwz, &&swizzle_zxwz, &&swizzle_wxwz,
		&&swizzle_xywz, &&swizzle_yywz, &&swizzle_zywz, &&swizzle_wywz,
		&&swizzle_xzwz, &&swizzle_yzwz, &&swizzle_zzwz, &&swizzle_wzwz,
		&&swizzle_xwwz, &&swizzle_ywwz, &&swizzle_zwwz, &&swizzle_wwwz,
		&&swizzle_xxxw, &&swizzle_yxxw, &&swizzle_zxxw, &&swizzle_wxxw,
		&&swizzle_xyxw, &&swizzle_yyxw, &&swizzle_zyxw, &&swizzle_wyxw,
		&&swizzle_xzxw, &&swizzle_yzxw, &&swizzle_zzxw, &&swizzle_wzxw,
		&&swizzle_xwxw, &&swizzle_ywxw, &&swizzle_zwxw, &&swizzle_wwxw,
		&&swizzle_xxyw, &&swizzle_yxyw, &&swizzle_zxyw, &&swizzle_wxyw,
		&&swizzle_xyyw, &&swizzle_yyyw, &&swizzle_zyyw, &&swizzle_wyyw,
		&&swizzle_xzyw, &&swizzle_yzyw, &&swizzle_zzyw, &&swizzle_wzyw,
		&&swizzle_xwyw, &&swizzle_ywyw, &&swizzle_zwyw, &&swizzle_wwyw,
		&&swizzle_xxzw, &&swizzle_yxzw, &&swizzle_zxzw, &&swizzle_wxzw,
		&&swizzle_xyzw, &&swizzle_yyzw, &&swizzle_zyzw, &&swizzle_wyzw,
		&&swizzle_xzzw, &&swizzle_yzzw, &&swizzle_zzzw, &&swizzle_wzzw,
		&&swizzle_xwzw, &&swizzle_ywzw, &&swizzle_zwzw, &&swizzle_wwzw,
		&&swizzle_xxww, &&swizzle_yxww, &&swizzle_zxww, &&swizzle_wxww,
		&&swizzle_xyww, &&swizzle_yyww, &&swizzle_zyww, &&swizzle_wyww,
		&&swizzle_xzww, &&swizzle_yzww, &&swizzle_zzww, &&swizzle_wzww,
		&&swizzle_xwww, &&swizzle_ywww, &&swizzle_zwww, &&swizzle_wwww,
	};
#undef swizzle
	static const pr_ivec4_t neg[16] = {
		{     0,     0,     0,     0 },
		{ 1<<31,     0,     0,     0 },
		{     0, 1<<31,     0,     0 },
		{ 1<<31, 1<<31,     0,     0 },
		{     0,     0, 1<<31,     0 },
		{ 1<<31,     0, 1<<31,     0 },
		{     0, 1<<31, 1<<31,     0 },
		{ 1<<31, 1<<31, 1<<31,     0 },
		{     0,     0,     0, 1<<31 },
		{ 1<<31,     0,     0, 1<<31 },
		{     0, 1<<31,     0, 1<<31 },
		{ 1<<31, 1<<31,     0, 1<<31 },
		{     0,     0, 1<<31, 1<<31 },
		{ 1<<31,     0, 1<<31, 1<<31 },
		{     0, 1<<31, 1<<31, 1<<31 },
		{ 1<<31, 1<<31, 1<<31, 1<<31 },
	};
	static const pr_ivec4_t zero[16] = {
		{ ~0, ~0, ~0, ~0 },
		{  0, ~0, ~0, ~0 },
		{ ~0,  0, ~0, ~0 },
		{  0,  0, ~0, ~0 },
		{ ~0, ~0,  0, ~0 },
		{  0, ~0,  0, ~0 },
		{ ~0,  0,  0, ~0 },
		{  0,  0,  0, ~0 },
		{ ~0, ~0, ~0,  0 },
		{  0, ~0, ~0,  0 },
		{ ~0,  0, ~0,  0 },
		{  0,  0, ~0,  0 },
		{ ~0, ~0,  0,  0 },
		{  0, ~0,  0,  0 },
		{ ~0,  0,  0,  0 },
		{  0,  0,  0,  0 },
	};

do_swizzle:
	goto *swizzle_table[swiz & 0xff];
negate:
	vec ^= neg[(swiz >> 8) & 0xf];
	vec &= zero[(swiz >> 12) & 0xf];
	return vec;
}

