diff --git a/Makefile b/Makefile
index f874cf3..8b0d09b 100644
--- a/Makefile
+++ b/Makefile
@@ -6,9 +6,12 @@ OBJCOPY = objcopy
 
 MCUFLAGS = -mthumb -mcpu=cortex-m4 -mfloat-abi=hard -mfpu=fpv4-sp-d16
 AFLAGS = $(MCUFLAGS) 
-CFLAGS = $(MCUFLAGS) -Iinclude -Iinclude/it -fno-builtin -fsigned-char -ffreestanding -Wall -Werror -Wextra -Wno-discarded-qualifiers -ggdb
-LFLAGS = 
-OFLAGS = -O ihex
+CFLAGS = $(MCUFLAGS) -ggdb \
+	-Iinclude -Iinclude/it \
+	-fno-builtin -fsigned-char -ffreestanding \
+	-Wall -Werror -Wextra -pedantic \
+	-Wno-overlength-strings -Wno-discarded-qualifiers
+LFLAGS = -T link.ld
 
 CFILES = $(wildcard src/*.c)
 AFILES = $(wildcard src/*.s) 
@@ -17,16 +20,15 @@ OUTDIR = out
 OFILES = $(patsubst src/%.c, $(OUTDIR)/%.o, $(CFILES)) \
 	 $(patsubst src/%.s, $(OUTDIR)/%.asm.o, $(AFILES))
 
-HEX = main.hex
+OUT = out/main.elf
 
-all: $(HEX)
+all: $(OUT)
 
-$(HEX): $(OFILES) initrd/init
-	@echo "  LINK   " $(HEX)
+$(OUT): $(OFILES) initrd/init libinterp.a
+	@echo "  LINK   " $(OUT)
 	@./mkinitrd.sh
 	@$(CROSS)$(OBJCOPY) -B arm -I binary -O elf32-littlearm initrd.img out/initrd.img.o
-	@$(CROSS)$(CC) $(CFLAGS) $(LFLAGS) -T link.ld out/*.o -o out/main.elf -L. -linterp
-	@$(CROSS)$(OBJCOPY) $(OFLAGS) out/main.elf $(HEX)
+	@$(CROSS)$(CC) $(CFLAGS) $(LFLAGS) out/*.o -o $(OUT) -L. -linterp
 
 $(OUTDIR)/%.o: src/%.c
 	@echo "  CC     " $<
diff --git a/include/heap.h b/include/heap.h
index 8118384..c93eb8e 100644
--- a/include/heap.h
+++ b/include/heap.h
@@ -4,7 +4,6 @@
 #include <stdint.h>
 
 void heap_init(void *buf);
-uint32_t heap_used(void);
 
 void *malloc(uint32_t size);
 void *calloc(uint32_t count, uint32_t size);
diff --git a/include/random.h b/include/random.h
new file mode 100644
index 0000000..6485f26
--- /dev/null
+++ b/include/random.h
@@ -0,0 +1,9 @@
+#ifndef RANDOM_H_
+#define RANDOM_H_
+
+#include <stdint.h>
+
+void random_init(void);
+uint32_t random_get(void);
+
+#endif // RANDOM_H_
diff --git a/include/script.h b/include/script.h
new file mode 100644
index 0000000..dc8c590
--- /dev/null
+++ b/include/script.h
@@ -0,0 +1,8 @@
+#ifndef SCRIPT_H_
+#define SCRIPT_H_
+
+#include <parser.h>
+
+void script_loadlib(interpreter *it);
+
+#endif // SCRIPT_H_
diff --git a/include/stdlib.h b/include/stdlib.h
new file mode 100644
index 0000000..c2658d6
--- /dev/null
+++ b/include/stdlib.h
@@ -0,0 +1,8 @@
+#ifndef STDLIB_H_
+#define STDLIB_H_
+
+char *snprintf(char *buf, unsigned int max, const char *format, ...);
+float strtof(const char *s, char **endptr);
+int atoi(const char *s);
+
+#endif // STDLIB_H_
diff --git a/include/task.h b/include/task.h
index 7077b00..1ef8639 100644
--- a/include/task.h
+++ b/include/task.h
@@ -8,6 +8,12 @@
 
 #include <stdint.h>
 
+typedef struct {
+	void *next;
+	uint32_t *stack;
+	uint32_t *sp;
+} task_t;
+
 /**
  * Enters multitasking mode. The given function acts as the initial thread.
  * This task is given a 4kb stack.
diff --git a/initrd/init b/initrd/init
index 694c151..9f6a9cf 100644
--- a/initrd/init
+++ b/initrd/init
@@ -1,33 +1,12 @@
-print "Hello."
-
-set fg 32767
-
-# draw bg, lines
-rect 50 50 380 220 6375
-line 50 160 430 160 fg
-line 240 50 240 270 fg
-
-set x 50
-do
-	line x 170 x 150 fg
-	set x (x + 20)
-while (x < 430)
-
-set y 50
 do
-	line 230 y 250 y fg
-	set y (y + 20)
-while (y < 270)
+	rand 479 > x
+	rand 319 > y
+	rand 479 > i
+	rand 319 > j
+	rand 32767 > purple
+		
+	line x y i j purple
+while (1)
 
-#line 80 250 380 90 511
-set purple 511
-do
-	rand 219 > x
-	rand 379 > y
-	rand 219 > i
-	rand 379 > j
+print "done"
 
