From b6b8703ecd693f177864c3bc9e216e4a252ab1be Mon Sep 17 00:00:00 2001
From: Robert Winkler <rob121618@gmail.com>
Date: Sat, 1 Jun 2024 12:38:23 -0700
Subject: [PATCH] Use pre-allocated array, remove cvector_float

---
 portablegl.h         | 377 ++-----------------------------------------
 portablegl_unsafe.h  | 375 ++----------------------------------------
 src/generate_gl_h.py |   5 +-
 src/gl_impl.c        |  14 +-
 src/gl_impl_unsafe.c |  12 +-
 src/gl_internal.c    |   6 +-
 src/gl_types.h       |   2 +-
 src/pgl_ext.c        |   5 +-
 8 files changed, 53 insertions(+), 743 deletions(-)

diff --git a/portablegl.h b/portablegl.h
index 504b5e2..4c90ace 100644
--- a/portablegl.h
+++ b/portablegl.h
@@ -1979,58 +1979,6 @@ static inline std::ostream& operator<<(std::ostream& stream, const vec4& a)
 /* CRSW_MATH_H */
 #endif
 
-#ifndef CVEC_SIZE_T
-#include <stdlib.h>
-#define CVEC_SIZE_T size_t
-#endif
-
-#ifndef CVEC_SZ
-#define CVEC_SZ
-typedef CVEC_SIZE_T cvec_sz;
-#endif
-
-
-/** Data structure for float vector. */
-typedef struct cvector_float
-{
-	float* a;           /**< Array. */
-	cvec_sz size;       /**< Current size (amount you use when manipulating array directly). */
-	cvec_sz capacity;   /**< Allocated size of array; always >= size. */
-} cvector_float;
-
-
-
-extern cvec_sz CVEC_float_SZ;
-
-int cvec_float(cvector_float* vec, cvec_sz size, cvec_sz capacity);
-int cvec_init_float(cvector_float* vec, float* vals, cvec_sz num);
-
-cvector_float* cvec_float_heap(cvec_sz size, cvec_sz capacity);
-cvector_float* cvec_init_float_heap(float* vals, cvec_sz num);
-int cvec_copyc_float(void* dest, void* src);
-int cvec_copy_float(cvector_float* dest, cvector_float* src);
-
-int cvec_push_float(cvector_float* vec, float a);
-float cvec_pop_float(cvector_float* vec);
-
-int cvec_extend_float(cvector_float* vec, cvec_sz num);
-int cvec_insert_float(cvector_float* vec, cvec_sz i, float a);
-int cvec_insert_array_float(cvector_float* vec, cvec_sz i, float* a, cvec_sz num);
-float cvec_replace_float(cvector_float* vec, cvec_sz i, float a);
-void cvec_erase_float(cvector_float* vec, cvec_sz start, cvec_sz end);
-int cvec_reserve_float(cvector_float* vec, cvec_sz size);
-#define cvec_shrink_to_fit_float(vec) cvec_set_cap_float((vec), (vec)->size)
-int cvec_set_cap_float(cvector_float* vec, cvec_sz size);
-void cvec_set_val_sz_float(cvector_float* vec, float val);
-void cvec_set_val_cap_float(cvector_float* vec, float val);
-
-float* cvec_back_float(cvector_float* vec);
-
-void cvec_clear_float(cvector_float* vec);
-void cvec_free_float_heap(void* vec);
-void cvec_free_float(void* vec);
-
-
 
 
 #include <stdint.h>
@@ -2649,7 +2597,7 @@ typedef struct Vertex_Shader_output
 	// but still easily add back functions as needed...
 	//
 	// or like comment in init_glContext says just allocate to the max size and be done
-	cvector_float output_buf;
+	float* output_buf;
 } Vertex_Shader_output;
 