static pr_lvec4_t
pr_swizzle_d (pr_lvec4_t vec, pr_ushort_t swiz)
{
	goto do_swizzle;
#define swizzle __builtin_shuffle
	swizzle_xxxx: vec = swizzle (vec, (pr_lvec4_t) { 0, 0, 0, 0 }); goto negate;
	swizzle_yxxx: vec = swizzle (vec, (pr_lvec4_t) { 1, 0, 0, 0 }); goto negate;
	swizzle_zxxx: vec = swizzle (vec, (pr_lvec4_t) { 2, 0, 0, 0 }); goto negate;
	swizzle_wxxx: vec = swizzle (vec, (pr_lvec4_t) { 3, 0, 0, 0 }); goto negate;
	swizzle_xyxx: vec = swizzle (vec, (pr_lvec4_t) { 0, 1, 0, 0 }); goto negate;
	swizzle_yyxx: vec = swizzle (vec, (pr_lvec4_t) { 1, 1, 0, 0 }); goto negate;
	swizzle_zyxx: vec = swizzle (vec, (pr_lvec4_t) { 2, 1, 0, 0 }); goto negate;
	swizzle_wyxx: vec = swizzle (vec, (pr_lvec4_t) { 3, 1, 0, 0 }); goto negate;
	swizzle_xzxx: vec = swizzle (vec, (pr_lvec4_t) { 0, 2, 0, 0 }); goto negate;
	swizzle_yzxx: vec = swizzle (vec, (pr_lvec4_t) { 1, 2, 0, 0 }); goto negate;
	swizzle_zzxx: vec = swizzle (vec, (pr_lvec4_t) { 2, 2, 0, 0 }); goto negate;
	swizzle_wzxx: vec = swizzle (vec, (pr_lvec4_t) { 3, 2, 0, 0 }); goto negate;
	swizzle_xwxx: vec = swizzle (vec, (pr_lvec4_t) { 0, 3, 0, 0 }); goto negate;
	swizzle_ywxx: vec = swizzle (vec, (pr_lvec4_t) { 1, 3, 0, 0 }); goto negate;
	swizzle_zwxx: vec = swizzle (vec, (pr_lvec4_t) { 2, 3, 0, 0 }); goto negate;
	swizzle_wwxx: vec = swizzle (vec, (pr_lvec4_t) { 3, 3, 0, 0 }); goto negate;
	swizzle_xxyx: vec = swizzle (vec, (pr_lvec4_t) { 0, 0, 1, 0 }); goto negate;
	swizzle_yxyx: vec = swizzle (vec, (pr_lvec4_t) { 1, 0, 1, 0 }); goto negate;
	swizzle_zxyx: vec = swizzle (vec, (pr_lvec4_t) { 2, 0, 1, 0 }); goto negate;
	swizzle_wxyx: vec = swizzle (vec, (pr_lvec4_t) { 3, 0, 1, 0 }); goto negate;
	swizzle_xyyx: vec = swizzle (vec, (pr_lvec4_t) { 0, 1, 1, 0 }); goto negate;
	swizzle_yyyx: vec = swizzle (vec, (pr_lvec4_t) { 1, 1, 1, 0 }); goto negate;
	swizzle_zyyx: vec = swizzle (vec, (pr_lvec4_t) { 2, 1, 1, 0 }); goto negate;
	swizzle_wyyx: vec = swizzle (vec, (pr_lvec4_t) { 3, 1, 1, 0 }); goto negate;
	swizzle_xzyx: vec = swizzle (vec, (pr_lvec4_t) { 0, 2, 1, 0 }); goto negate;
	swizzle_yzyx: vec = swizzle (vec, (pr_lvec4_t) { 1, 2, 1, 0 }); goto negate;
	swizzle_zzyx: vec = swizzle (vec, (pr_lvec4_t) { 2, 2, 1, 0 }); goto negate;
	swizzle_wzyx: vec = swizzle (vec, (pr_lvec4_t) { 3, 2, 1, 0 }); goto negate;
	swizzle_xwyx: vec = swizzle (vec, (pr_lvec4_t) { 0, 3, 1, 0 }); goto negate;
	swizzle_ywyx: vec = swizzle (vec, (pr_lvec4_t) { 1, 3, 1, 0 }); goto negate;
	swizzle_zwyx: vec = swizzle (vec, (pr_lvec4_t) { 2, 3, 1, 0 }); goto negate;
	swizzle_wwyx: vec = swizzle (vec, (pr_lvec4_t) { 3, 3, 1, 0 }); goto negate;
	swizzle_xxzx: vec = swizzle (vec, (pr_lvec4_t) { 0, 0, 2, 0 }); goto negate;
	swizzle_yxzx: vec = swizzle (vec, (pr_lvec4_t) { 1, 0, 2, 0 }); goto negate;
	swizzle_zxzx: vec = swizzle (vec, (pr_lvec4_t) { 2, 0, 2, 0 }); goto negate;
	swizzle_wxzx: vec = swizzle (vec, (pr_lvec4_t) { 3, 0, 2, 0 }); goto negate;
	swizzle_xyzx: vec = swizzle (vec, (pr_lvec4_t) { 0, 1, 2, 0 }); goto negate;
	swizzle_yyzx: vec = swizzle (vec, (pr_lvec4_t) { 1, 1, 2, 0 }); goto negate;
	swizzle_zyzx: vec = swizzle (vec, (pr_lvec4_t) { 2, 1, 2, 0 }); goto negate;
	swizzle_wyzx: vec = swizzle (vec, (pr_lvec4_t) { 3, 1, 2, 0 }); goto negate;
	swizzle_xzzx: vec = swizzle (vec, (pr_lvec4_t) { 0, 2, 2, 0 }); goto negate;
	swizzle_yzzx: vec = swizzle (vec, (pr_lvec4_t) { 1, 2, 2, 0 }); goto negate;
	swizzle_zzzx: vec = swizzle (vec, (pr_lvec4_t) { 2, 2, 2, 0 }); goto negate;
	swizzle_wzzx: vec = swizzle (vec, (pr_lvec4_t) { 3, 2, 2, 0 }); goto negate;
	swizzle_xwzx: vec = swizzle (vec, (pr_lvec4_t) { 0, 3, 2, 0 }); goto negate;
	swizzle_ywzx: vec = swizzle (vec, (pr_lvec4_t) { 1, 3, 2, 0 }); goto negate;
	swizzle_zwzx: vec = swizzle (vec, (pr_lvec4_t) { 2, 3, 2, 0 }); goto negate;
	swizzle_wwzx: vec = swizzle (vec, (pr_lvec4_t) { 3, 3, 2, 0 }); goto negate;
	swizzle_xxwx: vec = swizzle (vec, (pr_lvec4_t) { 0, 0, 3, 0 }); goto negate;
	swizzle_yxwx: vec = swizzle (vec, (pr_lvec4_t) { 1, 0, 3, 0 }); goto negate;
	swizzle_zxwx: vec = swizzle (vec, (pr_lvec4_t) { 2, 0, 3, 0 }); goto negate;
	swizzle_wxwx: vec = swizzle (vec, (pr_lvec4_t) { 3, 0, 3, 0 }); goto negate;
	swizzle_xywx: vec = swizzle (vec, (pr_lvec4_t) { 0, 1, 3, 0 }); goto negate;
	swizzle_yywx: vec = swizzle (vec, (pr_lvec4_t) { 1, 1, 3, 0 }); goto negate;
	swizzle_zywx: vec = swizzle (vec, (pr_lvec4_t) { 2, 1, 3, 0 }); goto negate;
	swizzle_wywx: vec = swizzle (vec, (pr_lvec4_t) { 3, 1, 3, 0 }); goto negate;
	swizzle_xzwx: vec = swizzle (vec, (pr_lvec4_t) { 0, 2, 3, 0 }); goto negate;
	swizzle_yzwx: vec = swizzle (vec, (pr_lvec4_t) { 1, 2, 3, 0 }); goto negate;
	swizzle_zzwx: vec = swizzle (vec, (pr_lvec4_t) { 2, 2, 3, 0 }); goto negate;
	swizzle_wzwx: vec = swizzle (vec, (pr_lvec4_t) { 3, 2, 3, 0 }); goto negate;
	swizzle_xwwx: vec = swizzle (vec, (pr_lvec4_t) { 0, 3, 3, 0 }); goto negate;
	swizzle_ywwx: vec = swizzle (vec, (pr_lvec4_t) { 1, 3, 3, 0 }); goto negate;
	swizzle_zwwx: vec = swizzle (vec, (pr_lvec4_t) { 2, 3, 3, 0 }); goto negate;
	swizzle_wwwx: vec = swizzle (vec, (pr_lvec4_t) { 3, 3, 3, 0 }); goto negate;
	swizzle_xxxy: vec = swizzle (vec, (pr_lvec4_t) { 0, 0, 0, 1 }); goto negate;
	swizzle_yxxy: vec = swizzle (vec, (pr_lvec4_t) { 1, 0, 0, 1 }); goto negate;
	swizzle_zxxy: vec = swizzle (vec, (pr_lvec4_t) { 2, 0, 0, 1 }); goto negate;
	swizzle_wxxy: vec = swizzle (vec, (pr_lvec4_t) { 3, 0, 0, 1 }); goto negate;
	swizzle_xyxy: vec = swizzle (vec, (pr_lvec4_t) { 0, 1, 0, 1 }); goto negate;
	swizzle_yyxy: vec = swizzle (vec, (pr_lvec4_t) { 1, 1, 0, 1 }); goto negate;
	swizzle_zyxy: vec = swizzle (vec, (pr_lvec4_t) { 2, 1, 0, 1 }); goto negate;
	swizzle_wyxy: vec = swizzle (vec, (pr_lvec4_t) { 3, 1, 0, 1 }); goto negate;
	swizzle_xzxy: vec = swizzle (vec, (pr_lvec4_t) { 0, 2, 0, 1 }); goto negate;
	swizzle_yzxy: vec = swizzle (vec, (pr_lvec4_t) { 1, 2, 0, 1 }); goto negate;
	swizzle_zzxy: vec = swizzle (vec, (pr_lvec4_t) { 2, 2, 0, 1 }); goto negate;
	swizzle_wzxy: vec = swizzle (vec, (pr_lvec4_t) { 3, 2, 0, 1 }); goto negate;
	swizzle_xwxy: vec = swizzle (vec, (pr_lvec4_t) { 0, 3, 0, 1 }); goto negate;
	swizzle_ywxy: vec = swizzle (vec, (pr_lvec4_t) { 1, 3, 0, 1 }); goto negate;
	swizzle_zwxy: vec = swizzle (vec, (pr_lvec4_t) { 2, 3, 0, 1 }); goto negate;
	swizzle_wwxy: vec = swizzle (vec, (pr_lvec4_t) { 3, 3, 0, 1 }); goto negate;
	swizzle_xxyy: vec = swizzle (vec, (pr_lvec4_t) { 0, 0, 1, 1 }); goto negate;
	swizzle_yxyy: vec = swizzle (vec, (pr_lvec4_t) { 1, 0, 1, 1 }); goto negate;
	swizzle_zxyy: vec = swizzle (vec, (pr_lvec4_t) { 2, 0, 1, 1 }); goto negate;
	swizzle_wxyy: vec = swizzle (vec, (pr_lvec4_t) { 3, 0, 1, 1 }); goto negate;
	swizzle_xyyy: vec = swizzle (vec, (pr_lvec4_t) { 0, 1, 1, 1 }); goto negate;
	swizzle_yyyy: vec = swizzle (vec, (pr_lvec4_t) { 1, 1, 1, 1 }); goto negate;
	swizzle_zyyy: vec = swizzle (vec, (pr_lvec4_t) { 2, 1, 1, 1 }); goto negate;
	swizzle_wyyy: vec = swizzle (vec, (pr_lvec4_t) { 3, 1, 1, 1 }); goto negate;
	swizzle_xzyy: vec = swizzle (vec, (pr_lvec4_t) { 0, 2, 1, 1 }); goto negate;
	swizzle_yzyy: vec = swizzle (vec, (pr_lvec4_t) { 1, 2, 1, 1 }); goto negate;
	swizzle_zzyy: vec = swizzle (vec, (pr_lvec4_t) { 2, 2, 1, 1 }); goto negate;
	swizzle_wzyy: vec = swizzle (vec, (pr_lvec4_t) { 3, 2, 1, 1 }); goto negate;
	swizzle_xwyy: vec = swizzle (vec, (pr_lvec4_t) { 0, 3, 1, 1 }); goto negate;
	swizzle_ywyy: vec = swizzle (vec, (pr_lvec4_t) { 1, 3, 1, 1 }); goto negate;
	swizzle_zwyy: vec = swizzle (vec, (pr_lvec4_t) { 2, 3, 1, 1 }); goto negate;
	swizzle_wwyy: vec = swizzle (vec, (pr_lvec4_t) { 3, 3, 1, 1 }); goto negate;
	swizzle_xxzy: vec = swizzle (vec, (pr_lvec4_t) { 0, 0, 2, 1 }); goto negate;
	swizzle_yxzy: vec = swizzle (vec, (pr_lvec4_t) { 1, 0, 2, 1 }); goto negate;
	swizzle_zxzy: vec = swizzle (vec, (pr_lvec4_t) { 2, 0, 2, 1 }); goto negate;
	swizzle_wxzy: vec = swizzle (vec, (pr_lvec4_t) { 3, 0, 2, 1 }); goto negate;
	swizzle_xyzy: vec = swizzle (vec, (pr_lvec4_t) { 0, 1, 2, 1 }); goto negate;
	swizzle_yyzy: vec = swizzle (vec, (pr_lvec4_t) { 1, 1, 2, 1 }); goto negate;
	swizzle_zyzy: vec = swizzle (vec, (pr_lvec4_t) { 2, 1, 2, 1 }); goto negate;
	swizzle_wyzy: vec = swizzle (vec, (pr_lvec4_t) { 3, 1, 2, 1 }); goto negate;
	swizzle_xzzy: vec = swizzle (vec, (pr_lvec4_t) { 0, 2, 2, 1 }); goto negate;
	swizzle_yzzy: vec = swizzle (vec, (pr_lvec4_t) { 1, 2, 2, 1 }); goto negate;
	swizzle_zzzy: vec = swizzle (vec, (pr_lvec4_t) { 2, 2, 2, 1 }); goto negate;