-	line (x + 50) (y + 50) (i + 50) (j + 50) purple
-	delay 1000
-	set purple (purple + 11)
-while (1)
diff --git a/libinterp.a b/libinterp.a
index 7884542..561bb11 100644
Binary files a/libinterp.a and b/libinterp.a differ
diff --git a/run.sh b/run.sh
index 318f1ff..f628d56 100755
--- a/run.sh
+++ b/run.sh
@@ -1,7 +1,4 @@
 #!/bin/bash
 
-#openocd -f /usr/share/openocd/scripts/board/st_nucleo_l476rg.cfg \
-#	-c "init; reset halt; flash write_image erase main.hex; reset run; exit"
-
 openocd -f /usr/share/openocd/scripts/board/st_nucleo_l476rg.cfg > /dev/null &
 gdb-multiarch -iex "target remote localhost:3333" out/main.elf
diff --git a/src/clock.c b/src/clock.c
index 7cb4ecf..3d31f91 100644
--- a/src/clock.c
+++ b/src/clock.c
@@ -22,8 +22,8 @@ void clock_init(void)
 	RCC->PLLCFGR |= RCC_PLLCFGR_PLLSRC_HSI;
 	RCC->PLLCFGR &= ~(RCC_PLLCFGR_PLLN | RCC_PLLCFGR_PLLM);
 	RCC->PLLCFGR |= 10 << RCC_PLLCFGR_PLLN_Pos;
-	RCC->PLLCFGR &= ~(RCC_PLLCFGR_PLLR);	// /2
-	RCC->PLLCFGR |= RCC_PLLCFGR_PLLREN;		// PLLR on
+	RCC->PLLCFGR &= ~(RCC_PLLCFGR_PLLR | RCC_PLLCFGR_PLLQ); // /2
+	RCC->PLLCFGR |= RCC_PLLCFGR_PLLREN | RCC_PLLCFGR_PLLQEN;
 
 	// start PLL
 	RCC->CR |= RCC_CR_PLLON;
@@ -54,7 +54,7 @@ void SysTick_Handler(void)
 	// just keep counting
 	ticks++;
 
-	if (!(ticks % 4))
+	if (!(ticks & 3))
 		SCB->ICSR |= SCB_ICSR_PENDSVSET_Msk;
 
 	asm("mov lr, %0; bx lr" :: "r" (lr));
diff --git a/src/display.c b/src/display.c
index 955cf8f..163a742 100644
--- a/src/display.c
+++ b/src/display.c
@@ -117,19 +117,19 @@ void dsp_init(void)
 	gpio_mode(LCD_WR, OUTPUT);
 	gpio_mode(LCD_RST, OUTPUT);
 	dsp_dmode(OUTPUT);
-	gpio_speed(LCD_CS, LOW);
-	gpio_speed(LCD_RS, LOW);
-	gpio_speed(LCD_RD, LOW);
-	gpio_speed(LCD_WR, LOW);
-	gpio_speed(LCD_RST, LOW);
-	gpio_speed(LCD_D0, LOW);
-	gpio_speed(LCD_D1, LOW);
-	gpio_speed(LCD_D2, LOW);
-	gpio_speed(LCD_D3, LOW);
-	gpio_speed(LCD_D4, LOW);
-	gpio_speed(LCD_D5, LOW);
-	gpio_speed(LCD_D6, LOW);
-	gpio_speed(LCD_D7, LOW);
+	gpio_speed(LCD_CS, VERYHIGH);
+	gpio_speed(LCD_RS, VERYHIGH);
+	gpio_speed(LCD_RD, VERYHIGH);
+	gpio_speed(LCD_WR, VERYHIGH);
+	gpio_speed(LCD_RST, VERYHIGH);
+	gpio_speed(LCD_D0, VERYHIGH);
+	gpio_speed(LCD_D1, VERYHIGH);
+	gpio_speed(LCD_D2, VERYHIGH);
+	gpio_speed(LCD_D3, VERYHIGH);
+	gpio_speed(LCD_D4, VERYHIGH);
+	gpio_speed(LCD_D5, VERYHIGH);
+	gpio_speed(LCD_D6, VERYHIGH);
+	gpio_speed(LCD_D7, VERYHIGH);
 	gpio_dout(LCD_CS, 0);
 	gpio_dout(LCD_RS, 1);
 	gpio_dout(LCD_RD, 1);
diff --git a/src/display_draw.c b/src/display_draw.c
index be4becf..27ed796 100644
--- a/src/display_draw.c
+++ b/src/display_draw.c
@@ -33,6 +33,7 @@ void dsp_cursoron(void)
 
 void dsp_putchar(int c)
 {
+	LOCK;
 	if (c == '\n') {
 		curx = 0;
 		if (++cury == 12) {
@@ -65,15 +66,14 @@ void dsp_putchar(int c)
 			cury = 0;
 		}
 	}
+	UNLOCK;
 }
 
 void dsp_puts(const char *s)
 {
 	unsigned int i = 0;
-	LOCK;
 	while (s[i])
 		dsp_putchar(s[i++]);
-	UNLOCK;
 }
 
 void dsp_cpos(int x, int y)
@@ -90,14 +90,14 @@ void dsp_coff(int x, int y)
 
 void dsp_rect(int x, int y, int w, int h, uint16_t color)
 {
-	LOCK;
 	dsp_set_addr(x, y, x + w - 1, y + h - 1);
 	int countdown = w * h;
 	do {
+		LOCK;
 		dsp_write_data(color >> 8);
 		dsp_write_data(color & 0xFF);
+		UNLOCK;
 	} while (countdown--);
-	UNLOCK;
 }
 
 void dsp_line(int x, int y, int i, int j, uint16_t color)
@@ -114,11 +114,12 @@ void dsp_line(int x, int y, int i, int j, uint16_t color)
 	int err = (dx > dy ? dx : -dy) / 2;
 	int e2;
 