 
@@ -5525,304 +5473,6 @@ void cvec_free_glVertex(void* vec)
 }
 
 
-#if defined(CVEC_MALLOC) && defined(CVEC_FREE) && defined(CVEC_REALLOC)
-/* ok */
-#elif !defined(CVEC_MALLOC) && !defined(CVEC_FREE) && !defined(CVEC_REALLOC)
-/* ok */
-#else
-#error "Must define all or none of CVEC_MALLOC, CVEC_FREE, and CVEC_REALLOC."
-#endif
-
-#ifndef CVEC_MALLOC
-#include <stdlib.h>
-#define CVEC_MALLOC(sz)      malloc(sz)
-#define CVEC_REALLOC(p, sz)  realloc(p, sz)
-#define CVEC_FREE(p)         free(p)
-#endif
-
-#ifndef CVEC_MEMMOVE
-#include <string.h>
-#define CVEC_MEMMOVE(dst, src, sz)  memmove(dst, src, sz)
-#endif
-
-#ifndef CVEC_ASSERT
-#include <assert.h>
-#define CVEC_ASSERT(x)       assert(x)
-#endif
-
-cvec_sz CVEC_float_SZ = 50;
-
-#define CVEC_float_ALLOCATOR(x) ((x+1) * 2)
-
-cvector_float* cvec_float_heap(cvec_sz size, cvec_sz capacity)
-{
-	cvector_float* vec;
-	if (!(vec = (cvector_float*)CVEC_MALLOC(sizeof(cvector_float)))) {
-		CVEC_ASSERT(vec != NULL);
-		return NULL;
-	}
-
-	vec->size = size;
-	vec->capacity = (capacity > vec->size || (vec->size && capacity == vec->size)) ? capacity : vec->size + CVEC_float_SZ;
-
-	if (!(vec->a = (float*)CVEC_MALLOC(vec->capacity*sizeof(float)))) {
-		CVEC_ASSERT(vec->a != NULL);
-		CVEC_FREE(vec);
-		return NULL;
-	}
-
-	return vec;
-}
-
-cvector_float* cvec_init_float_heap(float* vals, cvec_sz num)
-{
-	cvector_float* vec;
-	
-	if (!(vec = (cvector_float*)CVEC_MALLOC(sizeof(cvector_float)))) {
-		CVEC_ASSERT(vec != NULL);
-		return NULL;
-	}
-
-	vec->capacity = num + CVEC_float_SZ;
-	vec->size = num;
-	if (!(vec->a = (float*)CVEC_MALLOC(vec->capacity*sizeof(float)))) {
-		CVEC_ASSERT(vec->a != NULL);
-		CVEC_FREE(vec);
-		return NULL;
-	}
-
-	CVEC_MEMMOVE(vec->a, vals, sizeof(float)*num);
-
-	return vec;
-}
-
-int cvec_float(cvector_float* vec, cvec_sz size, cvec_sz capacity)
-{
-	vec->size = size;
-	vec->capacity = (capacity > vec->size || (vec->size && capacity == vec->size)) ? capacity : vec->size + CVEC_float_SZ;
-
-	if (!(vec->a = (float*)CVEC_MALLOC(vec->capacity*sizeof(float)))) {
-		CVEC_ASSERT(vec->a != NULL);
-		vec->size = vec->capacity = 0;
-		return 0;
-	}
-
-	return 1;
-}
-
-int cvec_init_float(cvector_float* vec, float* vals, cvec_sz num)
-{
-	vec->capacity = num + CVEC_float_SZ;
-	vec->size = num;
-	if (!(vec->a = (float*)CVEC_MALLOC(vec->capacity*sizeof(float)))) {
-		CVEC_ASSERT(vec->a != NULL);
-		vec->size = vec->capacity = 0;
-		return 0;
-	}
-
-	CVEC_MEMMOVE(vec->a, vals, sizeof(float)*num);
-
-	return 1;
-}
-
-int cvec_copyc_float(void* dest, void* src)
-{
-	cvector_float* vec1 = (cvector_float*)dest;
-	cvector_float* vec2 = (cvector_float*)src;
-
-	vec1->a = NULL;
-	vec1->size = 0;
-	vec1->capacity = 0;
-
-	return cvec_copy_float(vec1, vec2);
-}
-
-int cvec_copy_float(cvector_float* dest, cvector_float* src)
-{
-	float* tmp = NULL;
-	if (!(tmp = (float*)CVEC_REALLOC(dest->a, src->capacity*sizeof(float)))) {
-		CVEC_ASSERT(tmp != NULL);
-		return 0;
-	}
-	dest->a = tmp;
-
-	CVEC_MEMMOVE(dest->a, src->a, src->size*sizeof(float));
-	dest->size = src->size;
-	dest->capacity = src->capacity;
-	return 1;
-}
-
-
-int cvec_push_float(cvector_float* vec, float a)
-{
-	float* tmp;
-	cvec_sz tmp_sz;
-	if (vec->capacity > vec->size) {
-		vec->a[vec->size++] = a;
-	} else {
-		tmp_sz = CVEC_float_ALLOCATOR(vec->capacity);
-		if (!(tmp = (float*)CVEC_REALLOC(vec->a, sizeof(float)*tmp_sz))) {
