diff --git a/build.bat b/build.bat
index 1d0f9c0..e05bcf3 100644
--- a/build.bat
+++ b/build.bat
@@ -1,9 +1,9 @@
-@echo off
-rmdir /S /Q build
-mkdir build
-
-pushd build
-
-cl ../main.cpp /Zi
-
+@echo off
+rmdir /S /Q build
+mkdir build
+
+pushd build
+
+cl ../build.c /Zi /Fecgame.exe /Od /DDEBUG /MD /DEBUG /std:c11
+
 popd
\ No newline at end of file
diff --git a/build.c b/build.c
new file mode 100644
index 0000000..90aa346
--- /dev/null
+++ b/build.c
@@ -0,0 +1,19 @@
+
+///
+// Build config stuff
+#define VERY_DEBUG 0
+#define RUN_TESTS 0
+
+typedef struct Context_Extra {
+	int monkee;
+} Context_Extra;
+// This needs to be defined before oogabooga if we want extra stuff in context
+#define CONTEXT_EXTRA Context_Extra
+
+#include "oogabooga/oogabooga.c"
+
+
+// Includes for game goes here	
+// ...
+
+#include "main.c"
\ No newline at end of file
diff --git a/build_release.bat b/build_release.bat
new file mode 100644
index 0000000..dc025a8
--- /dev/null
+++ b/build_release.bat
@@ -0,0 +1,13 @@
+@echo off
+rmdir /S /Q build
+mkdir build
+
+pushd build
+
+mkdir release
+pushd release
+
+cl ../../main.c /Zi /Fecgame.exe /Ox /DRELEASE /std:c11
+
+popd
+popd
\ No newline at end of file
diff --git a/main.c b/main.c
new file mode 100644
index 0000000..829061a
--- /dev/null
+++ b/main.c
@@ -0,0 +1,60 @@
+
+int main() {
+	printf("Program started\n");
+	oogabooga_init(1024 * 1024 * 100); // Start with 100mb program memory
+	
+	// alloc calls to context.allocator.proc which by default is set to the
+	// heap allocator in memory.c
+	int *a = cast(int*)alloc(sizeof(int));
+	dealloc(a);
+	
+	// We can do an old school memory dump to save our game with the global variables
+	// (void*) program_memory and (u64) program_memory_size
+	// program_memory will ALWAYS have the same virtual memory base address, so all
+	// pointers will remain valid when read back from disk.
+	
+	// We can push allocator like jai
+	push_allocator(temp);
+	// all calls to alloc() here will be with the temporary allocator
+	pop_allocator(); // But this is C so we have to pop it manually!
+	
+	// or we can just do this for temporary allocation
+	int *b = talloc(sizeof(int));
+	
+	// Call each frame
+	reset_temporary_storage();
+	
+	Context c = context;
+	// ... modify c
+	push_context(c);
+	// ... do stuff in modified context
+	pop_context();
+	
+	// For now, context only has the allocator and CONTEXT_EXTRA which can be defined
+	// to add some custom stuff to context
+	// But it needs to be in build.c before including oogabooga.c.
+	// #define CONTEXT_EXTRA struct { int monkee; }
+	context.extra.monkee = 69;
+	
+
+	// This can be disabled in build.c
+#if RUN_TESTS
+	oogabooga_run_tests();
+#endif
+
+	int hello;
+	hello = 5;
+
+#ifdef DEBUG
+	printf("Hello, balls! (debug)\n");
+#endif
+
+#ifdef RELEASE
+	printf("Hello, balls! (release)\n");
+#endif
+
+	printf("Program exit as expected\n");
+
+  return 0;
+}
+
diff --git a/main.cpp b/main.cpp
deleted file mode 100644
index 6bd983f..0000000
--- a/main.cpp
+++ /dev/null
@@ -1,10 +0,0 @@
-#include <stdio.h>
-
-int main() {
-
-  int hello;
-  hello = 5;
-
-  printf("Hello, balls!\n");
-  return 0;
-}
\ No newline at end of file
diff --git a/oogabooga/base.c b/oogabooga/base.c
new file mode 100644
index 0000000..05773c2
--- /dev/null
+++ b/oogabooga/base.c
@@ -0,0 +1,92 @@
+
+#include <stdint.h>
+
+typedef uint8_t  u8;
+typedef uint16_t u16;
+typedef uint32_t u32;
+typedef uint64_t u64;
+typedef int8_t  s8;
+typedef int16_t s16;
+typedef int32_t s32;
+typedef int64_t s64;
+
+typedef float f32;
+typedef double f64;
+typedef f32 float32;
+typedef f64 float64;
+
+typedef u8 bool;
+#define false 0
+#define true 1
+
+#define thread_local _Thread_local
+
+#ifdef _MSC_VER
+	#define os_debug_break __debugbreak
+#else
+	#error "Only msvc compiler supported at the moment";
+#endif
+	