-	LOCK;
 	while (1) {
+		LOCK;
 		dsp_set_addr(x, y, x, y);
 		dsp_write_data(color >> 8);
 		dsp_write_data(color & 0xFF);
+		UNLOCK;
 		if (x == i && y == j)
 			break;
 		e2 = err;
@@ -131,6 +132,5 @@ void dsp_line(int x, int y, int i, int j, uint16_t color)
 			y += sy;
 		}
 	}
-	UNLOCK;
 }
 
diff --git a/src/heap.c b/src/heap.c
index 0502e76..92978b7 100644
--- a/src/heap.c
+++ b/src/heap.c
@@ -4,54 +4,52 @@
 #define HEAP_ALIGN 16
 
 typedef struct {
-	uint32_t next;
 	uint32_t size;
+	void *next;
 } __attribute__ ((packed)) alloc_t;
 
-static alloc_t root;
+static alloc_t *free_blocks;
 static void *heap_end;
+static uint32_t heap_used;
 
 void heap_init(void *buf)
 {
 	heap_end = buf;
-	root.next = 1;
-	root.size = 0;
-	// what to do...
-}
-
-uint32_t heap_used(void)
-{
-	uint32_t total = 0;
-	alloc_t *a = &root;
-	while (a->next > 1) {
-		total += a->size;
-		a = (void *)(a->next & ~(1));
-	}
-	return total;
+	free_blocks = 0;
+	heap_used = 0;
 }
 
 void *malloc(uint32_t size)
 {
-	task_hold(1);
-	if (size < HEAP_ALIGN)
-		size = HEAP_ALIGN;
-	alloc_t *node = &root;
-	while (node->next & 1 || node->size < size) {
-		if ((node->next & ~(1)) == 0) {
-			node->next |= (uint32_t)(heap_end + HEAP_ALIGN) & ~(HEAP_ALIGN - 1);
-			heap_end += HEAP_ALIGN + size;
-			node = (void *)(node->next & ~(1));
+	//task_hold(1);
+	size = (size + sizeof(alloc_t) + HEAP_ALIGN) & ~(HEAP_ALIGN - 1);
+
+	alloc_t *node = free_blocks;
+	alloc_t *prev = 0;
+	while (node != 0) {
+		if (node->size >= size) {
+			if (prev != 0)
+				prev->next = node->next;
+			else
+				free_blocks = node->next;
 			node->next = 0;
-			node->size = size;
-			break;
+			heap_used += node->size;
+			return (void *)((uint8_t *)node + sizeof(alloc_t));
 		}
-		node = (void *)(node->next & ~(1));
+
+		prev = node;
+		node = node->next;
 	}
 
-	node->next |= 1;
-	task_hold(0);
+	node = (alloc_t *)heap_end;
+	node->size = size;
+	node->next = 0;
 
-	return (void *)((uint32_t)node + sizeof(alloc_t));
+	heap_end = (void *)((uint8_t *)heap_end + size);
+	heap_used += size;
+
+	//task_hold(0);
+	return (void *)((uint8_t *)node + sizeof(alloc_t)); 
 }
 
 void *calloc(uint32_t count, uint32_t size)
@@ -66,6 +64,9 @@ void free(void *buf)
 {
 	if (buf == 0)
 		return;
-	alloc_t *alloc = (alloc_t *)((uint32_t)buf - sizeof(alloc_t));
-	alloc->next &= ~(1);
+
+	alloc_t *alloc = (alloc_t *)((uint8_t *)buf - sizeof(alloc_t));
+	heap_used -= alloc->size;
+	alloc->next = free_blocks;
+	free_blocks = alloc;
 }
diff --git a/src/initrd.c b/src/initrd.c
index 54bdc4b..e90f9f0 100644
--- a/src/initrd.c
+++ b/src/initrd.c
@@ -53,7 +53,7 @@ uint32_t initrd_getsize(initrd_file *file)
 
 initrd_file *initrd_getfileptr(const char *name)
 {
-	initrd_file *file = (initrd_file *)(initrd_start + sizeof(initrd_header));
+	initrd_file *file = (initrd_file *)((uint8_t *)initrd_start + sizeof(initrd_header));
 	uint32_t offset = sizeof(initrd_header);
 
 	while (offset < initrd_size) {
@@ -74,7 +74,7 @@ char *initrd_getfile(const char *name)
 		return 0;
 
 
-	char *ptr = (char *)((void *)file + sizeof(initrd_file));
+	char *ptr = (char *)file + sizeof(initrd_file);
 	ptr[initrd_getsize(file) - 1] = 0;
 	return ptr;
 }
diff --git a/src/main.c b/src/main.c
index a946609..54c2403 100644
--- a/src/main.c
+++ b/src/main.c
@@ -7,160 +7,65 @@
 #include <display.h>
 #include <display_draw.h>
 #include <initrd.h>
-#include <serial.h>
 #include <parser.h>
-#include <builtins.h>
-#include <stack.h>
+#include <serial.h>
 #include <string.h>
+#include <script.h>
+#include <random.h>
 