-			CVEC_ASSERT(tmp != NULL);
-			return 0;
-		}
-		vec->a = tmp;
-		vec->a[vec->size++] = a;
-		vec->capacity = tmp_sz;
-	}
-	return 1;
-}
-
-float cvec_pop_float(cvector_float* vec)
-{
-	return vec->a[--vec->size];
-}
-
-float* cvec_back_float(cvector_float* vec)
-{
-	return &vec->a[vec->size-1];
-}
-
-int cvec_extend_float(cvector_float* vec, cvec_sz num)
-{
-	float* tmp;
-	cvec_sz tmp_sz;
-	if (vec->capacity < vec->size + num) {
-		tmp_sz = vec->capacity + num + CVEC_float_SZ;
-		if (!(tmp = (float*)CVEC_REALLOC(vec->a, sizeof(float)*tmp_sz))) {
-			CVEC_ASSERT(tmp != NULL);
-			return 0;
-		}
-		vec->a = tmp;
-		vec->capacity = tmp_sz;
-	}
-
-	vec->size += num;
-	return 1;
-}
-
-int cvec_insert_float(cvector_float* vec, cvec_sz i, float a)
-{
-	float* tmp;
-	cvec_sz tmp_sz;
-	if (vec->capacity > vec->size) {
-		CVEC_MEMMOVE(&vec->a[i+1], &vec->a[i], (vec->size-i)*sizeof(float));
-		vec->a[i] = a;
-	} else {
-		tmp_sz = CVEC_float_ALLOCATOR(vec->capacity);
-		if (!(tmp = (float*)CVEC_REALLOC(vec->a, sizeof(float)*tmp_sz))) {
-			CVEC_ASSERT(tmp != NULL);
-			return 0;
-		}
-		vec->a = tmp;
-		CVEC_MEMMOVE(&vec->a[i+1], &vec->a[i], (vec->size-i)*sizeof(float));
-		vec->a[i] = a;
-		vec->capacity = tmp_sz;
-	}
-
-	vec->size++;
-	return 1;
-}
-
-int cvec_insert_array_float(cvector_float* vec, cvec_sz i, float* a, cvec_sz num)
-{
-	float* tmp;
-	cvec_sz tmp_sz;
-	if (vec->capacity < vec->size + num) {
-		tmp_sz = vec->capacity + num + CVEC_float_SZ;
-		if (!(tmp = (float*)CVEC_REALLOC(vec->a, sizeof(float)*tmp_sz))) {
-			CVEC_ASSERT(tmp != NULL);
-			return 0;
-		}
-		vec->a = tmp;
-		vec->capacity = tmp_sz;
-	}
-
-	CVEC_MEMMOVE(&vec->a[i+num], &vec->a[i], (vec->size-i)*sizeof(float));
-	CVEC_MEMMOVE(&vec->a[i], a, num*sizeof(float));
-	vec->size += num;
-	return 1;
-}
-
-float cvec_replace_float(cvector_float* vec, cvec_sz i, float a)
-{
-	float tmp = vec->a[i];
-	vec->a[i] = a;
-	return tmp;
-}
-
-void cvec_erase_float(cvector_float* vec, cvec_sz start, cvec_sz end)
-{
-	cvec_sz d = end - start + 1;
-	CVEC_MEMMOVE(&vec->a[start], &vec->a[end+1], (vec->size-1-end)*sizeof(float));
-	vec->size -= d;
-}
-
-
-int cvec_reserve_float(cvector_float* vec, cvec_sz size)
-{
-	float* tmp;
-	if (vec->capacity < size) {
-		if (!(tmp = (float*)CVEC_REALLOC(vec->a, sizeof(float)*(size+CVEC_float_SZ)))) {
-			CVEC_ASSERT(tmp != NULL);
-			return 0;
-		}
-		vec->a = tmp;
-		vec->capacity = size + CVEC_float_SZ;
-	}
-	return 1;
-}
-
-int cvec_set_cap_float(cvector_float* vec, cvec_sz size)
-{
-	float* tmp;
-	if (size < vec->size) {
-		vec->size = size;
-	}
-
-	if (!(tmp = (float*)CVEC_REALLOC(vec->a, sizeof(float)*size))) {
-		CVEC_ASSERT(tmp != NULL);
-		return 0;
-	}
-	vec->a = tmp;
-	vec->capacity = size;
-	return 1;
-}
-
-void cvec_set_val_sz_float(cvector_float* vec, float val)
-{
-	cvec_sz i;
-	for (i=0; i<vec->size; i++) {
-		vec->a[i] = val;
-	}
-}
-
-void cvec_set_val_cap_float(cvector_float* vec, float val)
-{
-	cvec_sz i;
-	for (i=0; i<vec->capacity; i++) {
-		vec->a[i] = val;
-	}
-}
-
-void cvec_clear_float(cvector_float* vec) { vec->size = 0; }
-
-void cvec_free_float_heap(void* vec)
-{
-	cvector_float* tmp = (cvector_float*)vec;
-	if (!tmp) return;
-	CVEC_FREE(tmp->a);
-	CVEC_FREE(tmp);
-}
-
-void cvec_free_float(void* vec)
-{
-	cvector_float* tmp = (cvector_float*)vec;
-	CVEC_FREE(tmp->a);
-	tmp->size = 0;
-	tmp->capacity = 0;
-}
-
-
 static glContext* c;
 