	swizzle_wzzy: vec = swizzle (vec, (pr_lvec4_t) { 3, 2, 2, 1 }); goto negate;
	swizzle_xwzy: vec = swizzle (vec, (pr_lvec4_t) { 0, 3, 2, 1 }); goto negate;
	swizzle_ywzy: vec = swizzle (vec, (pr_lvec4_t) { 1, 3, 2, 1 }); goto negate;
	swizzle_zwzy: vec = swizzle (vec, (pr_lvec4_t) { 2, 3, 2, 1 }); goto negate;
	swizzle_wwzy: vec = swizzle (vec, (pr_lvec4_t) { 3, 3, 2, 1 }); goto negate;
	swizzle_xxwy: vec = swizzle (vec, (pr_lvec4_t) { 0, 0, 3, 1 }); goto negate;
	swizzle_yxwy: vec = swizzle (vec, (pr_lvec4_t) { 1, 0, 3, 1 }); goto negate;
	swizzle_zxwy: vec = swizzle (vec, (pr_lvec4_t) { 2, 0, 3, 1 }); goto negate;
	swizzle_wxwy: vec = swizzle (vec, (pr_lvec4_t) { 3, 0, 3, 1 }); goto negate;
	swizzle_xywy: vec = swizzle (vec, (pr_lvec4_t) { 0, 1, 3, 1 }); goto negate;
	swizzle_yywy: vec = swizzle (vec, (pr_lvec4_t) { 1, 1, 3, 1 }); goto negate;
	swizzle_zywy: vec = swizzle (vec, (pr_lvec4_t) { 2, 1, 3, 1 }); goto negate;
	swizzle_wywy: vec = swizzle (vec, (pr_lvec4_t) { 3, 1, 3, 1 }); goto negate;
	swizzle_xzwy: vec = swizzle (vec, (pr_lvec4_t) { 0, 2, 3, 1 }); goto negate;
	swizzle_yzwy: vec = swizzle (vec, (pr_lvec4_t) { 1, 2, 3, 1 }); goto negate;
	swizzle_zzwy: vec = swizzle (vec, (pr_lvec4_t) { 2, 2, 3, 1 }); goto negate;
	swizzle_wzwy: vec = swizzle (vec, (pr_lvec4_t) { 3, 2, 3, 1 }); goto negate;
	swizzle_xwwy: vec = swizzle (vec, (pr_lvec4_t) { 0, 3, 3, 1 }); goto negate;
	swizzle_ywwy: vec = swizzle (vec, (pr_lvec4_t) { 1, 3, 3, 1 }); goto negate;
	swizzle_zwwy: vec = swizzle (vec, (pr_lvec4_t) { 2, 3, 3, 1 }); goto negate;
	swizzle_wwwy: vec = swizzle (vec, (pr_lvec4_t) { 3, 3, 3, 1 }); goto negate;
	swizzle_xxxz: vec = swizzle (vec, (pr_lvec4_t) { 0, 0, 0, 2 }); goto negate;
	swizzle_yxxz: vec = swizzle (vec, (pr_lvec4_t) { 1, 0, 0, 2 }); goto negate;
	swizzle_zxxz: vec = swizzle (vec, (pr_lvec4_t) { 2, 0, 0, 2 }); goto negate;
	swizzle_wxxz: vec = swizzle (vec, (pr_lvec4_t) { 3, 0, 0, 2 }); goto negate;
	swizzle_xyxz: vec = swizzle (vec, (pr_lvec4_t) { 0, 1, 0, 2 }); goto negate;
	swizzle_yyxz: vec = swizzle (vec, (pr_lvec4_t) { 1, 1, 0, 2 }); goto negate;
	swizzle_zyxz: vec = swizzle (vec, (pr_lvec4_t) { 2, 1, 0, 2 }); goto negate;
	swizzle_wyxz: vec = swizzle (vec, (pr_lvec4_t) { 3, 1, 0, 2 }); goto negate;
	swizzle_xzxz: vec = swizzle (vec, (pr_lvec4_t) { 0, 2, 0, 2 }); goto negate;
	swizzle_yzxz: vec = swizzle (vec, (pr_lvec4_t) { 1, 2, 0, 2 }); goto negate;
	swizzle_zzxz: vec = swizzle (vec, (pr_lvec4_t) { 2, 2, 0, 2 }); goto negate;
	swizzle_wzxz: vec = swizzle (vec, (pr_lvec4_t) { 3, 2, 0, 2 }); goto negate;
	swizzle_xwxz: vec = swizzle (vec, (pr_lvec4_t) { 0, 3, 0, 2 }); goto negate;
	swizzle_ywxz: vec = swizzle (vec, (pr_lvec4_t) { 1, 3, 0, 2 }); goto negate;
	swizzle_zwxz: vec = swizzle (vec, (pr_lvec4_t) { 2, 3, 0, 2 }); goto negate;
	swizzle_wwxz: vec = swizzle (vec, (pr_lvec4_t) { 3, 3, 0, 2 }); goto negate;
	swizzle_xxyz: vec = swizzle (vec, (pr_lvec4_t) { 0, 0, 1, 2 }); goto negate;
	swizzle_yxyz: vec = swizzle (vec, (pr_lvec4_t) { 1, 0, 1, 2 }); goto negate;
	swizzle_zxyz: vec = swizzle (vec, (pr_lvec4_t) { 2, 0, 1, 2 }); goto negate;
	swizzle_wxyz: vec = swizzle (vec, (pr_lvec4_t) { 3, 0, 1, 2 }); goto negate;
	swizzle_xyyz: vec = swizzle (vec, (pr_lvec4_t) { 0, 1, 1, 2 }); goto negate;
	swizzle_yyyz: vec = swizzle (vec, (pr_lvec4_t) { 1, 1, 1, 2 }); goto negate;
	swizzle_zyyz: vec = swizzle (vec, (pr_lvec4_t) { 2, 1, 1, 2 }); goto negate;
	swizzle_wyyz: vec = swizzle (vec, (pr_lvec4_t) { 3, 1, 1, 2 }); goto negate;
	swizzle_xzyz: vec = swizzle (vec, (pr_lvec4_t) { 0, 2, 1, 2 }); goto negate;
	swizzle_yzyz: vec = swizzle (vec, (pr_lvec4_t) { 1, 2, 1, 2 }); goto negate;
	swizzle_zzyz: vec = swizzle (vec, (pr_lvec4_t) { 2, 2, 1, 2 }); goto negate;
	swizzle_wzyz: vec = swizzle (vec, (pr_lvec4_t) { 3, 2, 1, 2 }); goto negate;
	swizzle_xwyz: vec = swizzle (vec, (pr_lvec4_t) { 0, 3, 1, 2 }); goto negate;
	swizzle_ywyz: vec = swizzle (vec, (pr_lvec4_t) { 1, 3, 1, 2 }); goto negate;
	swizzle_zwyz: vec = swizzle (vec, (pr_lvec4_t) { 2, 3, 1, 2 }); goto negate;
	swizzle_wwyz: vec = swizzle (vec, (pr_lvec4_t) { 3, 3, 1, 2 }); goto negate;
	swizzle_xxzz: vec = swizzle (vec, (pr_lvec4_t) { 0, 0, 2, 2 }); goto negate;
	swizzle_yxzz: vec = swizzle (vec, (pr_lvec4_t) { 1, 0, 2, 2 }); goto negate;
	swizzle_zxzz: vec = swizzle (vec, (pr_lvec4_t) { 2, 0, 2, 2 }); goto negate;
	swizzle_wxzz: vec = swizzle (vec, (pr_lvec4_t) { 3, 0, 2, 2 }); goto negate;
	swizzle_xyzz: vec = swizzle (vec, (pr_lvec4_t) { 0, 1, 2, 2 }); goto negate;
	swizzle_yyzz: vec = swizzle (vec, (pr_lvec4_t) { 1, 1, 2, 2 }); goto negate;
	swizzle_zyzz: vec = swizzle (vec, (pr_lvec4_t) { 2, 1, 2, 2 }); goto negate;
	swizzle_wyzz: vec = swizzle (vec, (pr_lvec4_t) { 3, 1, 2, 2 }); goto negate;
	swizzle_xzzz: vec = swizzle (vec, (pr_lvec4_t) { 0, 2, 2, 2 }); goto negate;
	swizzle_yzzz: vec = swizzle (vec, (pr_lvec4_t) { 1, 2, 2, 2 }); goto negate;
	swizzle_zzzz: vec = swizzle (vec, (pr_lvec4_t) { 2, 2, 2, 2 }); goto negate;
	swizzle_wzzz: vec = swizzle (vec, (pr_lvec4_t) { 3, 2, 2, 2 }); goto negate;
	swizzle_xwzz: vec = swizzle (vec, (pr_lvec4_t) { 0, 3, 2, 2 }); goto negate;
	swizzle_ywzz: vec = swizzle (vec, (pr_lvec4_t) { 1, 3, 2, 2 }); goto negate;
	swizzle_zwzz: vec = swizzle (vec, (pr_lvec4_t) { 2, 3, 2, 2 }); goto negate;
	swizzle_wwzz: vec = swizzle (vec, (pr_lvec4_t) { 3, 3, 2, 2 }); goto negate;
	swizzle_xxwz: vec = swizzle (vec, (pr_lvec4_t) { 0, 0, 3, 2 }); goto negate;
	swizzle_yxwz: vec = swizzle (vec, (pr_lvec4_t) { 1, 0, 3, 2 }); goto negate;
	swizzle_zxwz: vec = swizzle (vec, (pr_lvec4_t) { 2, 0, 3, 2 }); goto negate;
	swizzle_wxwz: vec = swizzle (vec, (pr_lvec4_t) { 3, 0, 3, 2 }); goto negate;
	swizzle_xywz: vec = swizzle (vec, (pr_lvec4_t) { 0, 1, 3, 2 }); goto negate;
	swizzle_yywz: vec = swizzle (vec, (pr_lvec4_t) { 1, 1, 3, 2 }); goto negate;
	swizzle_zywz: vec = swizzle (vec, (pr_lvec4_t) { 2, 1, 3, 2 }); goto negate;
	swizzle_wywz: vec = swizzle (vec, (pr_lvec4_t) { 3, 1, 3, 2 }); goto negate;
	swizzle_xzwz: vec = swizzle (vec, (pr_lvec4_t) { 0, 2, 3, 2 }); goto negate;
	swizzle_yzwz: vec = swizzle (vec, (pr_lvec4_t) { 1, 2, 3, 2 }); goto negate;
	swizzle_zzwz: vec = swizzle (vec, (pr_lvec4_t) { 2, 2, 3, 2 }); goto negate;
	swizzle_wzwz: vec = swizzle (vec, (pr_lvec4_t) { 3, 2, 3, 2 }); goto negate;
	swizzle_xwwz: vec = swizzle (vec, (pr_lvec4_t) { 0, 3, 3, 2 }); goto negate;
	swizzle_ywwz: vec = swizzle (vec, (pr_lvec4_t) { 1, 3, 3, 2 }); goto negate;
	swizzle_zwwz: vec = swizzle (vec, (pr_lvec4_t) { 2, 3, 3, 2 }); goto negate;
	swizzle_wwwz: vec = swizzle (vec, (pr_lvec4_t) { 3, 3, 3, 2 }); goto negate;
	swizzle_xxxw: vec = swizzle (vec, (pr_lvec4_t) { 0, 0, 0, 3 }); goto negate;
	swizzle_yxxw: vec = swizzle (vec, (pr_lvec4_t) { 1, 0, 0, 3 }); goto negate;
	swizzle_zxxw: vec = swizzle (vec, (pr_lvec4_t) { 2, 0, 0, 3 }); goto negate;
	swizzle_wxxw: vec = swizzle (vec, (pr_lvec4_t) { 3, 0, 0, 3 }); goto negate;
	swizzle_xyxw: vec = swizzle (vec, (pr_lvec4_t) { 0, 1, 0, 3 }); goto negate;
	swizzle_yyxw: vec = swizzle (vec, (pr_lvec4_t) { 1, 1, 0, 3 }); goto negate;
	swizzle_zyxw: vec = swizzle (vec, (pr_lvec4_t) { 2, 1, 0, 3 }); goto negate;
	swizzle_wyxw: vec = swizzle (vec, (pr_lvec4_t) { 3, 1, 0, 3 }); goto negate;
	swizzle_xzxw: vec = swizzle (vec, (pr_lvec4_t) { 0, 2, 0, 3 }); goto negate;
	swizzle_yzxw: vec = swizzle (vec, (pr_lvec4_t) { 1, 2, 0, 3 }); goto negate;
	swizzle_zzxw: vec = swizzle (vec, (pr_lvec4_t) { 2, 2, 0, 3 }); goto negate;
	swizzle_wzxw: vec = swizzle (vec, (pr_lvec4_t) { 3, 2, 0, 3 }); goto negate;
	swizzle_xwxw: vec = swizzle (vec, (pr_lvec4_t) { 0, 3, 0, 3 }); goto negate;
	swizzle_ywxw: vec = swizzle (vec, (pr_lvec4_t) { 1, 3, 0, 3 }); goto negate;
	swizzle_zwxw: vec = swizzle (vec, (pr_lvec4_t) { 2, 3, 0, 3 }); goto negate;
	swizzle_wwxw: vec = swizzle (vec, (pr_lvec4_t) { 3, 3, 0, 3 }); goto negate;
	swizzle_xxyw: vec = swizzle (vec, (pr_lvec4_t) { 0, 0, 1, 3 }); goto negate;
	swizzle_yxyw: vec = swizzle (vec, (pr_lvec4_t) { 1, 0, 1, 3 }); goto negate;
	swizzle_zxyw: vec = swizzle (vec, (pr_lvec4_t) { 2, 0, 1, 3 }); goto negate;
	swizzle_wxyw: vec = swizzle (vec, (pr_lvec4_t) { 3, 0, 1, 3 }); goto negate;
	swizzle_xyyw: vec = swizzle (vec, (pr_lvec4_t) { 0, 1, 1, 3 }); goto negate;
	swizzle_yyyw: vec = swizzle (vec, (pr_lvec4_t) { 1, 1, 1, 3 }); goto negate;
	swizzle_zyyw: vec = swizzle (vec, (pr_lvec4_t) { 2, 1, 1, 3 }); goto negate;
	swizzle_wyyw: vec = swizzle (vec, (pr_lvec4_t) { 3, 1, 1, 3 }); goto negate;