+#define assert(cond, ...) if (!(cond)) { printf("Assertion failed for condition: " #cond ". Message: " __VA_ARGS__); os_debug_break(); }
+
+#ifndef max
+	#define max(a, b) ((a) > (b) ? (a) : (b))
+	#define min(a, b) ((a) < (b) ? (a) : (b))
+#endif
+
+#define cast(t) (t)
+
+typedef struct Nothing {int nothing;} Nothing;
+
+#ifndef CONTEXT_EXTRA
+	#define CONTEXT_EXTRA Nothing
+#endif
+
+typedef enum Allocator_Message {
+	ALLOCATOR_ALLOCATE,
+	ALLOCATOR_DEALLOCATE,
+} Allocator_Message;
+typedef void*(*Allocator_Proc)(u64, void*, Allocator_Message);
+
+typedef struct Allocator {
+	Allocator_Proc proc;
+	void *data;	
+} Allocator;
+
+typedef struct Context {
+	Allocator allocator;
+	
+	CONTEXT_EXTRA extra;
+} Context;
+
+
+#define CONTEXT_STACK_MAX 512
+thread_local Context context;
+thread_local Context context_stack[CONTEXT_STACK_MAX];
+thread_local u64 num_contexts = 0;
+
+
+void* alloc(u64 size) { return context.allocator.proc(size, NULL, ALLOCATOR_ALLOCATE); }
+void dealloc(void *p) { context.allocator.proc(0, p, ALLOCATOR_DEALLOCATE); }
+
+
+
+void push_context(Context c) {
+	assert(num_contexts < CONTEXT_STACK_MAX, "Context stack overflow");
+	
+	context_stack[num_contexts] = context;
+	context = c;
+	num_contexts += 1;
+}
+void pop_context() {
+	assert(num_contexts > 0, "No contexts to pop!");
+	num_contexts -= 1;
+	context = context_stack[num_contexts];
+}
+
+void push_allocator(Allocator a) {
+	Context c = context;
+	c.allocator = a;
+	push_context(c);
+}
+void pop_allocator() { pop_context(); }
diff --git a/oogabooga/memory.c b/oogabooga/memory.c
new file mode 100644
index 0000000..f4ad2cf
--- /dev/null
+++ b/oogabooga/memory.c
@@ -0,0 +1,433 @@
+
+
+///
+///
+// Basic general heap allocator, free list
+///
+// Technically thread safe but synchronization is horrible.
+// Fragmentation is catastrophic.
+// We could fix it by merging free nodes every now and then
+// BUT: We aren't really supposed to allocate/deallocate directly on the heap too much anyways...
+
+#define DEFAULT_HEAP_BLOCK_SIZE min((os.page_size * 1024ULL * 50ULL), program_memory_size)
+#define HEAP_ALIGNMENT (sizeof(Heap_Free_Node))
+typedef struct Heap_Free_Node Heap_Free_Node;
+typedef struct Heap_Block Heap_Block;
+
+typedef struct Heap_Free_Node {
+	u64 size;
+	Heap_Free_Node *next;
+} Heap_Free_Node;
+
+typedef struct Heap_Block {
+	u64 size;
+	Heap_Free_Node *free_head;
+	void* start;
+	Heap_Block *next;
+	// 32 bytes !!
+} Heap_Block;
+
+typedef struct {
+	u64 size;
+	Heap_Block *block;
+} Heap_Allocation_Metadata;
+
+Heap_Block *heap_head;
+bool heap_initted = false;
+Mutex_Handle heap_mutex; // This is terrible but I don't care for now
+
+
+u64 get_heap_block_size_excluding_metadata(Heap_Block *block) {
+	return block->size - sizeof(Heap_Block);
+}
+u64 get_heap_block_size_including_metadata(Heap_Block *block) {
+	return block->size;
+}
+
+bool is_pointer_in_program_memory(void *p) {
+	return (u8*)p >= (u8*)program_memory && (u8*)p<((u8*)program_memory+program_memory_size);
+}
+
+// Meant for debug
+void santiy_check_free_node_tree(Heap_Block *block) {
+	Heap_Free_Node *node = block->free_head;
+	
+	u64 total_free = 0;
+	while (node != 0) {
+		Heap_Free_Node *other_node = node->next;
+		
+		while (other_node != 0) {
+			assert(other_node != node, "Circular reference in heap free node tree. That's bad.");
+			other_node = other_node->next;
+		}
+		total_free += node->size;
+		assert(total_free <= block->size, "Free nodes are fucky wucky");
+		node = node->next;
+	}
+	
+}
+
+typedef struct {
+	Heap_Free_Node *best_fit;
+	Heap_Free_Node *previous;
+	u64 delta;
+} Heap_Search_Result;
+Heap_Search_Result search_heap_block(Heap_Block *block, u64 size) {
+	
+	if (block->free_head == 0)  return (Heap_Search_Result){0, 0, 0};
+		
+	Heap_Free_Node *node = block->free_head;
+	Heap_Free_Node *previous = 0;
+	
+	Heap_Free_Node *best_fit = 0;