+extern uint8_t _ebss;
 extern char *itoa(int, char *, int);
 
 void kmain(void);
+void task_interpreter(void);
 
 int main(void)
 {
 	asm("cpsid i");
+	// disable write buffer
+	*((uint32_t *)0xE000E008) |= 2;
 
-	// prepare flash latency for 40MHz operation
+	// prepare flash latency for 80MHz operation
 	FLASH->ACR &= ~(FLASH_ACR_LATENCY);
 	FLASH->ACR |= FLASH_ACR_LATENCY_4WS;
 
 	//MPU->CTRL |= MPU_CTRL_ENABLE_Msk | MPU_CTRL_PRIVDEFENA_Msk;
 	clock_init();
-	extern uint8_t _ebss;
 	heap_init(&_ebss);
 	gpio_init();
+	serial_init();
+	random_init();
 
 	gpio_mode(GPIOA, 5, OUTPUT);
 
-	serial_init();
-
 	// enable FPU
 	SCB->CPACR |= (0xF << 20);
 
 	task_init(kmain);
-
 	while (1);
 }
 
-int script_puts(interpreter *it)
-{
-	char *s = igetarg_string(it, 0);
-	dsp_puts(s);
-	//dsp_puts("\n");
-	//asm("mov r0, %0; svc 2" :: "r" (s));
-	return 0;
-}
-
-int script_gets(interpreter *it)
-{
-	char *s = malloc(64), c[2] = {0, 0};
-	uint16_t index = 0;
-
-	do {
-		c[0] = serial_get();
-		s[index] = c[0];
-		if (c[0] != '\r')
-			dsp_puts(c);
-	} while (s[index] != '\r' && index++ < 23);
-	s[index] = '\0';
-
-	variable *v = igetarg(it, 0);
-	v->valtype = STRING;
-	v->svalue = s;
-	return 0;
-}
-
-int script_delay(interpreter *it)
-{
-	int ms = igetarg_integer(it, 0);
-	delay(ms);
-	return 0;
-}
-
-int script_rect(interpreter *it)
-{
-	dsp_rect(igetarg_integer(it, 0), igetarg_integer(it, 1),
-		igetarg_integer(it, 2), igetarg_integer(it, 3),
-		igetarg_integer(it, 4));
-	return 0;
-}
-
-int script_line(interpreter *it)
-{
-	dsp_line(igetarg_integer(it, 0), igetarg_integer(it, 1),
-		igetarg_integer(it, 2), igetarg_integer(it, 3),
-		igetarg_integer(it, 4));
-	return 0;
-}
-
-int script_ppos(interpreter *it)
-{
-	dsp_cpos(0, 0);
-	dsp_coff(igetarg_integer(it, 0), igetarg_integer(it, 1));
-	return 0;
-}
-
-int script_color(interpreter *it)
+void kmain(void)
 {
-	uint16_t c = dsp_color(igetarg_integer(it, 0), igetarg_integer(it, 1),
-		igetarg_integer(it, 2));
-	variable v;
-	v.valtype = INTEGER;
-	INT(&v) = c;
-	v.svalue = 0;
-	isetstr(&v);
-	iret(it, &v);
-	return 0;
-}
+	asm("cpsie i");
+	dsp_init();
+	dsp_rect(0, 0, LCD_WIDTH, LCD_HEIGHT, dsp_color(0, 0, 0));
+	dsp_cursoron();
+	task_start(task_interpreter, 4096);
 
-int script_rand(interpreter *it)
-{
-	static uint32_t next = 1;
-	next = (next * 182 + 1829) % igetarg_integer(it, 0);
-	variable v;
-	v.valtype = INTEGER;
-	INT(&v) = next;
-	v.svalue = 0;
-	isetstr(&v);
-	iret(it, &v);
-	return 0;
+	while (1) {
+		gpio_dout(GPIOA, 5, 1);
+		delay(250);
+		gpio_dout(GPIOA, 5, 0);
+		delay(250);
+	}
 }
 
 void task_interpreter(void)
 {
 	interpreter it;
 	iinit(&it);
-	inew_cfunc(&it, "print", script_puts);
-	inew_cfunc(&it, "gets", script_gets);
-	inew_cfunc(&it, "delay", script_delay);
-	inew_cfunc(&it, "rect", script_rect);
-	inew_cfunc(&it, "ppos", script_ppos);
-	inew_cfunc(&it, "line", script_line);
-	inew_cfunc(&it, "color", script_color);
-	inew_cfunc(&it, "rand", script_rand);
-
-	/*int ret = 0;
-	char *linebuf = malloc(100), c[2] = {0, 0};
-	while (1) {
-		uint16_t index = 0;
-		if (it.indent > 0)
-			dsp_puts(">");
-		dsp_puts("> ");
-		do {
-			c[0] = serial_get();
-			if (c[0] >= ' ' || c[0] == '\r') {
-				linebuf[index] = c[0];
-				if (c[0] >= ' ')
-					dsp_puts(c);
-			}
-		} while (linebuf[index] != '\r' && index++ < 100);
-		linebuf[index] = '\0';
-		dsp_puts("\n");
-		ret = idoline(&it, linebuf);
-		if (ret < 0)
-			break;
-	}*/
+	script_loadlib(&it);
 