	swizzle_xzyw: vec = swizzle (vec, (pr_lvec4_t) { 0, 2, 1, 3 }); goto negate;
	swizzle_yzyw: vec = swizzle (vec, (pr_lvec4_t) { 1, 2, 1, 3 }); goto negate;
	swizzle_zzyw: vec = swizzle (vec, (pr_lvec4_t) { 2, 2, 1, 3 }); goto negate;
	swizzle_wzyw: vec = swizzle (vec, (pr_lvec4_t) { 3, 2, 1, 3 }); goto negate;
	swizzle_xwyw: vec = swizzle (vec, (pr_lvec4_t) { 0, 3, 1, 3 }); goto negate;
	swizzle_ywyw: vec = swizzle (vec, (pr_lvec4_t) { 1, 3, 1, 3 }); goto negate;
	swizzle_zwyw: vec = swizzle (vec, (pr_lvec4_t) { 2, 3, 1, 3 }); goto negate;
	swizzle_wwyw: vec = swizzle (vec, (pr_lvec4_t) { 3, 3, 1, 3 }); goto negate;
	swizzle_xxzw: vec = swizzle (vec, (pr_lvec4_t) { 0, 0, 2, 3 }); goto negate;
	swizzle_yxzw: vec = swizzle (vec, (pr_lvec4_t) { 1, 0, 2, 3 }); goto negate;
	swizzle_zxzw: vec = swizzle (vec, (pr_lvec4_t) { 2, 0, 2, 3 }); goto negate;
	swizzle_wxzw: vec = swizzle (vec, (pr_lvec4_t) { 3, 0, 2, 3 }); goto negate;
	swizzle_xyzw: vec = swizzle (vec, (pr_lvec4_t) { 0, 1, 2, 3 }); goto negate;
	swizzle_yyzw: vec = swizzle (vec, (pr_lvec4_t) { 1, 1, 2, 3 }); goto negate;
	swizzle_zyzw: vec = swizzle (vec, (pr_lvec4_t) { 2, 1, 2, 3 }); goto negate;
	swizzle_wyzw: vec = swizzle (vec, (pr_lvec4_t) { 3, 1, 2, 3 }); goto negate;
	swizzle_xzzw: vec = swizzle (vec, (pr_lvec4_t) { 0, 2, 2, 3 }); goto negate;
	swizzle_yzzw: vec = swizzle (vec, (pr_lvec4_t) { 1, 2, 2, 3 }); goto negate;
	swizzle_zzzw: vec = swizzle (vec, (pr_lvec4_t) { 2, 2, 2, 3 }); goto negate;
	swizzle_wzzw: vec = swizzle (vec, (pr_lvec4_t) { 3, 2, 2, 3 }); goto negate;
	swizzle_xwzw: vec = swizzle (vec, (pr_lvec4_t) { 0, 3, 2, 3 }); goto negate;
	swizzle_ywzw: vec = swizzle (vec, (pr_lvec4_t) { 1, 3, 2, 3 }); goto negate;
	swizzle_zwzw: vec = swizzle (vec, (pr_lvec4_t) { 2, 3, 2, 3 }); goto negate;
	swizzle_wwzw: vec = swizzle (vec, (pr_lvec4_t) { 3, 3, 2, 3 }); goto negate;
	swizzle_xxww: vec = swizzle (vec, (pr_lvec4_t) { 0, 0, 3, 3 }); goto negate;
	swizzle_yxww: vec = swizzle (vec, (pr_lvec4_t) { 1, 0, 3, 3 }); goto negate;
	swizzle_zxww: vec = swizzle (vec, (pr_lvec4_t) { 2, 0, 3, 3 }); goto negate;
	swizzle_wxww: vec = swizzle (vec, (pr_lvec4_t) { 3, 0, 3, 3 }); goto negate;
	swizzle_xyww: vec = swizzle (vec, (pr_lvec4_t) { 0, 1, 3, 3 }); goto negate;
	swizzle_yyww: vec = swizzle (vec, (pr_lvec4_t) { 1, 1, 3, 3 }); goto negate;
	swizzle_zyww: vec = swizzle (vec, (pr_lvec4_t) { 2, 1, 3, 3 }); goto negate;
	swizzle_wyww: vec = swizzle (vec, (pr_lvec4_t) { 3, 1, 3, 3 }); goto negate;
	swizzle_xzww: vec = swizzle (vec, (pr_lvec4_t) { 0, 2, 3, 3 }); goto negate;
	swizzle_yzww: vec = swizzle (vec, (pr_lvec4_t) { 1, 2, 3, 3 }); goto negate;
	swizzle_zzww: vec = swizzle (vec, (pr_lvec4_t) { 2, 2, 3, 3 }); goto negate;
	swizzle_wzww: vec = swizzle (vec, (pr_lvec4_t) { 3, 2, 3, 3 }); goto negate;
	swizzle_xwww: vec = swizzle (vec, (pr_lvec4_t) { 0, 3, 3, 3 }); goto negate;
	swizzle_ywww: vec = swizzle (vec, (pr_lvec4_t) { 1, 3, 3, 3 }); goto negate;
	swizzle_zwww: vec = swizzle (vec, (pr_lvec4_t) { 2, 3, 3, 3 }); goto negate;
	swizzle_wwww: vec = swizzle (vec, (pr_lvec4_t) { 3, 3, 3, 3 }); goto negate;
	static void *swizzle_table[256] = {
		&&swizzle_xxxx, &&swizzle_yxxx, &&swizzle_zxxx, &&swizzle_wxxx,
		&&swizzle_xyxx, &&swizzle_yyxx, &&swizzle_zyxx, &&swizzle_wyxx,
		&&swizzle_xzxx, &&swizzle_yzxx, &&swizzle_zzxx, &&swizzle_wzxx,
		&&swizzle_xwxx, &&swizzle_ywxx, &&swizzle_zwxx, &&swizzle_wwxx,
		&&swizzle_xxyx, &&swizzle_yxyx, &&swizzle_zxyx, &&swizzle_wxyx,
		&&swizzle_xyyx, &&swizzle_yyyx, &&swizzle_zyyx, &&swizzle_wyyx,
		&&swizzle_xzyx, &&swizzle_yzyx, &&swizzle_zzyx, &&swizzle_wzyx,
		&&swizzle_xwyx, &&swizzle_ywyx, &&swizzle_zwyx, &&swizzle_wwyx,
		&&swizzle_xxzx, &&swizzle_yxzx, &&swizzle_zxzx, &&swizzle_wxzx,
		&&swizzle_xyzx, &&swizzle_yyzx, &&swizzle_zyzx, &&swizzle_wyzx,
		&&swizzle_xzzx, &&swizzle_yzzx, &&swizzle_zzzx, &&swizzle_wzzx,
		&&swizzle_xwzx, &&swizzle_ywzx, &&swizzle_zwzx, &&swizzle_wwzx,
		&&swizzle_xxwx, &&swizzle_yxwx, &&swizzle_zxwx, &&swizzle_wxwx,
		&&swizzle_xywx, &&swizzle_yywx, &&swizzle_zywx, &&swizzle_wywx,
		&&swizzle_xzwx, &&swizzle_yzwx, &&swizzle_zzwx, &&swizzle_wzwx,
		&&swizzle_xwwx, &&swizzle_ywwx, &&swizzle_zwwx, &&swizzle_wwwx,
		&&swizzle_xxxy, &&swizzle_yxxy, &&swizzle_zxxy, &&swizzle_wxxy,
		&&swizzle_xyxy, &&swizzle_yyxy, &&swizzle_zyxy, &&swizzle_wyxy,
		&&swizzle_xzxy, &&swizzle_yzxy, &&swizzle_zzxy, &&swizzle_wzxy,
		&&swizzle_xwxy, &&swizzle_ywxy, &&swizzle_zwxy, &&swizzle_wwxy,
		&&swizzle_xxyy, &&swizzle_yxyy, &&swizzle_zxyy, &&swizzle_wxyy,
		&&swizzle_xyyy, &&swizzle_yyyy, &&swizzle_zyyy, &&swizzle_wyyy,
		&&swizzle_xzyy, &&swizzle_yzyy, &&swizzle_zzyy, &&swizzle_wzyy,
		&&swizzle_xwyy, &&swizzle_ywyy, &&swizzle_zwyy, &&swizzle_wwyy,
		&&swizzle_xxzy, &&swizzle_yxzy, &&swizzle_zxzy, &&swizzle_wxzy,
		&&swizzle_xyzy, &&swizzle_yyzy, &&swizzle_zyzy, &&swizzle_wyzy,
		&&swizzle_xzzy, &&swizzle_yzzy, &&swizzle_zzzy, &&swizzle_wzzy,
		&&swizzle_xwzy, &&swizzle_ywzy, &&swizzle_zwzy, &&swizzle_wwzy,
		&&swizzle_xxwy, &&swizzle_yxwy, &&swizzle_zxwy, &&swizzle_wxwy,
		&&swizzle_xywy, &&swizzle_yywy, &&swizzle_zywy, &&swizzle_wywy,
		&&swizzle_xzwy, &&swizzle_yzwy, &&swizzle_zzwy, &&swizzle_wzwy,
		&&swizzle_xwwy, &&swizzle_ywwy, &&swizzle_zwwy, &&swizzle_wwwy,
		&&swizzle_xxxz, &&swizzle_yxxz, &&swizzle_zxxz, &&swizzle_wxxz,
		&&swizzle_xyxz, &&swizzle_yyxz, &&swizzle_zyxz, &&swizzle_wyxz,
		&&swizzle_xzxz, &&swizzle_yzxz, &&swizzle_zzxz, &&swizzle_wzxz,
		&&swizzle_xwxz, &&swizzle_ywxz, &&swizzle_zwxz, &&swizzle_wwxz,
		&&swizzle_xxyz, &&swizzle_yxyz, &&swizzle_zxyz, &&swizzle_wxyz,
		&&swizzle_xyyz, &&swizzle_yyyz, &&swizzle_zyyz, &&swizzle_wyyz,
		&&swizzle_xzyz, &&swizzle_yzyz, &&swizzle_zzyz, &&swizzle_wzyz,
		&&swizzle_xwyz, &&swizzle_ywyz, &&swizzle_zwyz, &&swizzle_wwyz,
		&&swizzle_xxzz, &&swizzle_yxzz, &&swizzle_zxzz, &&swizzle_wxzz,
		&&swizzle_xyzz, &&swizzle_yyzz, &&swizzle_zyzz, &&swizzle_wyzz,
		&&swizzle_xzzz, &&swizzle_yzzz, &&swizzle_zzzz, &&swizzle_wzzz,
		&&swizzle_xwzz, &&swizzle_ywzz, &&swizzle_zwzz, &&swizzle_wwzz,
		&&swizzle_xxwz, &&swizzle_yxwz, &&swizzle_zxwz, &&swizzle_wxwz,
		&&swizzle_xywz, &&swizzle_yywz, &&swizzle_zywz, &&swizzle_wywz,
		&&swizzle_xzwz, &&swizzle_yzwz, &&swizzle_zzwz, &&swizzle_wzwz,
		&&swizzle_xwwz, &&swizzle_ywwz, &&swizzle_zwwz, &&swizzle_wwwz,
		&&swizzle_xxxw, &&swizzle_yxxw, &&swizzle_zxxw, &&swizzle_wxxw,
		&&swizzle_xyxw, &&swizzle_yyxw, &&swizzle_zyxw, &&swizzle_wyxw,
		&&swizzle_xzxw, &&swizzle_yzxw, &&swizzle_zzxw, &&swizzle_wzxw,
		&&swizzle_xwxw, &&swizzle_ywxw, &&swizzle_zwxw, &&swizzle_wwxw,
		&&swizzle_xxyw, &&swizzle_yxyw, &&swizzle_zxyw, &&swizzle_wxyw,
		&&swizzle_xyyw, &&swizzle_yyyw, &&swizzle_zyyw, &&swizzle_wyyw,
		&&swizzle_xzyw, &&swizzle_yzyw, &&swizzle_zzyw, &&swizzle_wzyw,
		&&swizzle_xwyw, &&swizzle_ywyw, &&swizzle_zwyw, &&swizzle_wwyw,
		&&swizzle_xxzw, &&swizzle_yxzw, &&swizzle_zxzw, &&swizzle_wxzw,
		&&swizzle_xyzw, &&swizzle_yyzw, &&swizzle_zyzw, &&swizzle_wyzw,
		&&swizzle_xzzw, &&swizzle_yzzw, &&swizzle_zzzw, &&swizzle_wzzw,
		&&swizzle_xwzw, &&swizzle_ywzw, &&swizzle_zwzw, &&swizzle_wwzw,
		&&swizzle_xxww, &&swizzle_yxww, &&swizzle_zxww, &&swizzle_wxww,
		&&swizzle_xyww, &&swizzle_yyww, &&swizzle_zyww, &&swizzle_wyww,
		&&swizzle_xzww, &&swizzle_yzww, &&swizzle_zzww, &&swizzle_wzww,
		&&swizzle_xwww, &&swizzle_ywww, &&swizzle_zwww, &&swizzle_wwww,
	};
#undef swizzle
	static const pr_lvec4_t neg[16] = {
		{     0l,     0l,     0l,     0l },
		{ 1l<<63,     0l,     0l,     0l },
		{     0l, 1l<<63,     0l,     0l },
		{ 1l<<63, 1l<<63,     0l,     0l },
		{     0l,     0l, 1l<<63,     0l },
		{ 1l<<63,     0l, 1l<<63,     0l },
		{     0l, 1l<<63, 1l<<63,     0l },
		{ 1l<<63, 1l<<63, 1l<<63,     0l },
		{     0l,     0l,     0l, 1l<<63 },
		{ 1l<<63,     0l,     0l, 1l<<63 },
		{     0l, 1l<<63,     0l, 1l<<63 },
		{ 1l<<63, 1l<<63,     0l, 1l<<63 },
		{     0l,     0l, 1l<<63, 1l<<63 },
		{ 1l<<63,     0l, 1l<<63, 1l<<63 },
		{     0l, 1l<<63, 1l<<63, 1l<<63 },
		{ 1l<<63, 1l<<63, 1l<<63, 1l<<63 },
	};
	static const pr_lvec4_t zero[16] = {
		{ ~0l, ~0l, ~0l, ~0l },
		{  0l, ~0l, ~0l, ~0l },
		{ ~0l,  0l, ~0l, ~0l },
		{  0l,  0l, ~0l, ~0l },
		{ ~0l, ~0l,  0l, ~0l },
		{  0l, ~0l,  0l, ~0l },
		{ ~0l,  0l,  0l, ~0l },
		{  0l,  0l,  0l, ~0l },
		{ ~0l, ~0l, ~0l,  0l },
		{  0l, ~0l, ~0l,  0l },
		{ ~0l,  0l, ~0l,  0l },
		{  0l,  0l, ~0l,  0l },
		{ ~0l, ~0l,  0l,  0l },
		{  0l, ~0l,  0l,  0l },
		{ ~0l,  0l,  0l,  0l },
		{  0l,  0l,  0l,  0l },
	};