+	Heap_Free_Node *before_best_fit = 0;
+	u64 best_fit_delta = 0;
+	
+	while (node != 0) {
+		
+		if (node->size == size) {
+			Heap_Search_Result result;
+			result.best_fit = node;
+			result.previous = previous;
+			result.delta = 0;
+			assert(result.previous != result.best_fit);
+			return result;
+		}
+		
+		if (node->size >= size) {
+			u64 delta = node->size - size;
+			
+			if (delta < best_fit_delta || !best_fit) {
+				before_best_fit = previous;
+				best_fit = node;
+				best_fit_delta = delta;
+			}
+		}
+		
+		if (node->next) previous = node;
+		node = node->next;
+	}
+	
+	if (!best_fit)  return (Heap_Search_Result){0, 0, 0};
+
+	Heap_Search_Result result;
+	result.best_fit = best_fit;
+	result.previous = before_best_fit;
+	result.delta = best_fit_delta;
+	assert(result.previous != result.best_fit);
+	return result;
+}
+
+Heap_Block *make_heap_block(Heap_Block *parent, u64 size) {
+
+	size = (size) & ~(HEAP_ALIGNMENT-1);	
+
+	Heap_Block *block;
+	if (parent) {
+		block = (Heap_Block*)(((u8*)parent)+get_heap_block_size_including_metadata(parent));
+		parent->next = block;
+	} else {
+		block = (Heap_Block*)program_memory;
+	}
+	
+	
+	
+	// #Speed #Cleanup
+	if (((u8*)block)+size >= ((u8*)program_memory)+program_memory_size) {
+		u64 minimum_size = ((u8*)block+size) - (u8*)program_memory + 1;
+		u64 new_program_size = (cast(u64)(minimum_size * 1.5));
+		assert(new_program_size >= minimum_size, "Bröd");
+		printf("Growing program memory to %llu bytes\n", new_program_size);
+		const u64 ATTEMPTS = 1000;
+		for (u64 i = 0; i <= ATTEMPTS; i++) {
+			if (program_memory_size >= new_program_size) break; // Another thread might have resized already, causing it to fail here.
+			assert(i < ATTEMPTS, "OS is not letting us allocate more memory. Maybe we are out of memory?");
+			if (os_grow_program_memory(new_program_size))
+				break;
+		}
+	}
+	
+	block->start = ((u8*)block)+sizeof(Heap_Block);
+	block->size = size;
+	block->next = 0;
+	block->free_head = (Heap_Free_Node*)block->start;
+	block->free_head->size = get_heap_block_size_excluding_metadata(block);
+	block->free_head->next = 0;
+	
+	return block;
+}
+
+void heap_init() {
+	heap_head = make_heap_block(0, DEFAULT_HEAP_BLOCK_SIZE);
+	heap_mutex = os_make_mutex();
+	heap_initted = true;
+}
+
+void *heap_alloc(u64 size) {
+	// #Sync #Speed oof
+	os_lock_mutex(heap_mutex);
+
+	if (!heap_initted) heap_init();
+	
+	size += sizeof(Heap_Allocation_Metadata);
+	
+	size = (size+HEAP_ALIGNMENT) & ~(HEAP_ALIGNMENT-1);
+	
+	assert(size < DEFAULT_HEAP_BLOCK_SIZE, "Past Charlie has been lazy and did not handle large allocations like this. I apologize on behalf of past Charlie. A quick fix could be to increase the heap block size for now.");
+	
+	Heap_Block *block = heap_head;
+	Heap_Block *last_block = 0;
+	Heap_Free_Node *best_fit = 0;
+	Heap_Block *best_fit_block = 0;
+	Heap_Free_Node *previous = 0;
+	u64 best_fit_delta = 0;
+	// #Speed
+	// Maybe instead of going through EVERY free node to find best fit we do a good-enough fit
+	while (block != 0) {
+		Heap_Search_Result result = search_heap_block(block, size);
+		Heap_Free_Node *node = result.best_fit;
+		if (node) {
+			if (node->size < size) continue;
+			if (node->size == size) {
+				best_fit = node;
+				best_fit_block = block;
+				previous = result.previous;
+				best_fit_delta = 0;
+				break;
+			}
+			
+			u64 delta = node->size-size;
+			if (delta < best_fit_delta || !best_fit) {
+				best_fit = node;
+				best_fit_block = block;
+				previous = result.previous;
+				best_fit_delta = delta;
+			}
+		}
+		
+		last_block = block;
+		block = block->next;
+	}
+	
+	if (!best_fit) {
+		block = make_heap_block(last_block, DEFAULT_HEAP_BLOCK_SIZE);
+		previous = 0;
+		best_fit = block->free_head;
+		best_fit_block = block;
+	}
+		
+	
+	// Ideally this should not be possible.