 	char *s = initrd_getfile("init");
 	if (s == 0) {
@@ -171,7 +76,7 @@ void task_interpreter(void)
 	char *linebuf = (char *)malloc(120);
 	uint32_t i = 0, prev = 0, lc;
 	uint32_t size = initrd_getfilesize("init");
-	int ret;
+	int ret = 0;
 	while (i < size) {
 		for (; s[i] != '\n' && s[i] != '\0'; i++);
 		lc = i - prev;
@@ -181,9 +86,7 @@ void task_interpreter(void)
 		}
 		strncpy(linebuf, s + prev, lc + 1);
 		linebuf[lc] = '\0';
-		//task_hold(1);
 		ret = idoline(&it, linebuf);
-		//task_hold(0);
 		if (ret < 0)
 			break;
 		prev = ++i;
@@ -201,20 +104,26 @@ end:
 		delay(10);
 }
 
-void kmain(void)
-{
-	asm("cpsie i");
-	dsp_init();
-	dsp_rect(0, 0, LCD_WIDTH, LCD_HEIGHT, dsp_color(0, 0, 0));
-	dsp_cursoron();
-	//dsp_puts("Hey.");
-	task_start(task_interpreter, 4096);
-
-	while (1) {
-		gpio_dout(GPIOA, 5, 1);
-		delay(500);
-		gpio_dout(GPIOA, 5, 0);
-		delay(500);
-	}
-}
+// for interactive use
+/*int ret = 0;
+char *linebuf = malloc(100), c[2] = {0, 0};
+while (1) {
+	uint16_t index = 0;
+	if (it.indent > 0)
+		dsp_puts(">");
+	dsp_puts("> ");
+	do {
+		c[0] = serial_get();
+		if (c[0] >= ' ' || c[0] == '\r') {
+			linebuf[index] = c[0];
+			if (c[0] >= ' ')
+				dsp_puts(c);
+		}
+	} while (linebuf[index] != '\r' && index++ < 100);
+	linebuf[index] = '\0';
+	dsp_puts("\n");
+	ret = idoline(&it, linebuf);
+	if (ret < 0)
+		break;
+}*/
 