do_swizzle:
	goto *swizzle_table[swiz & 0xff];
negate:
	vec ^= neg[(swiz >> 8) & 0xf];
	vec &= zero[(swiz >> 12) & 0xf];
	return vec;
}

static void
pr_exec_ruamoko (progs_t *pr, int exitdepth)
{
	int         profile, startprofile;
	dstatement_t *st;
	pr_type_t   old_val = {0};

	// make a stack frame
	startprofile = profile = 0;

	st = pr->pr_statements + pr->pr_xstatement;

	if (pr->watch) {
		old_val = *pr->watch;
	}

	while (1) {
		st++;
		++pr->pr_xstatement;
		if (pr->pr_xstatement != st - pr->pr_statements)
			PR_RunError (pr, "internal error");
		if (++profile > 1000000 && !pr->no_exec_limit) {
			PR_RunError (pr, "runaway loop error");
		}

		if (pr->pr_trace) {
			if (pr->debug_handler) {
				pr->debug_handler (prd_trace, 0, pr->debug_data);
			} else {
				PR_PrintStatement (pr, st, 1);
			}
		}

		if (st->op & OP_BREAK) {
			if (pr->debug_handler) {
				pr->debug_handler (prd_breakpoint, 0, pr->debug_data);
			} else {
				PR_RunError (pr, "breakpoint hit");
			}
		}

		pointer_t   st_a = st->a + PR_BASE (pr, st, A);
		pointer_t   st_b = st->b + PR_BASE (pr, st, B);
		pointer_t   st_c = st->c + PR_BASE (pr, st, C);


		pr_type_t  *op_a = pr->pr_globals + st_a;
		pr_type_t  *op_b = pr->pr_globals + st_b;
		pr_type_t  *op_c = pr->pr_globals + st_c;