+	// If we run out of program_memory, we should just grow it and if that fails
+	// we crash because out of memory.
+	assert(best_fit != 0, "Internal heap allocation failed");
+	
+	Heap_Free_Node *new_free_node = 0;
+	if (size != best_fit->size) {
+		u64 remainder = best_fit->size - size;
+		new_free_node = (Heap_Free_Node*)(((u8*)best_fit)+size);
+		new_free_node->size = remainder;
+		new_free_node->next = best_fit->next;
+	}
+	
+	if (previous && new_free_node) {
+		assert(previous->next == best_fit, "Bro what");
+		previous->next = new_free_node;
+	} else if (previous) {
+		assert(previous->next == best_fit, "Bro what");
+		previous->next = best_fit->next;
+	}
+	
+	if (best_fit_block->free_head == best_fit) {
+		// If we allocated the first free node then replace with new free node or just
+		// remove it if perfect fit.
+		if (new_free_node) {
+			new_free_node->next = best_fit_block->free_head->next;
+			best_fit_block->free_head = new_free_node;
+		} else best_fit_block->free_head = best_fit_block->free_head->next;
+	}
+	
+	Heap_Allocation_Metadata *meta = (Heap_Allocation_Metadata*)best_fit;
+	meta->size = size;
+	meta->block = best_fit_block;
+
+#if VERY_DEBUG
+	santiy_check_free_node_tree(meta->block);
+#endif
+	
+	// #Sync #Speed oof
+	os_unlock_mutex(heap_mutex);
+	
+	return ((u8*)meta)+sizeof(Heap_Allocation_Metadata);
+}
+void heap_dealloc(void *p) {
+	// #Sync #Speed oof
+	os_lock_mutex(heap_mutex);
+	
+	if (!heap_initted) heap_init();
+	
+	assert(is_pointer_in_program_memory(p), "Garbage pointer; out of program memory bounds!"); 
+	p = (u8*)p-sizeof(Heap_Allocation_Metadata);
+	Heap_Allocation_Metadata *meta = (Heap_Allocation_Metadata*)(p);
+	
+	// If > 256GB then prolly not legit lol
+	assert(meta->size < 1024ULL*1024ULL*1024ULL*256ULL, "Garbage pointer passed to heap_dealloc !!!");	
+	assert(is_pointer_in_program_memory(meta->block), "Garbage pointer passed to heap_dealloc !!!"); 
+	
+	// Yoink meta data before we start overwriting it
+	Heap_Block *block = meta->block;
+	u64 size = meta->size;
+	
+	Heap_Free_Node *new_node = cast(Heap_Free_Node*)p;
+	new_node->size = size;
+	
+	if (new_node < block->free_head) {
+		if ((u8*)new_node+size == (u8*)block->free_head) {
+			new_node->size = size + block->free_head->size;
+			new_node->next = block->free_head->next;
+			block->free_head = new_node;
+		} else {
+			new_node->next = block->free_head;
+			block->free_head = new_node;
+		}
+	} else {
+		Heap_Free_Node *node = block->free_head;
+	
+		
+		while (true) {
+		
+			assert(node != 0, "We didn't find where the free node should be! uh oh");
+			
+			if (new_node > node) {
+				u8* node_tail = (u8*)node + node->size;
+				if (cast(u8*)new_node == node_tail) {
+					node->size += new_node->size;
+					break;
+				} else {
+					new_node->next = node->next;
+					node->next = new_node;
+					
+					u8* new_node_tail = (u8*)new_node + new_node->size;
+					if (new_node->next && (u8*)new_node->next == new_node_tail) {
+						new_node->size += new_node->next->size;
+						new_node->next = new_node->next->next;
+					}
+					break;
+				}
+			}
+			
+			node = node->next;
+		}
+	}
+	
+#if VERY_DEBUG
+	santiy_check_free_node_tree(block);
+#endif
+
+	// #Sync #Speed oof
+	os_unlock_mutex(heap_mutex);
+}
+void log_heap() {
+	os_lock_mutex(heap_mutex);
+	printf("\nHEAP:\n");
+	
+	Heap_Block *block = heap_head;
+	
+	while (block != 0) {
+		
+		printf("\tBLOCK @ 0x%I64x, %llu bytes\n", (u64)block, block->size);
+		
+		Heap_Free_Node *node = block->free_head;
+
+		u64 total_free = 0;
+		
+		while (node != 0) {
+		
+			printf("\t\tFREE NODE @ 0x%I64x, %llu bytes\n", (u64)node, node->size);
+			
+			total_free += node->size;
+		
+			node = node->next;
+		}
+		
+		printf("\t TOTAL FREE: %llu\n\n", total_free);
+		
+		block = block->next;
+	}
+	os_unlock_mutex(heap_mutex);
+}
+
+