diff --git a/src/main.c.bak b/src/main.c.bak
deleted file mode 100644
index 95f3210..0000000
--- a/src/main.c.bak
+++ /dev/null
@@ -1,203 +0,0 @@
-#include <stm32l476xx.h>
-#include <clock.h>
-#include <heap.h>
-#include <task.h>
-#include <gpio.h>
-#include <lcd.h>
-#include <display.h>
-#include <display_draw.h>
-#include <initrd.h>
-#include <serial.h>
-#include <parser.h>
-#include <stack.h>
-
-#include <stdlib.h>
-#include <string.h>
-
-void kmain(void);
-void task_interpreter(void);
-
-int main(void)
-{
-	asm("cpsid i");
-
-	// prepare flash latency for 40MHz operation
-	FLASH->ACR &= ~(FLASH_ACR_LATENCY);
-	FLASH->ACR |= FLASH_ACR_LATENCY_4WS;
-
-	//MPU->CTRL |= MPU_CTRL_ENABLE_Msk | MPU_CTRL_PRIVDEFENA_Msk;
-	clock_init();
-	gpio_init();
-
-	gpio_mode(GPIOA, 5, OUTPUT);
-
-	serial_init();
-
-	// enable FPU
-	SCB->CPACR |= (0xF << 20);
-
-	task_init(kmain);
-
-	while (1);
-}
-
-int script_puts(interpreter *it)
-{
-	char *s = igetarg_string(it, 0);
-	dsp_puts(s);
-	//dsp_puts("\n");
-	//asm("mov r0, %0; svc 2" :: "r" (s));
-	return 0;
-}
-
-int script_gets(interpreter *it)
-{
-	char *s = malloc(64), c[2] = {0, 0};
-	uint16_t index = 0;
-
-	do {
-		c[0] = serial_get();
-		s[index] = c[0];
-		if (c[0] != '\r')
-			dsp_puts(c);
-	} while (s[index] != '\r' && index++ < 23);
-	s[index] = '\0';
-
-	variable *v = igetarg(it, 0);
-	v->valtype = STRING;
-	v->svalue = s;
-	return 0;
-}
-
-int script_delay(interpreter *it)
-{
-	int ms = igetarg_integer(it, 0);
-	delay(ms);
-	return 0;
-}
-
-int script_rect(interpreter *it)
-{
-	dsp_rect(igetarg_integer(it, 0), igetarg_integer(it, 1),
-		igetarg_integer(it, 2), igetarg_integer(it, 3),
-		igetarg_integer(it, 4));
-	return 0;
-}
-
-int script_line(interpreter *it)
-{
-	dsp_line(igetarg_integer(it, 0), igetarg_integer(it, 1),
-		igetarg_integer(it, 2), igetarg_integer(it, 3),
-		igetarg_integer(it, 4));
-	return 0;
-}
-
-int script_ppos(interpreter *it)
-{
-	dsp_cpos(0, 0);
-	dsp_coff(igetarg_integer(it, 0), igetarg_integer(it, 1));
-	return 0;
-}
-
-void task_interpreter(void)
-{
-	interpreter it;
-	iinit(&it);
-	inew_cfunc(&it, "print", script_puts);
-	inew_cfunc(&it, "gets", script_gets);
-	inew_cfunc(&it, "delay", script_delay);
-	inew_cfunc(&it, "rect", script_rect);
-	inew_cfunc(&it, "ppos", script_ppos);
-	inew_cfunc(&it, "line", script_line);
-
-	/*int ret = 0;
-	char *linebuf = malloc(100), c[2] = {0, 0};
-	while (1) {
-		uint16_t index = 0;
-		if (it.indent > 0)
-			dsp_puts(">");
-		dsp_puts("> ");
-		do {
-			c[0] = serial_get();
-			if (c[0] >= ' ' || c[0] == '\r') {
-				linebuf[index] = c[0];
-				if (c[0] >= ' ')
-					dsp_puts(c);
-			}
-		} while (linebuf[index] != '\r' && index++ < 100);
-		linebuf[index] = '\0';
-		dsp_puts("\n");
-		ret = idoline(&it, linebuf);
-		if (ret < 0)
-			break;
-	}*/
-
-	char *s = initrd_getfile("init");
-	if (s == 0)
-		goto end;
-
-	char *linebuf = (char *)malloc(120);
-	uint32_t i = 0, prev = 0, lc;
-	uint32_t size = initrd_getfilesize("init");
-	int ret;
-	while (i < size) {
-		for (; s[i] != '\n' && s[i] != '\0'; i++);
-		lc = i - prev;
-		if (lc == 0) {
-			prev = ++i;
-			continue;
-		}
-		strncpy(linebuf, s + prev, lc + 1);
-		linebuf[lc] = '\0';
-		ret = idoline(&it, linebuf);
-		if (ret < 0)
-			break;
-		prev = ++i;
-	}
-
-	if (ret < 0) {
-		dsp_puts("\nError: ");
-		dsp_puts(itoa(ret, linebuf, 10));
-	}
-	free(linebuf);
-	//iend(&it); // nah
-
-end:
-	while (1)
-		delay(10);
-}
-
-void kmain(void)
-{
-	asm("cpsie i");
-
-	dsp_init();
-
-	//dsp_rect(0, 0, 40, 40, dsp_color(0x7F, 0, 0x7F));
-
-	//dsp_set_addr_read(0, 0, 39, 39);
-	//dsp_dmode(INPUT);
-	//uint8_t *buf = (uint8_t *)malloc(40 * 40 * 2);
-	//for (int i = 0; i < 180; i++)
-	//	buf[i] = dsp_read_data();
-	//dsp_dmode(OUTPUT);
-	//dsp_set_addr(40, 40, 79, 79);
-	//for (int i = 0; i < 320; i++)
-	//	dsp_write_data(buf[i]);
-	//dsp_rect(80, 80, 40, 40, dsp_color(0x7F, 0x7F, 0));
-
-	dsp_rect(0, 0, LCD_WIDTH, LCD_HEIGHT, dsp_color(0, 0, 0));
-	dsp_cursoron();
-
-	task_start(task_interpreter, 4096);
-
-	//char *s = initrd_getfile("test.txt");
-
-	while (1) {
-		gpio_dout(GPIOA, 5, 1);
-		delay(500);
-		gpio_dout(GPIOA, 5, 0);
-		delay(500);
-	}
-}
-
diff --git a/src/random.c b/src/random.c
new file mode 100644
index 0000000..9c1e0d1
--- /dev/null
+++ b/src/random.c
@@ -0,0 +1,24 @@
+#include <random.h>
+
+#include <stm32l476xx.h>
+#include <clock.h>
+
+void random_init(void)
+{
+	// setup and enable RNG clock
+	RCC->CCIPR &= ~(RCC_CCIPR_CLK48SEL);
+	RCC->CCIPR |= RCC_CCIPR_CLK48SEL_1;
+	RCC->AHB2ENR |= RCC_AHB2ENR_RNGEN;
+	RNG->CR |= RNG_CR_RNGEN;
+}
+
+uint32_t random_get(void)
+{
+	if (RNG->SR & (RNG_SR_SEIS | RNG_SR_CEIS))
+		return 0;
+
+	while (!(RNG->SR & RNG_SR_DRDY))
+		delay(1);
+
+	return RNG->DR;
+}
diff --git a/src/script.c b/src/script.c
new file mode 100644
index 0000000..adf402b
--- /dev/null
+++ b/src/script.c
@@ -0,0 +1,120 @@
+#include <script.h>
+
+#include <builtins.h>
+#include <clock.h>
+#include <display.h>
+#include <display_draw.h>
+#include <heap.h>
+#include <random.h>
+#include <serial.h>
+#include <stack.h>
+
+int script_puts(interpreter *it);
+int script_gets(interpreter *it);
+int script_delay(interpreter *it);
+int script_rect(interpreter *it);
+int script_ppos(interpreter *it);
+int script_line(interpreter *it);
+int script_color(interpreter *it);
+int script_rand(interpreter *it);
+
+void script_loadlib(interpreter *it)
+{
+	inew_cfunc(it, "print", script_puts);
+	inew_cfunc(it, "gets", script_gets);
+	inew_cfunc(it, "delay", script_delay);
+	inew_cfunc(it, "rect", script_rect);
+	inew_cfunc(it, "ppos", script_ppos);
+	inew_cfunc(it, "line", script_line);
+	inew_cfunc(it, "color", script_color);
+	inew_cfunc(it, "rand", script_rand);
+}
+
+int script_puts(interpreter *it)
+{
+	char *s = igetarg_string(it, 0);
+	dsp_puts(s);
+	//dsp_puts("\n");
+	//asm("mov r0, %0; svc 2" :: "r" (s));
+	return 0;
+}
+
+int script_gets(interpreter *it)
+{
+	char *s = malloc(64), c[2] = {0, 0};
+	uint16_t index = 0;
+
+	do {
+		c[0] = serial_get();
+		s[index] = c[0];
+		if (c[0] != '\r')
+			dsp_puts(c);
+	} while (s[index] != '\r' && index++ < 23);
+	s[index] = '\0';
+
+	variable *v = igetarg(it, 0);
+	v->valtype = STRING;
+	v->svalue = s;
+	return 0;
+}
+
+int script_delay(interpreter *it)
+{
+	int ms = igetarg_integer(it, 0);
+	delay(ms);
+	return 0;
+}
+
+int script_rect(interpreter *it)
+{
+	dsp_rect(igetarg_integer(it, 0), igetarg_integer(it, 1),
+		igetarg_integer(it, 2), igetarg_integer(it, 3),
+		igetarg_integer(it, 4));
+	return 0;
+}
+
+int script_line(interpreter *it)
+{
+	int x = igetarg_integer(it, 0);
+	int y = igetarg_integer(it, 1);
+	int i = igetarg_integer(it, 2);
+	int j = igetarg_integer(it, 3);
+	int c = igetarg_integer(it, 4);
+	dsp_line(x, y, i, j, c);
+	return 0;
+}
+
+int script_ppos(interpreter *it)
+{
+	dsp_cpos(0, 0);
+	dsp_coff(igetarg_integer(it, 0), igetarg_integer(it, 1));
+	return 0;
+}
+
+int script_color(interpreter *it)
+{
+	uint16_t c = dsp_color(igetarg_integer(it, 0), igetarg_integer(it, 1),
+		igetarg_integer(it, 2));
+	variable v;
+	v.valtype = INTEGER;
+	INT(&v) = c;
+	v.svalue = 0;
+	isetstr(&v);
+	iret(it, &v);
+	return 0;
+}
+
+int script_rand(interpreter *it)
+{
+	variable *v = (variable *)malloc(sizeof(variable));
+	unsigned int mod = igetarg_integer(it, 0);
+	unsigned int val = random_get();
+	v->valtype = INTEGER;
+	INT(v) = val % mod;
+	v->svalue = 0;
+	isetstr(v);
+	iret(it, v);
+	free(v->svalue);
+	free(v);
+	return 0;
+}
diff --git a/src/stdlib.c b/src/stdlib.c
index 3f11d57..303db4d 100644
--- a/src/stdlib.c
+++ b/src/stdlib.c
@@ -1,6 +1,80 @@
+#include <stdlib.h>
+
+#include <ctype.h>
+#include <stdarg.h>
+#include <stdint.h>
+#include <string.h>
+
+extern char *itoa(int, char *, int);
+
 void _exit(int code)
 {
 	(void)code;
 	for (;;);
 }
 