 static Color blend_pixel(vec4 src, vec4 dst);
@@ -5972,7 +5622,7 @@ static void do_vertex(glVertex_Attrib* v, int* enabled, unsigned int num_enabled
 		c->vertex_attribs_vs[enabled[j]] = get_v_attrib(&v[enabled[j]], i);
 	}
 
-	float* vs_out = &c->vs_output.output_buf.a[vert*c->vs_output.size];
+	float* vs_out = &c->vs_output.output_buf[vert*c->vs_output.size];
 	c->programs.a[c->cur_program].vertex_shader(vs_out, c->vertex_attribs_vs, &c->builtins, c->programs.a[c->cur_program].uniform);
 
 	c->glverts.a[vert].vs_out = vs_out;
@@ -6217,7 +5867,7 @@ static int depthtest(float zval, float zbufval)
 
 static void setup_fs_input(float t, float* v1_out, float* v2_out, float wa, float wb, unsigned int provoke)
 {
-	float* vs_output = &c->vs_output.output_buf.a[0];
+	float* vs_output = &c->vs_output.output_buf[0];
 
 	float inv_wa = 1.0/wa;
 	float inv_wb = 1.0/wb;
@@ -7570,7 +7220,7 @@ static void draw_triangle_fill(glVertex* v0, glVertex* v1, glVertex* v2, unsigne
 	float alpha, beta, gamma, tmp, tmp2, z;
 	float fs_input[GL_MAX_VERTEX_OUTPUT_COMPONENTS];
 	float perspective[GL_MAX_VERTEX_OUTPUT_COMPONENTS*3];
-	float* vs_output = &c->vs_output.output_buf.a[0];
+	float* vs_output = &c->vs_output.output_buf[0];
 