+///
+///
+// Temporary storage
+///
+
+#ifndef TEMPORARY_STORAGE_SIZE
+	#define TEMPORARY_STORAGE_SIZE (1024ULL*1024ULL*16ULL) // 16mb
+#endif
+
+void* talloc(u64);
+void* temp_allocator_proc(u64 size, void *p, Allocator_Message message);
+
+thread_local void * temporary_storage = 0;
+thread_local bool   temporary_storage_initted = false;
+thread_local void * temporary_storage_pointer = 0;
+thread_local bool   has_warned_temporary_storage_overflow = false;
+thread_local Allocator temp;
+
+
+void* temp_allocator_proc(u64 size, void *p, Allocator_Message message) {
+	switch (message) {
+		case ALLOCATOR_ALLOCATE: {
+			return talloc(size);
+			break;
+		}
+		case ALLOCATOR_DEALLOCATE: {
+			return 0;
+		}
+	}
+	return 0;
+}
+
+void temporary_storage_init() {
+	if (temporary_storage_initted) return;
+	
+	temporary_storage = heap_alloc(TEMPORARY_STORAGE_SIZE);
+	assert(temporary_storage, "Failed allocating temporary storage");
+	temporary_storage_pointer = temporary_storage;
+
+	temp.proc = temp_allocator_proc;
+	temp.data = 0;
+	
+	temporary_storage_initted = true;
+}
+
+void* talloc(u64 size) {
+	if (!temporary_storage_initted) temporary_storage_init();
+	
+	assert(size < TEMPORARY_STORAGE_SIZE, "Bruddah this is too large for temp allocator");
+	
+	void* p = temporary_storage_pointer;
+	
+	temporary_storage_pointer = (u8*)temporary_storage_pointer + size;
+	
+	if ((u8*)temporary_storage_pointer >= (u8*)temporary_storage+TEMPORARY_STORAGE_SIZE) {
+		if (!has_warned_temporary_storage_overflow) {
+			printf("WARNING: temporary storage was overflown, we wrap around at the start.\n");
+		}
+		temporary_storage_pointer = temporary_storage;
+		return talloc(size);;
+	}
+	
+	return p;
+}
+
+void reset_temporary_storage() {
+	if (!temporary_storage_initted) temporary_storage_init();
+	
+	temporary_storage_pointer = temporary_storage;
+	
+	has_warned_temporary_storage_overflow = true;
+}
+
diff --git a/oogabooga/oogabooga.c b/oogabooga/oogabooga.c
new file mode 100644
index 0000000..6d4d5b2
--- /dev/null
+++ b/oogabooga/oogabooga.c
@@ -0,0 +1,29 @@
+
+#include <stdio.h> // #Cleanup just using this for printf
+#include <stdlib.h>
+
+#include "base.c"
+
+#include "os_interface.c"
+
+#include "memory.c"
+
+
+
+
+#ifdef OS_WINDOWS
+	#include "os_impl_windows.c"
+#elif defined (OS_LINUX)
+
+#elif defined (OS_MAC)
+
+#endif
+
+#include "tests.c"
+
+
+void oogabooga_init(u64 program_memory_size) {
+	os_init(program_memory_size);
+	heap_init();
+	temporary_storage_init();
+}
\ No newline at end of file
diff --git a/oogabooga/os_impl_windows.c b/oogabooga/os_impl_windows.c
new file mode 100644
index 0000000..62b44fb
--- /dev/null
+++ b/oogabooga/os_impl_windows.c
@@ -0,0 +1,174 @@
+
+
+#define VIRTUAL_MEMORY_BASE ((void*)0x0000010000000000ULL)
+
+void* heap_alloc(u64);
+void heap_dealloc(void*);
+
+void* heap_allocator_proc(u64 size, void *p, Allocator_Message message) {
+	switch (message) {
+		case ALLOCATOR_ALLOCATE: {
+			return heap_alloc(size);
+			break;
+		}
+		case ALLOCATOR_DEALLOCATE: {
+			heap_dealloc(p);
+			return 0;
+		}
+	}
+	return 0;
+}
+
+void os_init(u64 program_memory_size) {
+	
+	SYSTEM_INFO si;
+    GetSystemInfo(&si);
+	os.granularity = cast(u64)si.dwAllocationGranularity;
+	os.page_size = cast(u64)si.dwPageSize;
+
+
+	program_memory_mutex = os_make_mutex();
+	os_grow_program_memory(program_memory_size);
+	
+	Allocator heap_allocator;
+	
+	heap_allocator.proc = heap_allocator_proc;
+	heap_allocator.data = 0;
+	
+	context.allocator = heap_allocator;
+	
+}
+
+bool os_grow_program_memory(u64 new_size) {
+	os_lock_mutex(program_memory_mutex); // #Sync
+	if (program_memory_size >= new_size) {
+		os_unlock_mutex(program_memory_mutex); // #Sync
+		return true;
+	}
+
+	
+	
+	bool is_first_time = program_memory == 0;
+	
+	if (is_first_time) {
+		u64 aligned_size = (new_size+os.granularity) & ~(os.granularity);