+char *snprintf(char *buf, unsigned int max, const char *format, ...)
+{
+	va_list args;
+	va_start(args, format);
+
+	if (buf == 0 || max == 0)
+		return 0;
+
+	static char nbuf[32];
+	unsigned int off = 0;
+	for (unsigned int i = 0; format[i] != '\0' && off < max; i++) {
+		if (format[i] == '%') {
+			i++;
+			nbuf[0] = '\0';
+			switch (format[i]) {
+			case 'd':
+				itoa(va_arg(args, int), nbuf, 10);
+				break;
+			case 'u':
+				itoa(va_arg(args, unsigned int), nbuf, 10);
+				break;
+			case 'x':
+			case 'p':
+				itoa(va_arg(args, unsigned int), nbuf, 16);
+				break;
+			case 'f':
+				itoa((int)va_arg(args, double), nbuf, 10);
+				break;
+			default:
+				buf[off++] = format[i];
+				nbuf[0] = '\0';
+				break;
+			}
+			if (nbuf[0] != '\0') {
+				for (unsigned int j = 0; off < max &&
+					nbuf[j] != '\0'; off++, j++)
+					buf[off] = nbuf[j];
+			}
+		} else {
+			buf[off++] = format[i];
+		}
+	}
+
+
+	buf[off >= max ? max - 1 : off] = '\0';
+	va_end(args);
+	return buf;
+}
+
+float strtof(const char *s, char **endptr)
+{
+	(void)s;
+	(void)endptr;
+	return 0.0f;
+}
+
+int atoi(const char *s)
+{
+	int n = 0;
+	for (unsigned int i = 0; isdigit(s[i]); i++) {
+		n *= 10;
+		n += s[i] - '0';
+	}
+	return n;
+}
diff --git a/src/stm32l4xx_it.c b/src/stm32l4xx_it.c
index ef10406..f643db3 100644
--- a/src/stm32l4xx_it.c
+++ b/src/stm32l4xx_it.c
@@ -1,20 +1,37 @@
 #include <stm32l476xx.h>
-#include <lcd.h>
+#include <serial.h>
+#include <stdlib.h>
+#include <task.h>
+
+void serial_puts(const char *s)
+{
+	int i = 0;
+	while (s[i] != 0)
+		serial_put(s[i++]);
+}
 
 void perror(const char *s)
 {
-	(void)s;//lcd_puts(s);
+	extern task_t *current;
+	serial_puts(s);
+	char buf[200];
+	snprintf(buf, 200, "xPSR: %x\r\nPC: %x\r\nLR: %x\r\n", current->sp[0],
+		current->sp[1], current->sp[2]);
+	serial_puts(buf);
 }
 