 	for (int i=0; i<c->vs_output.size; ++i) {
 		perspective[i] = v0->vs_out[i]/p0.w;
@@ -8262,9 +7912,8 @@ int init_glContext(glContext* context, u32** back, int w, int h, int bitdepth, u
 	cvec_glTexture(&c->textures, 0, 1);
 	cvec_glVertex(&c->glverts, 0, 10);
 
-	//TODO might as well just set it to PGL_MAX_VERTICES * MAX_OUTPUT_COMPONENTS
-	cvec_float(&c->vs_output.output_buf, 0, 0);
-
+	// If not pre-allocating max, need to track size and edit glUseProgram and pglSetInterp
+	c->vs_output.output_buf = (float*)PGL_MALLOC(PGL_MAX_VERTICES * GL_MAX_VERTEX_OUTPUT_COMPONENTS * sizeof(float));
 
 	c->clear_stencil = 0;
 	c->clear_color = make_Color(0, 0, 0, 0);
@@ -8276,7 +7925,6 @@ int init_glContext(glContext* context, u32** back, int w, int h, int bitdepth, u
 	c->depth_range_far = 1.0f;
 	make_viewport_matrix(c->vp_mat, 0, 0, w, h, 1);
 
-
 	//set flags
 	//TODO match order in structure definition
 	c->provoking_vert = GL_LAST_VERTEX_CONVENTION;
@@ -8400,7 +8048,7 @@ void free_glContext(glContext* ctx)
 	cvec_free_glTexture(&ctx->textures);
 	cvec_free_glVertex(&ctx->glverts);
 
-	cvec_free_float(&ctx->vs_output.output_buf);
+	PGL_FREE(ctx->vs_output.output_buf);
 
 	if (c == ctx) {
 		c = NULL;
@@ -10265,7 +9913,7 @@ void glDeleteProgram(GLuint program)
 
 void glUseProgram(GLuint program)
 {
-	// Not a problem is program is marked "deleted" already
+	// Not a problem if program is marked "deleted" already
 	if (program >= c->programs.size) {
 		if (!c->error)
 			c->error = GL_INVALID_VALUE;
@@ -10273,7 +9921,9 @@ void glUseProgram(GLuint program)
 	}
 
 	c->vs_output.size = c->programs.a[program].vs_output_size;
-	cvec_reserve_float(&c->vs_output.output_buf, c->vs_output.size * PGL_MAX_VERTICES);
+	// c->vs_output.output_buf was pre-allocated to max size needed in init_glContext
+	// otherwise would need to assure it's at least
+	// c->vs_output_size * PGL_MAX_VERTS * sizeof(float) right here
 	c->vs_output.interpolation = c->programs.a[program].interpolation;
 	c->fragdepth_or_discard = c->programs.a[program].fragdepth_or_discard;
 
@@ -11290,7 +10940,10 @@ void pglSetInterp(GLsizei n, GLenum* interpolation)
 	c->vs_output.size = n;
 
 	memcpy(c->programs.a[c->cur_program].interpolation, interpolation, n*sizeof(GLenum));
-	cvec_reserve_float(&c->vs_output.output_buf, n * PGL_MAX_VERTICES);
+
+	// c->vs_output.output_buf was pre-allocated to max size needed in init_glContext
+	// otherwise would need to assure it's at least
+	// c->vs_output_size * PGL_MAX_VERTS * sizeof(float) right here
 
 	//vs_output.interpolation would be already pointing at current program's array
 	//unless the programs array was realloced since the last glUseProgram because
diff --git a/portablegl_unsafe.h b/portablegl_unsafe.h
index d678dd9..8398d92 100644
--- a/portablegl_unsafe.h
+++ b/portablegl_unsafe.h
@@ -1979,58 +1979,6 @@ static inline std::ostream& operator<<(std::ostream& stream, const vec4& a)
 /* CRSW_MATH_H */
 #endif
 