+		void* aligned_base = (void*)(((u64)VIRTUAL_MEMORY_BASE+os.granularity) & ~(os.granularity-1));
+
+		u64 m = aligned_size & os.granularity;
+		assert(m == 0);
+		
+		program_memory = VirtualAlloc(aligned_base, aligned_size, MEM_RESERVE | MEM_COMMIT, PAGE_READWRITE);
+		if (program_memory == 0) { 
+			os_unlock_mutex(program_memory_mutex); // #Sync
+			return false;
+		}
+		program_memory_size = aligned_size;
+	} else {
+		void* tail = (u8*)program_memory + program_memory_size;
+		u64 m = ((u64)program_memory_size % os.granularity);
+		assert(m == 0, "program_memory_size is not aligned to granularity!");
+		m = ((u64)tail % os.granularity);
+		assert(m == 0, "Tail is not aligned to granularity!");
+		u64 amount_to_allocate = new_size-program_memory_size;
+		amount_to_allocate = ((amount_to_allocate+os.granularity)&~(os.granularity-1));
+		m = ((u64)amount_to_allocate % os.granularity);
+		assert(m == 0, "amount_to_allocate is not aligned to granularity!");
+		// Just keep allocating at the tail of the current chunk
+		void* result = VirtualAlloc(tail, amount_to_allocate, MEM_RESERVE | MEM_COMMIT, PAGE_READWRITE);
+		if (result == 0) { 
+			os_unlock_mutex(program_memory_mutex); // #Sync
+			return false;
+		}
+		assert(tail == result, "It seems tail is not aligned properly. o nein");
+		
+		program_memory_size += amount_to_allocate;
+
+		m = ((u64)program_memory_size % os.granularity);
+		assert(m == 0, "program_memory_size is not aligned to granularity!");
+	}
+
+	
+	
+	os_unlock_mutex(program_memory_mutex); // #Sync
+	return true;
+}
+
+Mutex_Handle os_make_mutex() {
+	return CreateMutex(0, FALSE, 0);
+}
+void os_destroy_mutex(Mutex_Handle m) {
+	CloseHandle(m);
+}
+void os_lock_mutex(Mutex_Handle m) {
+	DWORD wait_result = WaitForSingleObject(m, INFINITE);
+	
+	switch (wait_result) {
+        case WAIT_OBJECT_0:
+            break;
+
+        case WAIT_ABANDONED:
+            break;
+
+        default:
+        	assert(false, "Unexpected mutex lock result");
+            break;
+    }
+}
+void os_unlock_mutex(Mutex_Handle m) {
+	BOOL result = ReleaseMutex(m);
+	assert(result, "Unlock mutex failed");
+}
+
+DWORD WINAPI win32_thread_invoker(LPVOID param) {
+	Thread *t = (Thread*)param;
+	temporary_storage_init();
+	context = t->initial_context;
+	t->proc(t);
+	return 0;
+}
+
+Thread* os_make_thread(Thread_Proc proc) {
+	Thread *t = (Thread*)alloc(sizeof(Thread));
+	t->id = 0; // This is set when we start it
+	t->proc = proc;
+	t->initial_context = context;
+	
+	return t;
+}
+void os_start_thread(Thread *t) {
+	t->os_handle = CreateThread(
+        0,
+        0,
+        win32_thread_invoker,
+        t,
+        0,
+        (DWORD*)&t->id
+    );
+    
+    assert(t->os_handle, "Failed creating thread");
+}
+void os_join_thread(Thread *t) {
+	WaitForSingleObject(t->os_handle, INFINITE);
+	CloseHandle(t->os_handle);
+}
+
+void os_sleep(u32 ms) {
+    Sleep(ms);
+}
+
+void os_yield_thread() {
+    SwitchToThread();
+}
+
+#include <intrin.h>
+u64 os_get_current_cycle_count() {
+	return __rdtsc();
+}
+
+float64 os_get_current_time_in_seconds() {
+    LARGE_INTEGER frequency, counter;
+    if (!QueryPerformanceFrequency(&frequency) || !QueryPerformanceCounter(&counter)) {
+        return -1.0;
+    }
+    return (double)counter.QuadPart / (double)frequency.QuadPart;
+}
\ No newline at end of file
diff --git a/oogabooga/os_interface.c b/oogabooga/os_interface.c
new file mode 100644
index 0000000..4d69c87
--- /dev/null
+++ b/oogabooga/os_interface.c
@@ -0,0 +1,83 @@
+
+
+#ifdef _WIN32
+	#include <Windows.h>
+	#define OS_WINDOWS
+	
+	typedef HANDLE Mutex_Handle;
+	typedef HANDLE Thread_Handle;
+	
+#elif defined(__linux__)
+	// Include whatever #Incomplete #Portability
+	#define OS_LINUX
+	
+	typedef SOMETHING Mutex_Handle;
+	typedef SOMETHING Thread_Handle;