+__attribute__ ((naked))
 void NMI_Handler(void) {}
 
+__attribute__ ((naked))
 void HardFault_Handler(void)
 {
 	GPIOA->BSRR |= (1 << 5);
-	perror("Hard Fault!");
+	//perror("Hard Fault!");
 	while (1);
 }
 
+__attribute__ ((naked))
 void MemManage_Handler(void)
 {
 	GPIOA->BSRR |= (1 << 5);
@@ -22,6 +39,7 @@ void MemManage_Handler(void)
 	while (1);
 }
 
+__attribute__ ((naked))
 void BusFault_Handler(void)
 {
 	GPIOA->BSRR |= (1 << 5);
@@ -29,6 +47,7 @@ void BusFault_Handler(void)
 	while (1);
 }
 
+__attribute__ ((naked))
 void UsageFault_Handler(void)
 {
 	GPIOA->BSRR |= (1 << 5);
@@ -36,5 +55,6 @@ void UsageFault_Handler(void)
 	while (1);
 }
 
+__attribute__ ((naked))
 void DebugMon_Handler(void) {}
 
diff --git a/src/task.c b/src/task.c
index d0cb3f1..6a755de 100644
--- a/src/task.c
+++ b/src/task.c
@@ -2,13 +2,7 @@
 #include <heap.h>
 #include <stm32l476xx.h>
 
-typedef struct {
-	void *next;
-	uint32_t *stack;
-	uint32_t *sp;
-} task_t;
-
-static task_t *current;
+task_t *current;
 static uint8_t task_disable = 0;
 
 void task_hold(uint8_t hold)
@@ -32,10 +26,20 @@ task_t *task_create(void (*code)(void), uint32_t stackSize)
 	task_t *t = (task_t *)malloc(sizeof(task_t));
 	t->next = 0;
 	t->stack = (uint32_t *)malloc(stackSize);
-	t->sp = (void *)t->stack + stackSize - 64;//16;
-	t->sp[13] = (uint32_t)task_exit;
-	t->sp[14] = (uint32_t)code;
-	t->sp[15] = 0x01000000;
+	void *sp = (uint8_t *)t->stack + stackSize - 68; // excep. stack + regs
+	t->sp = (uint32_t *)sp;
+	for (uint8_t i = 0; i < 14; i++)
+		t->sp[i] = 0;
+	t->sp[8] = 0xFFFFFFFD;
+	t->sp[14] = 0xFFFFFFFD;
+	t->sp[15] = (uint32_t)code;
+	t->sp[16] = 0x01000000;
+
+	//void *sp = (uint8_t *)t->stack + stackSize - 64; // 16 words
+	//t->sp = (uint32_t *)sp;
+	//t->sp[13] = (uint32_t)task_exit;
+	//t->sp[14] = (uint32_t)code;
+	//t->sp[15] = 0x01000000;
 	return t;
 }
 
@@ -43,6 +47,8 @@ void task_init(void (*init)(void))
 {
 	current = task_create(init, 4096);
 	current->next = current;
+	task_disable = 0;
+
 	// bit 0 - priv, bit 1 - psp/msp
 	asm("\
 		msr psp, %0; \
@@ -50,14 +56,8 @@ void task_init(void (*init)(void))
 		orr r0, r0, #2; \
 		msr control, r0; \
 		isb; \
-	" :: "r" (current->sp));
-
-	task_disable = 0;
-	init();
-	/*asm("\
-		cpsie i; \
-		mov pc, %0; \
-	" :: "r" (init + 4));*/
+		bx %1; \
+	" :: "r" (current->sp), "r" (init));
 }
 
 void task_start(void (*task)(void), uint16_t stackSize)
@@ -75,13 +75,29 @@ void PendSV_Handler(void)
 	if (task_disable != 0)
 		asm("bx lr");
 
-	// save state
+	asm("\
+		mrs r0, psp; \
+		isb; \
+		ldr r1, =current; \
+		ldr r2, [r1]; \
+		stmdb r0!, {r4-r11, r14}; \
+		str r0, [r2, #8]; \
+		ldr r0, [r2, #0]; \
+		str r0, [r1]; \
+		ldr r2, [r1]; \
+		ldr r0, [r2, #8]; \
+		ldmia r0!, {r4-r11, r14}; \
+		msr psp, r0; \
+		bx lr; \
+	");
+
+	/*// save state
 	asm("\
 		cpsid i; \
 		isb; \
 		dsb; \
 		mrs r0, psp; \
-		stmdb r0!, {r4-r11}; \
+		stmdb r0!, {r4-r11, lr}; \
 		mov %0, r0; \
 	" : "=r" (current->sp));
 
@@ -90,12 +106,12 @@ void PendSV_Handler(void)
 	// restore
 	asm("\
 		mov r0, %0; \
-		ldmia r0!, {r4-r11}; \
+		ldmia r0!, {r4-r11, lr}; \
 		msr psp, r0; \
 		isb; \
 		dsb; \
 		cpsie i; \
 		bx lr; \
-	" :: "r" (current->sp));
+	" :: "r" (current->sp));*/
 }