		pr_type_t  *stk;
		pr_type_t  *mm;
		func_t      function;
		pr_opcode_e st_op = st->op & OP_MASK;
		switch (st_op) {
			// 0 0000
			case OP_LOAD_E_1:
			case OP_LOAD_B_1:
			case OP_LOAD_C_1:
			case OP_LOAD_D_1:
				mm = pr_entity_mode (pr, st, 2);
				OPC(int) = MM(int);
				break;
			case OP_LOAD_E_2:
			case OP_LOAD_B_2:
			case OP_LOAD_C_2:
			case OP_LOAD_D_2:
				mm = pr_entity_mode (pr, st, 2);
				OPC(ivec2) = MM(ivec2);
				break;
			case OP_LOAD_E_3:
			case OP_LOAD_B_3:
			case OP_LOAD_C_3:
			case OP_LOAD_D_3:
				mm = pr_entity_mode (pr, st, 2);
				VectorCopy (&MM(int), &OPC(int));
				break;
			case OP_LOAD_E_4:
			case OP_LOAD_B_4:
			case OP_LOAD_C_4:
			case OP_LOAD_D_4:
				mm = pr_entity_mode (pr, st, 2);
				OPC(ivec4) = MM(ivec4);
				break;
			// 0 0001
			case OP_STORE_A_1:
			case OP_STORE_B_1:
			case OP_STORE_C_1:
			case OP_STORE_D_1:
				mm = pr_address_mode (pr, st, 2);
				MM(int) = OPC(int);
				break;
			case OP_STORE_A_2:
			case OP_STORE_B_2:
			case OP_STORE_C_2:
			case OP_STORE_D_2:
				mm = pr_address_mode (pr, st, 2);
				MM(ivec2) = OPC(ivec2);
				break;
			case OP_STORE_A_3:
			case OP_STORE_B_3:
			case OP_STORE_C_3:
			case OP_STORE_D_3:
				mm = pr_address_mode (pr, st, 2);
				VectorCopy (&OPC(int), &MM(int));
				break;
			case OP_STORE_A_4:
			case OP_STORE_B_4:
			case OP_STORE_C_4:
			case OP_STORE_D_4:
				mm = pr_address_mode (pr, st, 2);
				MM(ivec4) = OPC(ivec4);
				break;
			// 0 0010
			case OP_PUSH_A_1:
			case OP_PUSH_B_1:
			case OP_PUSH_C_1:
			case OP_PUSH_D_1:
				mm = pr_address_mode (pr, st, 2);
				stk = pr_stack_push (pr);
				STK(int) = MM(int);
				break;
			case OP_PUSH_A_2:
			case OP_PUSH_B_2:
			case OP_PUSH_C_2:
			case OP_PUSH_D_2:
				mm = pr_address_mode (pr, st, 2);
				stk = pr_stack_push (pr);
				STK(ivec2) = MM(ivec2);
				break;
			case OP_PUSH_A_3:
			case OP_PUSH_B_3:
			case OP_PUSH_C_3:
			case OP_PUSH_D_3:
				mm = pr_address_mode (pr, st, 2);
				stk = pr_stack_push (pr);
				VectorCopy (&MM(int), &STK(int));
				break;
			case OP_PUSH_A_4:
			case OP_PUSH_B_4:
			case OP_PUSH_C_4:
			case OP_PUSH_D_4:
				mm = pr_address_mode (pr, st, 2);
				stk = pr_stack_push (pr);
				STK(ivec4) = MM(ivec4);
				break;
			// 0 0011
			case OP_POP_A_1:
			case OP_POP_B_1:
			case OP_POP_C_1:
			case OP_POP_D_1:
				mm = pr_address_mode (pr, st, 2);
				stk = pr_stack_pop (pr);
				MM(int) = STK(int);
				break;
			case OP_POP_A_2:
			case OP_POP_B_2:
			case OP_POP_C_2:
			case OP_POP_D_2:
				mm = pr_address_mode (pr, st, 2);
				stk = pr_stack_pop (pr);
				MM(ivec2) = STK(ivec2);
				break;
			case OP_POP_A_3:
			case OP_POP_B_3:
			case OP_POP_C_3:
			case OP_POP_D_3:
				mm = pr_address_mode (pr, st, 2);
				stk = pr_stack_pop (pr);
				VectorCopy (&STK(int), &MM(int));
				break;
			case OP_POP_A_4:
			case OP_POP_B_4:
			case OP_POP_C_4:
			case OP_POP_D_4:
				mm = pr_address_mode (pr, st, 2);
				stk = pr_stack_pop (pr);
				MM(ivec4) = STK(ivec4);
				break;
			// 0 0100
			case OP_IFNOT_A:
			case OP_IFNOT_B:
			case OP_IFNOT_C:
			case OP_IFNOT_D:
				if (!OPC(int)) {
					pr->pr_xstatement = pr_jump_mode (pr, st);
					st = pr->pr_statements + pr->pr_xstatement;
				}
				break;
			case OP_IF_A:
			case OP_IF_B:
			case OP_IF_C:
			case OP_IF_D:
				if (OPC(int)) {
					pr->pr_xstatement = pr_jump_mode (pr, st);
					st = pr->pr_statements + pr->pr_xstatement;
				}
				break;
			case OP_JUMP_A:
			case OP_JUMP_B:
			case OP_JUMP_C:
			case OP_JUMP_D:
				pr->pr_xstatement = pr_jump_mode (pr, st);
				st = pr->pr_statements + pr->pr_xstatement;
				break;
			case OP_CALL_A:
			case OP_CALL_B:
			case OP_CALL_C:
			case OP_CALL_D:
				mm = pr_address_mode (pr, st, 0);
				function = mm->func_var;
				pr->pr_argc = 0;
			op_call:
				pr->pr_xfunction->profile += profile - startprofile;
				startprofile = profile;
				PR_CallFunction (pr, function);
				st = pr->pr_statements + pr->pr_xstatement;
				break;
			// 0 0101
			case OP_CALL_2: case OP_CALL_3: case OP_CALL_4:
			case OP_CALL_5: case OP_CALL_6: case OP_CALL_7: case OP_CALL_8:
				pr->pr_params[1] = op_c;
				goto op_call_n;
			case OP_CALL_1:
				pr->pr_params[1] = pr->pr_real_params[1];
			op_call_n:
				pr->pr_params[0] = op_b;
				function = op_a->func_var;
				pr->pr_argc = st->op - OP_CALL_1 + 1;
				goto op_call;
			case OP_RETURN_4:
				memcpy (&R_INT (pr), op_a, 4 * sizeof (*op_a));
				goto op_return;
			case OP_RETURN_3:
				memcpy (&R_INT (pr), op_a, 3 * sizeof (*op_a));
				goto op_return;
			case OP_RETURN_2:
				memcpy (&R_INT (pr), op_a, 2 * sizeof (*op_a));
				goto op_return;
			case OP_RETURN_1:
				memcpy (&R_INT (pr), op_a, 1 * sizeof (*op_a));
				goto op_return;
			case OP_RETURN_0:
			op_return:
				pr->pr_xfunction->profile += profile - startprofile;
				startprofile = profile;
				PR_LeaveFunction (pr, pr->pr_depth == exitdepth);
				st = pr->pr_statements + pr->pr_xstatement;
				if (pr->pr_depth== exitdepth) {
					if (pr->pr_trace && pr->pr_depth <= pr->pr_trace_depth) {
						pr->pr_trace = false;
					}
					goto exit_program;
				}
				break;
			case OP_WITH:
				pr->pr_bases[st->c & 3] = pr_with (pr, st);
				break;
			case OP_STATE_ft:
				{
					int         self = *pr->globals.self;
					int         nextthink = pr->fields.nextthink + self;
					int         frame = pr->fields.frame + self;
					int         think = pr->fields.think + self;
					float       time = *pr->globals.time + 0.1;
					pr->pr_edict_area[nextthink].float_var = time;
					pr->pr_edict_area[frame].float_var = OPA(float);
					pr->pr_edict_area[think].func_var = op_b->func_var;
				}
				break;
			case OP_STATE_ftt:
				{
					int         self = *pr->globals.self;
					int         nextthink = pr->fields.nextthink + self;
					int         frame = pr->fields.frame + self;
					int         think = pr->fields.think + self;
					float       time = *pr->globals.time + OPC(float);
					pr->pr_edict_area[nextthink].float_var = time;
					pr->pr_edict_area[frame].float_var = OPA(float);
					pr->pr_edict_area[think].func_var = op_b->func_var;
				}
				break;
			// 0 0110
			case OP_IFA_A:
			case OP_IFA_B:
			case OP_IFA_C:
			case OP_IFA_D:
				if (OPC(int) > 0) {
					pr->pr_xstatement = pr_jump_mode (pr, st);
					st = pr->pr_statements + pr->pr_xstatement;
				}
				break;
			case OP_IFBE_A:
			case OP_IFBE_B:
			case OP_IFBE_C:
			case OP_IFBE_D:
				if (OPC(int) <= 0) {
					pr->pr_xstatement = pr_jump_mode (pr, st);
					st = pr->pr_statements + pr->pr_xstatement;
				}
				break;
			case OP_IFB_A:
			case OP_IFB_B:
			case OP_IFB_C:
			case OP_IFB_D:
				if (OPC(int) < 0) {
					pr->pr_xstatement = pr_jump_mode (pr, st);
					st = pr->pr_statements + pr->pr_xstatement;
				}
				break;
			case OP_IFAE_A:
			case OP_IFAE_B:
			case OP_IFAE_C:
			case OP_IFAE_D:
				if (OPC(int) >= 0) {
					pr->pr_xstatement = pr_jump_mode (pr, st);
					st = pr->pr_statements + pr->pr_xstatement;
				}
				break;
			// 0 0111
			case OP_DOT_CC_F:
				OPC(vec2) = dot2f (OPA(vec2), OPB(vec2));
				break;
			case OP_DOT_VV_F:
				{
					vec_t       d = DotProduct (&OPA(float),
												&OPB(float));
					VectorSet (d, d, d, &OPC(float));
				}
				break;
			case OP_DOT_QQ_F:
				OPC(vec4) = dotf (OPA(vec4), OPB(vec4));
				break;
			case OP_CROSS_VV_F:
				{
					pr_vec4_t   a = loadvec3f (&OPA(float));
					pr_vec4_t   b = loadvec3f (&OPB(float));
					pr_vec4_t   c = crossf (a, b);
					storevec3f (&OPC(float), c);
				}
				break;
			case OP_MUL_CC_F:
				OPC(vec2) = cmulf (OPA(vec2), OPB(vec2));
				break;
			case OP_MUL_QV_F:
				{
					pr_vec4_t   v = loadvec3f (&OPB(float));
					v = qvmulf (OPA(vec4), v);
					storevec3f (&OPC(float), v);
				}
				break;
			case OP_MUL_VQ_F:
				{
					pr_vec4_t   v = loadvec3f (&OPA(float));
					v = vqmulf (v, OPB(vec4));
					storevec3f (&OPC(float), v);
				}
				break;
			case OP_MUL_QQ_F:
				OPC(vec4) = qmulf (OPA(vec4), OPB(vec4));
				break;
			case OP_DOT_CC_D:
				OPC(dvec2) = dot2d (OPA(dvec2), OPB(dvec2));
				break;
			case OP_DOT_VV_D:
				{
					double      d = DotProduct (&OPA(double),
												&OPB(double));
					VectorSet (d, d, d, &OPC(double));
				}
				break;
			case OP_DOT_QQ_D:
				OPC(dvec4) = dotd (OPA(dvec4), OPB(dvec4));
				break;
			case OP_CROSS_VV_D:
				{
					pr_dvec4_t  a = loadvec3d (&OPA(double));
					pr_dvec4_t  b = loadvec3d (&OPB(double));
					pr_dvec4_t  c = crossd (a, b);
					storevec3d (&OPC(double), c);
				}
				break;
			case OP_MUL_CC_D:
				OPC(dvec2) = cmuld (OPA(dvec2), OPB(dvec2));
				break;
			case OP_MUL_QV_D:
				{
					pr_dvec4_t   v = loadvec3d (&OPB(double));
					v = qvmuld (OPA(dvec4), v);
					storevec3d (&OPC(double), v);
				}
				break;
			case OP_MUL_VQ_D:
				{
					pr_dvec4_t  v = loadvec3d (&OPA(double));
					v = vqmuld (v, OPB(dvec4));
					storevec3d (&OPC(double), v);
				}
				break;
			case OP_MUL_QQ_D:
				OPC(dvec4) = qmuld (OPA(dvec4), OPB(dvec4));
				break;