-#ifndef CVEC_SIZE_T
-#include <stdlib.h>
-#define CVEC_SIZE_T size_t
-#endif
-
-#ifndef CVEC_SZ
-#define CVEC_SZ
-typedef CVEC_SIZE_T cvec_sz;
-#endif
-
-
-/** Data structure for float vector. */
-typedef struct cvector_float
-{
-	float* a;           /**< Array. */
-	cvec_sz size;       /**< Current size (amount you use when manipulating array directly). */
-	cvec_sz capacity;   /**< Allocated size of array; always >= size. */
-} cvector_float;
-
-
-
-extern cvec_sz CVEC_float_SZ;
-
-int cvec_float(cvector_float* vec, cvec_sz size, cvec_sz capacity);
-int cvec_init_float(cvector_float* vec, float* vals, cvec_sz num);
-
-cvector_float* cvec_float_heap(cvec_sz size, cvec_sz capacity);
-cvector_float* cvec_init_float_heap(float* vals, cvec_sz num);
-int cvec_copyc_float(void* dest, void* src);
-int cvec_copy_float(cvector_float* dest, cvector_float* src);
-
-int cvec_push_float(cvector_float* vec, float a);
-float cvec_pop_float(cvector_float* vec);
-
-int cvec_extend_float(cvector_float* vec, cvec_sz num);
-int cvec_insert_float(cvector_float* vec, cvec_sz i, float a);
-int cvec_insert_array_float(cvector_float* vec, cvec_sz i, float* a, cvec_sz num);
-float cvec_replace_float(cvector_float* vec, cvec_sz i, float a);
-void cvec_erase_float(cvector_float* vec, cvec_sz start, cvec_sz end);
-int cvec_reserve_float(cvector_float* vec, cvec_sz size);
-#define cvec_shrink_to_fit_float(vec) cvec_set_cap_float((vec), (vec)->size)
-int cvec_set_cap_float(cvector_float* vec, cvec_sz size);
-void cvec_set_val_sz_float(cvector_float* vec, float val);
-void cvec_set_val_cap_float(cvector_float* vec, float val);
-
-float* cvec_back_float(cvector_float* vec);
-
-void cvec_clear_float(cvector_float* vec);
-void cvec_free_float_heap(void* vec);
-void cvec_free_float(void* vec);
-
-
 
 
 #include <stdint.h>
@@ -2649,7 +2597,7 @@ typedef struct Vertex_Shader_output
 	// but still easily add back functions as needed...
 	//
 	// or like comment in init_glContext says just allocate to the max size and be done
-	cvector_float output_buf;
+	float* output_buf;
 } Vertex_Shader_output;
 