+	
+#elif defined(__APPLE__) && defined(__MACH__)
+	// Include whatever #Incomplete #Portability
+	#define OS_MAC
+	
+	typedef SOMETHING Mutex_Handle;
+	typedef SOMETHING Thread_Handle;
+#else
+	#error "Current OS not supported!";
+#endif
+
+typedef struct Os_Info {
+	u64 page_size;
+	u64 granularity;
+} Os_Info;
+Os_Info os;
+
+
+
+
+void* program_memory = 0;
+u64 program_memory_size = 0;
+Mutex_Handle program_memory_mutex = 0;
+
+bool os_grow_program_memory(size_t new_size);
+
+///
+///
+// Threading
+///
+
+typedef struct Thread Thread;
+
+typedef void(*Thread_Proc)(Thread*);
+
+typedef struct Thread {
+	u64 id;
+	Context initial_context;
+	void* data;
+	Thread_Proc proc;
+	Thread_Handle os_handle;
+	
+} Thread;
+
+///
+// Thread primitive
+Thread* os_make_thread();
+void os_start_thread(Thread* t);
+void os_join_thread(Thread* t);
+
+void os_sleep(u32 ms);
+void os_yield_thread();
+
+///
+// Mutex primitive
+Mutex_Handle os_make_mutex();
+void os_destroy_mutex(Mutex_Handle m);
+void os_lock_mutex(Mutex_Handle m);
+void os_unlock_mutex(Mutex_Handle m);
+
+
+///
+///
+// Time
+///
+
+u64 os_get_current_cycle_count();
+float64 os_get_current_time_in_seconds();
\ No newline at end of file
diff --git a/oogabooga/tests.c b/oogabooga/tests.c
new file mode 100644
index 0000000..1ba3594
--- /dev/null
+++ b/oogabooga/tests.c
@@ -0,0 +1,243 @@
+
+
+void test_allocator(bool do_log_heap) {
+	// Basic allocation and free
+    int* a = (int*)alloc(sizeof(int));
+    int* b = (int*)alloc(sizeof(int));
+    int* c = (int*)alloc(sizeof(int));
+    
+    *a = 69;
+    *b = 420;
+    *c = 1337;
+    
+    assert(*a == 69, "Test failed: Memory corrupted");
+    assert(*b == 420, "Test failed: Memory corrupted");
+    assert(*c == 1337, "Test failed: Memory corrupted");
+
+    // Test growing memory
+    os_grow_program_memory(1024 * 1024 * 1000);
+    
+    assert(*a == 69, "Test failed: Memory corrupted");
+    assert(*b == 420, "Test failed: Memory corrupted");
+    assert(*c == 1337, "Test failed: Memory corrupted");
+
+    // Allocate and free large block
+    void* large_block = alloc(1024 * 1024 * 100);
+    dealloc(large_block);
+
+    // Allocate multiple small blocks
+    void* blocks[100];
+    for (int i = 0; i < 100; ++i) {
+        blocks[i] = alloc(128);
+        assert(blocks[i] != NULL, "Failed to allocate small block");
+    }
+    
+    for (int i = 0; i < 100; ++i) {
+        dealloc(blocks[i]);
+    }
+
+    // Stress test with various sizes
+    for (int i = 1; i <= 1000; ++i) {
+        void* p = alloc(i * 64);
+        assert(p != NULL, "Failed to allocate varying size block");
+        dealloc(p);
+    }
+
+    // Free in reverse order
+    for (int i = 0; i < 100; ++i) {
+        blocks[i] = alloc(128);
+        assert(blocks[i] != NULL, "Failed to allocate small block");
+    }
+
+    for (int i = 99; i >= 0; --i) {
+        dealloc(blocks[i]);
+    }
+
+    // Test memory integrity with various allocation patterns
+    int* nums[10];
+    for (int i = 0; i < 10; ++i) {
+        nums[i] = (int*)alloc(sizeof(int) * 10);
+        for (int j = 0; j < 10; ++j) {
+            nums[i][j] = i * 10 + j;
+        }
+    }
+
+    for (int i = 0; i < 10; ++i) {
+        for (int j = 0; j < 10; ++j) {
+            assert(nums[i][j] == i * 10 + j, "Memory corruption detected");
+        }
+         dealloc(nums[i]);
+    }
+    
+    
+    push_allocator(temp);
+    
+    int* foo = (int*)alloc(72);
+    *foo = 1337;
+    void* bar = alloc(69);
+    void* baz = alloc(420);
+    
+    assert(*foo == 1337, "Temp memory corruptada");
+    
+    int* old_foo = foo;
+    
+    reset_temporary_storage();
+    
+    foo = (int*)alloc(72);
+    
+    assert(old_foo == foo, "Temp allocator goof");
+    
+    pop_allocator();
+    
+    // Repeated Allocation and Free
+    for (int i = 0; i < 10000; ++i) {