#define OP_cmp_1(OP, T, rt, cmp, ct) \
			case OP_##OP##_##T##_1: \
				OPC(rt) = -(OPA(ct) cmp OPB(ct)); \
				break
#define OP_cmp_2(OP, T, rt, cmp, ct) \
			case OP_##OP##_##T##_2: \
				OPC(rt) = (OPA(ct) cmp OPB(ct)); \
				break
#define OP_cmp_3(OP, T, rt, cmp, ct) \
			case OP_##OP##_##T##_3: \
				VectorCompCompare (&OPC(rt), -, &OPA(ct), cmp, &OPB(ct)); \
				break;
#define OP_cmp_4(OP, T, rt, cmp, ct) \
			case OP_##OP##_##T##_4: \
				OPC(rt) = (OPA(ct) cmp OPB(ct)); \
				break
#define OP_cmp_T(OP, T, rt1, rt2, rt4, cmp, ct1, ct2, ct4) \
			OP_cmp_1 (OP, T, rt1, cmp, ct1); \
			OP_cmp_2 (OP, T, rt2, cmp, ct2); \
			OP_cmp_3 (OP, T, rt1, cmp, ct1); \
			OP_cmp_4 (OP, T, rt4, cmp, ct4)
#define OP_cmp(OP, cmp) \
			OP_cmp_T (OP, I, int, ivec2, ivec4, cmp, int, ivec2, ivec4); \
			OP_cmp_T (OP, F, int, ivec2, ivec4, cmp, float, vec2, vec4); \
			OP_cmp_T (OP, L, long, lvec2, lvec4, cmp, long, lvec2, lvec4); \
			OP_cmp_T (OP, D, long, lvec2, lvec4, cmp, double, dvec2, dvec4)

			// 0 1000
			OP_cmp(EQ, ==);
			// 0 1001
			OP_cmp(LT, <);
			// 0 1010
			OP_cmp(GT, >);
			// 0 1011
			//FIXME conversion 1
			// 0 1100
			OP_cmp(NE, !=);
			// 0 1101
			OP_cmp(GE, >=);
			// 0 1110
			OP_cmp(LE, <=);
			// 0 1011
			//FIXME conversion 2

#define OP_op_1(OP, T, t, op) \
			case OP_##OP##_##T##_1: \
				OPC(t) = (OPA(t) op OPB(t)); \
				break
#define OP_op_2(OP, T, t, op) \
			case OP_##OP##_##T##_2: \
				OPC(t) = (OPA(t) op OPB(t)); \
				break
#define OP_op_3(OP, T, t, op) \
			case OP_##OP##_##T##_3: \
				VectorCompOp (&OPC(t), &OPA(t), op, &OPB(t)); \
				break;
#define OP_op_4(OP, T, t, op) \
			case OP_##OP##_##T##_4: \
				OPC(t) = (OPA(t) op OPB(t)); \
				break
#define OP_op_T(OP, T, t1, t2, t4, op) \
			OP_op_1 (OP, T, t1, op); \
			OP_op_2 (OP, T, t2, op); \
			OP_op_3 (OP, T, t1, op); \
			OP_op_4 (OP, T, t4, op)
#define OP_op(OP, op) \
			OP_op_T (OP, I, int, ivec2, ivec4, op); \
			OP_op_T (OP, F, float, vec2, vec4, op); \
			OP_op_T (OP, L, long, lvec2, lvec4, op); \
			OP_op_T (OP, D, double, dvec2, dvec4, op)
#define OP_uop_1(OP, T, t, op) \
			case OP_##OP##_##T##_1: \
				OPC(t) = op (OPA(t)); \
				break
#define OP_uop_2(OP, T, t, op) \
			case OP_##OP##_##T##_2: \
				OPC(t) = op (OPA(t)); \
				break
#define OP_uop_3(OP, T, t, op) \
			case OP_##OP##_##T##_3: \
				VectorCompUop (&OPC(t), op, &OPA(t)); \
				break;
#define OP_uop_4(OP, T, t, op) \
			case OP_##OP##_##T##_4: \
				OPC(t) = op (OPA(t)); \
				break
#define OP_uop_T(OP, T, t1, t2, t4, op) \
			OP_uop_1 (OP, T, t1, op); \
			OP_uop_2 (OP, T, t2, op); \
			OP_uop_3 (OP, T, t1, op); \
			OP_uop_4 (OP, T, t4, op)

			// 1 0000
			OP_op(MUL, *);
			// 1 0001
			OP_op(DIV, /);

// implement remainder (c %) for integers:
//  5 rem  3 =  2
// -5 rem  3 = -2
//  5 rem -3 =  2
// -5 rem -3 = -2
#define OP_store(d, s) *(d) = s
#define OP_remmod_T(OP, T, n, t, l, f, s) \
			case OP_##OP##_##T##_##n: \
				{ \
					__auto_type a = l (&OPA(t)); \
					__auto_type b = l (&OPB(t)); \
					s (&OPC(t), a - b * f(a / b)); \
				} \
				break
#define OP_rem_T(T, n, t, l, f, s) \
			OP_remmod_T(REM, T, n, t, l, f, s)

			// 1 0010
			OP_op_T (REM, I, int, ivec2, ivec4, %);
			OP_rem_T (F, 1, float, *, truncf, OP_store);
			OP_rem_T (F, 2, vec2, *, vtrunc2f, OP_store);
			OP_rem_T (F, 3, float, loadvec3f, vtrunc4f, storevec3f);
			OP_rem_T (F, 4, vec4, *, vtrunc4f, OP_store);
			OP_op_T (REM, L, long, lvec2, lvec4, %);
			OP_rem_T (D, 1, double, *, trunc, OP_store);
			OP_rem_T (D, 2, dvec2, *, vtrunc2d, OP_store);
			OP_rem_T (D, 3, double, loadvec3d, vtrunc4d, storevec3d);
			OP_rem_T (D, 4, dvec4, *, vtrunc4d, OP_store);

// implement true modulo (python %) for integers:
//  5 mod  3 =  2
// -5 mod  3 =  1
//  5 mod -3 = -1
// -5 mod -3 = -2
#define OP_mod_Ti(T, n, t, l, m, s) \
			case OP_MOD_##T##_##n: \
				{ \
					__auto_type a = l(&OPA(t)); \
					__auto_type b = l(&OPB(t)); \
					__auto_type c = a % b; \
					/* % is really remainder and so has the same sign rules */\
					/* as division: -5 % 3 = -2, so need to add b (3 here)  */\
					/* if c's sign is incorrect, but only if c is non-zero  */\
					__auto_type mask = m((a ^ b) < 0); \
					mask &= m(c != 0); \
					s(&OPC(t), c + (mask & b)); \
				} \
				break
// floating point modulo is so much easier :P (just use floor instead of trunc)
#define OP_mod_Tf(T, n, t, l, f, s) \
			OP_remmod_T(MOD, T, n, t, l, f, s)

			// 1 0011
			OP_mod_Ti (I, 1, int, *, -, OP_store);
			OP_mod_Ti (I, 2, ivec2, *, +, OP_store);
			OP_mod_Ti (I, 3, int, loadvec3i1, +, storevec3i);
			OP_mod_Ti (I, 4, ivec4, *, +, OP_store);
			OP_mod_Tf (F, 1, float, *, floorf, OP_store);
			OP_mod_Tf (F, 2, vec2, *, vfloor2f, OP_store);
			OP_mod_Tf (F, 3, float, loadvec3f, vfloor4f, storevec3f);
			OP_mod_Tf (F, 4, vec4, *, vfloor4f, OP_store);
			OP_mod_Ti (L, 1, long, *, -, OP_store);
			OP_mod_Ti (L, 2, lvec2, *, +, OP_store);
			OP_mod_Ti (L, 3, long, loadvec3l1, +, storevec3l);
			OP_mod_Ti (L, 4, lvec4, *, +, OP_store);
			OP_mod_Tf (D, 1, double, *, floor, OP_store);
			OP_mod_Tf (D, 2, dvec2, *, vfloor2d, OP_store);
			OP_mod_Tf (D, 3, double, loadvec3d, vfloor4d, storevec3d);
			OP_mod_Tf (D, 4, dvec4, *, vfloor4d, OP_store);