 
@@ -5525,304 +5473,6 @@ void cvec_free_glVertex(void* vec)
 }
 
 
-#if defined(CVEC_MALLOC) && defined(CVEC_FREE) && defined(CVEC_REALLOC)
-/* ok */
-#elif !defined(CVEC_MALLOC) && !defined(CVEC_FREE) && !defined(CVEC_REALLOC)
-/* ok */
-#else
-#error "Must define all or none of CVEC_MALLOC, CVEC_FREE, and CVEC_REALLOC."
-#endif
-
-#ifndef CVEC_MALLOC
-#include <stdlib.h>
-#define CVEC_MALLOC(sz)      malloc(sz)
-#define CVEC_REALLOC(p, sz)  realloc(p, sz)
-#define CVEC_FREE(p)         free(p)
-#endif
-
-#ifndef CVEC_MEMMOVE
-#include <string.h>
-#define CVEC_MEMMOVE(dst, src, sz)  memmove(dst, src, sz)
-#endif
-
-#ifndef CVEC_ASSERT
-#include <assert.h>
-#define CVEC_ASSERT(x)       assert(x)
-#endif
-
-cvec_sz CVEC_float_SZ = 50;
-
-#define CVEC_float_ALLOCATOR(x) ((x+1) * 2)
-
-cvector_float* cvec_float_heap(cvec_sz size, cvec_sz capacity)
-{
-	cvector_float* vec;
-	if (!(vec = (cvector_float*)CVEC_MALLOC(sizeof(cvector_float)))) {
-		CVEC_ASSERT(vec != NULL);
-		return NULL;
-	}
-
-	vec->size = size;
-	vec->capacity = (capacity > vec->size || (vec->size && capacity == vec->size)) ? capacity : vec->size + CVEC_float_SZ;
-
-	if (!(vec->a = (float*)CVEC_MALLOC(vec->capacity*sizeof(float)))) {
-		CVEC_ASSERT(vec->a != NULL);
-		CVEC_FREE(vec);
-		return NULL;
-	}
-
-	return vec;
-}
-
-cvector_float* cvec_init_float_heap(float* vals, cvec_sz num)
-{
-	cvector_float* vec;
-	
-	if (!(vec = (cvector_float*)CVEC_MALLOC(sizeof(cvector_float)))) {
-		CVEC_ASSERT(vec != NULL);
-		return NULL;
-	}
-
-	vec->capacity = num + CVEC_float_SZ;
-	vec->size = num;
-	if (!(vec->a = (float*)CVEC_MALLOC(vec->capacity*sizeof(float)))) {
-		CVEC_ASSERT(vec->a != NULL);
-		CVEC_FREE(vec);
-		return NULL;
-	}
-
-	CVEC_MEMMOVE(vec->a, vals, sizeof(float)*num);
-
-	return vec;
-}
-
-int cvec_float(cvector_float* vec, cvec_sz size, cvec_sz capacity)
-{
-	vec->size = size;
-	vec->capacity = (capacity > vec->size || (vec->size && capacity == vec->size)) ? capacity : vec->size + CVEC_float_SZ;
-
-	if (!(vec->a = (float*)CVEC_MALLOC(vec->capacity*sizeof(float)))) {
-		CVEC_ASSERT(vec->a != NULL);
-		vec->size = vec->capacity = 0;
-		return 0;
-	}
-
-	return 1;
-}
-
-int cvec_init_float(cvector_float* vec, float* vals, cvec_sz num)
-{
-	vec->capacity = num + CVEC_float_SZ;
-	vec->size = num;
-	if (!(vec->a = (float*)CVEC_MALLOC(vec->capacity*sizeof(float)))) {
-		CVEC_ASSERT(vec->a != NULL);
-		vec->size = vec->capacity = 0;
-		return 0;
-	}
-
-	CVEC_MEMMOVE(vec->a, vals, sizeof(float)*num);
-
-	return 1;
-}
-
-int cvec_copyc_float(void* dest, void* src)
-{
-	cvector_float* vec1 = (cvector_float*)dest;
-	cvector_float* vec2 = (cvector_float*)src;
-
-	vec1->a = NULL;
-	vec1->size = 0;
-	vec1->capacity = 0;
-
-	return cvec_copy_float(vec1, vec2);
-}
-
-int cvec_copy_float(cvector_float* dest, cvector_float* src)
-{
-	float* tmp = NULL;
-	if (!(tmp = (float*)CVEC_REALLOC(dest->a, src->capacity*sizeof(float)))) {