+        void* temp = alloc(128);
+        assert(temp != NULL && "Repeated allocation failed");
+        dealloc(temp);
+    }
+
+    // Mixed Size Allocations
+    void* mixed_blocks[200];
+    for (int i = 0; i < 200; ++i) {
+        if (i % 2 == 0) {
+            mixed_blocks[i] = alloc(128);
+        } else {
+            mixed_blocks[i] = alloc(1024 * 1024); // 1MB blocks
+        }
+        assert(mixed_blocks[i] != NULL && "Mixed size allocation failed");
+    }
+
+    for (int i = 0; i < 200; ++i) {
+        if (i % 2 == 0) {
+            dealloc(mixed_blocks[i]);
+        }
+    }
+
+    for (int i = 0; i < 200; ++i) {
+        if (i % 2 != 0) {
+            dealloc(mixed_blocks[i]);
+        }
+    }
+
+    // Fragmentation Stress Test
+    for (int i = 0; i < 50; ++i) {
+        blocks[i] = alloc(256);
+        assert(blocks[i] != NULL && "Failed to allocate small block for fragmentation test");
+    }
+
+    for (int i = 0; i < 50; i += 2) {
+        dealloc(blocks[i]);
+    }
+
+    for (int i = 50; i < 100; ++i) {
+        blocks[i] = alloc(128);
+        assert(blocks[i] != NULL && "Failed to allocate small block in fragmented heap");
+    }
+
+    for (int i = 50; i < 100; ++i) {
+        dealloc(blocks[i]);
+    }
+
+    for (int i = 1; i < 50; i += 2) {
+        dealloc(blocks[i]);
+    }
+    
+    if (do_log_heap) log_heap();
+}
+
+void test_thread_proc1(Thread* t) {
+	os_sleep(5);
+	printf("Hello from thread %llu\n", t->id);
+	os_sleep(5);
+	printf("Hello from thread %llu\n", t->id);
+	os_sleep(5);
+	printf("Hello from thread %llu\n", t->id);
+	os_sleep(5);
+	printf("Hello from thread %llu\n", t->id);
+	os_sleep(5);
+	printf("Hello from thread %llu\n", t->id);
+}
+Mutex_Handle test_mutex;
+void test_thread_proc2(Thread* t) {
+	os_lock_mutex(test_mutex);
+	printf("Thread %llu part 1\n", t->id);
+	os_sleep(1);
+	printf("Thread %llu part 2\n", t->id);
+	os_sleep(1);
+	printf("Thread %llu part 3\n", t->id);
+	os_unlock_mutex(test_mutex);
+}
+void test_threads() {
+	Thread* t = os_make_thread(test_thread_proc1);
+	os_start_thread(t);
+	os_sleep(10);
+	printf("This should be printed in middle of thread execution\n");
+	os_join_thread(t);
+	printf("Thread is joined\n");
+	
+	Mutex_Handle m = os_make_mutex();
+	os_lock_mutex(m);
+	os_unlock_mutex(m);
+	
+	
+	test_mutex = os_make_mutex();
+	Thread *threads[100];
+	for (int i = 0; i < 100; i++) {
+		threads[i] = os_make_thread(test_thread_proc2);
+		os_start_thread(threads[i]);
+	}
+	for (int i = 0; i < 100; i++) {
+		os_join_thread(threads[i]);
+	}
+}
+
+void test_allocator_threaded(Thread *t) {
+    for (int i = 0; i < 1000; ++i) {
+        void* temp = alloc(128);
+        assert(temp != NULL && "Repeated allocation failed");
+        dealloc(temp);
+    }
+
+    void* mixed_blocks[40];
+    for (int i = 0; i < 40; ++i) {
+        if (i % 2 == 0) {
+            mixed_blocks[i] = alloc(128);
+        } else {
+            mixed_blocks[i] = alloc(1024 * 1024); // 1MB blocks
+        }
+        assert(mixed_blocks[i] != NULL && "Mixed size allocation failed");
+    }
+
+    for (int i = 0; i < 40; ++i) {
+        if (i % 2 == 0) {
+            dealloc(mixed_blocks[i]);
+        }
+    }
+
+    for (int i = 0; i < 40; ++i) {
+        if (i % 2 != 0) {
+            dealloc(mixed_blocks[i]);
+        }
+    }
+}
+
+void oogabooga_run_tests() {
+	printf("Testing allocator...\n");
+	test_allocator(true);
+	printf("OK!\n");
+	
+	printf("Testing threads...\n");
+	test_threads();
+	printf("OK!\n");
+
+	printf("Thread bombing allocator...\n");
+	Thread* threads[100];
+	for (int i = 0; i < 100; i++) {
+		threads[i] = os_make_thread(test_allocator_threaded);
+		os_start_thread(threads[i]);
+	}
+	for (int i = 0; i < 100; i++) {
+		os_join_thread(threads[i]);
+	}
+	printf("OK!\n");
+}
\ No newline at end of file