			// 1 0100
			OP_op(ADD, +);
			// 1 0101
			OP_op(SUB, -);
			// 1 0110
			OP_op_T (SHL, I, int, ivec2, ivec4, <<);
			OP_op_T (SHL, L, long, lvec2, lvec4, <<);
			case OP_EQ_S:
			case OP_LT_S:
			case OP_GT_S:
			case OP_CMP_S:
			case OP_GE_S:
			case OP_LE_S:
				{
					int         cmp = strcmp (PR_GetString (pr, OPA(string)),
											  PR_GetString (pr, OPB(string)));
					switch (st_op) {
						case OP_EQ_S: cmp = (cmp == 0); break;
						case OP_LT_S: cmp = (cmp <  0); break;
						case OP_GT_S: cmp = (cmp >  0); break;
						case OP_GE_S: cmp = (cmp >= 0); break;
						case OP_LE_S: cmp = (cmp <= 0); break;
						case OP_NOT_S: break;
						default: break;
					}
					OPC(int) = cmp;
				}
				break;
			case OP_ADD_S:
				OPC(string) = PR_CatStrings(pr, PR_GetString (pr, OPA(string)),
											PR_GetString (pr, OPB(string)));
				break;
			case OP_NOT_S:
				OPC(int) = !OPA(string) || !*PR_GetString (pr, OPA(string));
				break;
			// 1 0111
			OP_op_T (SHR, I, int, ivec2, ivec4, >>);
			OP_op_T (SHR, u, uint, uivec2, uivec4, >>);
			OP_op_T (SHR, L, long, lvec2, lvec4, >>);
			OP_op_T (SHR, U, ulong, ulvec2, ulvec4, >>);
			// 1 1000
			OP_op_T (BITAND, I, int, ivec2, ivec4, &);
			OP_op_T (BITOR, I, int, ivec2, ivec4, |);
			OP_op_T (BITXOR, I, int, ivec2, ivec4, ^);
			OP_uop_T (BITNOT, I, int, ivec2, ivec4, ~);
			// 1 1001
			OP_op_T (LT, u, uint, uivec2, uivec4, <);
			case OP_SWIZZLE_F:
				OPC(ivec4) = pr_swizzle_f (OPA(ivec4), st->b);
				break;
			//FIXME scale ops
			OP_op_T (LT, U, ulong, ulvec2, ulvec4, <);
			case OP_SWIZZLE_D:
				OPC(lvec4) = pr_swizzle_d (OPA(lvec4), st->b);
				break;
			//FIXME scale ops
			// 1 1010
			OP_op_T (GT, u, uint, uivec2, uivec4, >);
			//FIXME conversion ops
			OP_op_T (GT, U, ulong, ulvec2, ulvec4, >);
			//FIXME conversion ops
			// 1 1011
			case OP_LEA_A:
			case OP_LEA_B:
			case OP_LEA_C:
			case OP_LEA_D:
				mm = pr_address_mode (pr, st, 0);
				op_c->pointer_var = mm - pr->pr_globals;
				break;
			case OP_LEA_E:
				// ensures OP_LEA_E is compatible with OP_LOAD_E_n and thus
				// with pr_entity_mode
				mm = __builtin_choose_expr (
						(OP_LEA_E & 3) == 0,
						pr_entity_mode (pr, st, 0),
						(void) 0);
				op_c->pointer_var = mm - pr->pr_globals;
				break;
			case OP_ANY_2:
				OPC(int) = any2i (OPA(ivec2));
				break;
			case OP_ANY_3:
				{
					__auto_type v = loadvec3i (&OPA(int));
					OPC(int) = any4i (v);
				}
				break;
			case OP_ANY_4:
				OPC(int) = any4i (OPA(ivec4));
				break;
			case OP_PUSHREG:
				stk = pr_stack_push (pr);
				STK(uivec4) = pr->pr_bases;
				break;
			case OP_ALL_2:
				OPC(int) = all2i (OPA(ivec2));
				break;
			case OP_ALL_3:
				{
					__auto_type v = loadvec3i (&OPA(int));
					v[3] = -1;
					OPC(int) = all4i (v);
				}
				break;
			case OP_ALL_4:
				OPC(int) = all4i (OPA(ivec4));
				break;
			case OP_POPREG:
				stk = pr_stack_pop (pr);
				pr->pr_bases = STK(uivec4);
				break;
			case OP_NONE_2:
				OPC(int) = none2i (OPA(ivec2));
				break;
			case OP_NONE_3:
				{
					__auto_type v = loadvec3i (&OPA(int));
					OPC(int) = none4i (v);
				}
				break;
			case OP_NONE_4:
				OPC(int) = none4i (OPA(ivec4));
				break;

#define OP_bool_n(OP, t, n, op, m) \
			case OP_##OP##_I_##n: \
				OPC(t) = m((OPA(t) != 0) op (OPB(t) != 0)); \
				break
#define OP_bool_3(OP, t, n, op, m) \
			case OP_##OP##_I_##n: \
				{ \
					__auto_type a = loadvec3i (&OPA(int)); \
					__auto_type b = loadvec3i (&OPB(int)); \
					storevec3i (&OPC(int), (a != 0) op (b != 0)); \
				} \
				break
#define OP_bool(OP, op) \
			OP_bool_n (OP, int, 1, op, -); \
			OP_bool_n (OP, ivec2, 2, op, +); \
			OP_bool_3 (OP, int, 3, op, +); \
			OP_bool_n (OP, ivec4, 4, op, +)
#define OP_not_n(OP, t, n, m) \
			case OP_##OP##_I_##n: \
				OPC(t) = m((OPA(t) == 0)); \
				break
#define OP_not_3(OP, t, n, m) \
			case OP_##OP##_I_##n: \
				{ \
					__auto_type a = loadvec3i (&OPA(int)); \
					storevec3i (&OPC(int), (a == 0)); \
				} \
				break
			// 1 1100
			OP_bool (AND, &);
			OP_bool (OR, |);
			OP_bool (XOR, ^);
			OP_not_n (NOT, int, 1, -);
			OP_not_n (NOT, ivec2, 2, +);
			OP_not_3 (NOT, int, 3, +);
			OP_not_n (NOT, ivec4, 4, +);
			// 1 1101
			OP_op_T (GE, u, uint, uivec2, uivec4, >=);
			case OP_MUL_QV4_F:
				OPC(vec4) = qvmulf (OPA(vec4), OPB(vec4));
				break;
			case OP_MOVE_I:
				memmove (op_c, op_a, st->b * sizeof (pr_type_t));
				break;
			case OP_MOVE_P:
				memmove (pr->pr_globals + OPC(int), pr->pr_globals + OPA(int),
						 OPB(uint) * sizeof (pr_type_t));
				break;
			case OP_MOVE_PI:
				memmove (pr->pr_globals + OPC(int), pr->pr_globals + OPA(int),
						 st->b * sizeof (pr_type_t));
				break;
			OP_op_T (GE, U, ulong, ulvec2, ulvec4, >=);
			case OP_MUL_QV4_D:
				OPC(dvec4) = qvmuld (OPA(dvec4), OPB(dvec4));
				break;
			case OP_MEMSET_I:
				pr_memset (op_c, OPA(int), st->b);
				break;
			case OP_MEMSET_P:
				pr_memset (pr->pr_globals + OPC(int), OPA(int), OPB(uint));
				break;
			case OP_MEMSET_PI:
				pr_memset (pr->pr_globals + OPC(int), OPA(int), st->b);
				break;
			// 1 1110
			OP_op_T (LE, u, uint, uivec2, uivec4, <=);
			case OP_MUL_V4Q_F:
				OPC(vec4) = vqmulf (OPA(vec4), OPB(vec4));
				break;

			OP_op_T (LE, U, ulong, ulvec2, ulvec4, <=);
			case OP_MUL_V4Q_D:
				OPC(dvec4) = vqmuld (OPA(dvec4), OPB(dvec4));
				break;

			// 1 1111

			default:
				PR_RunError (pr, "Bad opcode o%03o", st->op & OP_MASK);
		}
		if (pr->watch && pr->watch->integer_var != old_val.integer_var) {
			if (!pr->wp_conditional
				|| pr->watch->integer_var == pr->wp_val.integer_var) {
				if (pr->debug_handler) {
					pr->debug_handler (prd_watchpoint, 0, pr->debug_data);
				} else {
					PR_RunError (pr, "watchpoint hit: %d -> %d",
								 old_val.integer_var, pr->watch->integer_var);
				}
			}
			old_val.integer_var = pr->watch->integer_var;
		}
	}
exit_program:
}
/*
	PR_ExecuteProgram

	The interpretation main loop
*/
VISIBLE void
PR_ExecuteProgram (progs_t *pr, func_t fnum)
{
	Sys_PushSignalHook (signal_hook, pr);
	Sys_PushErrorHandler (error_handler, pr);

	if (pr->debug_handler) {
		pr->debug_handler (prd_subenter, &fnum, pr->debug_data);
	}

	int         exitdepth = pr->pr_depth;
	if (!PR_CallFunction (pr, fnum)) {
		// called a builtin instead of progs code
		goto exit_program;
	}
	if (pr->progs->version < PROG_VERSION) {
		pr_exec_quakec (pr, exitdepth);
	} else {
		pr_exec_ruamoko (pr, exitdepth);
	}
exit_program:
	if (pr->debug_handler) {
		pr->debug_handler (prd_subexit, 0, pr->debug_data);
	}
	pr->pr_argc = 0;
	Sys_PopErrorHandler ();
	Sys_PopSignalHook ();
}