-		CVEC_ASSERT(tmp != NULL);
-		return 0;
-	}
-	dest->a = tmp;
-
-	CVEC_MEMMOVE(dest->a, src->a, src->size*sizeof(float));
-	dest->size = src->size;
-	dest->capacity = src->capacity;
-	return 1;
-}
-
-
-int cvec_push_float(cvector_float* vec, float a)
-{
-	float* tmp;
-	cvec_sz tmp_sz;
-	if (vec->capacity > vec->size) {
-		vec->a[vec->size++] = a;
-	} else {
-		tmp_sz = CVEC_float_ALLOCATOR(vec->capacity);
-		if (!(tmp = (float*)CVEC_REALLOC(vec->a, sizeof(float)*tmp_sz))) {
-			CVEC_ASSERT(tmp != NULL);
-			return 0;
-		}
-		vec->a = tmp;
-		vec->a[vec->size++] = a;
-		vec->capacity = tmp_sz;
-	}
-	return 1;
-}
-
-float cvec_pop_float(cvector_float* vec)
-{
-	return vec->a[--vec->size];
-}
-
-float* cvec_back_float(cvector_float* vec)
-{
-	return &vec->a[vec->size-1];
-}
-
-int cvec_extend_float(cvector_float* vec, cvec_sz num)
-{
-	float* tmp;
-	cvec_sz tmp_sz;
-	if (vec->capacity < vec->size + num) {
-		tmp_sz = vec->capacity + num + CVEC_float_SZ;
-		if (!(tmp = (float*)CVEC_REALLOC(vec->a, sizeof(float)*tmp_sz))) {
-			CVEC_ASSERT(tmp != NULL);
-			return 0;
-		}
-		vec->a = tmp;
-		vec->capacity = tmp_sz;
-	}
-
-	vec->size += num;
-	return 1;
-}
-
-int cvec_insert_float(cvector_float* vec, cvec_sz i, float a)
-{
-	float* tmp;
-	cvec_sz tmp_sz;
-	if (vec->capacity > vec->size) {
-		CVEC_MEMMOVE(&vec->a[i+1], &vec->a[i], (vec->size-i)*sizeof(float));
-		vec->a[i] = a;
-	} else {
-		tmp_sz = CVEC_float_ALLOCATOR(vec->capacity);
-		if (!(tmp = (float*)CVEC_REALLOC(vec->a, sizeof(float)*tmp_sz))) {
-			CVEC_ASSERT(tmp != NULL);
-			return 0;
-		}
-		vec->a = tmp;
-		CVEC_MEMMOVE(&vec->a[i+1], &vec->a[i], (vec->size-i)*sizeof(float));
-		vec->a[i] = a;
-		vec->capacity = tmp_sz;
-	}
-
-	vec->size++;
-	return 1;
-}
-
-int cvec_insert_array_float(cvector_float* vec, cvec_sz i, float* a, cvec_sz num)
-{
-	float* tmp;
-	cvec_sz tmp_sz;
-	if (vec->capacity < vec->size + num) {
-		tmp_sz = vec->capacity + num + CVEC_float_SZ;
-		if (!(tmp = (float*)CVEC_REALLOC(vec->a, sizeof(float)*tmp_sz))) {
-			CVEC_ASSERT(tmp != NULL);
-			return 0;
-		}
-		vec->a = tmp;
-		vec->capacity = tmp_sz;
-	}
-
-	CVEC_MEMMOVE(&vec->a[i+num], &vec->a[i], (vec->size-i)*sizeof(float));
-	CVEC_MEMMOVE(&vec->a[i], a, num*sizeof(float));
-	vec->size += num;
-	return 1;
-}
-
-float cvec_replace_float(cvector_float* vec, cvec_sz i, float a)
-{
-	float tmp = vec->a[i];
-	vec->a[i] = a;
-	return tmp;
-}
-
-void cvec_erase_float(cvector_float* vec, cvec_sz start, cvec_sz end)
-{
-	cvec_sz d = end - start + 1;
-	CVEC_MEMMOVE(&vec->a[start], &vec->a[end+1], (vec->size-1-end)*sizeof(float));
-	vec->size -= d;
-}
-
-
-int cvec_reserve_float(cvector_float* vec, cvec_sz size)
-{
-	float* tmp;
-	if (vec->capacity < size) {
-		if (!(tmp = (float*)CVEC_REALLOC(vec->a, sizeof(float)*(size+CVEC_float_SZ)))) {
-			CVEC_ASSERT(tmp != NULL);
-			return 0;
-		}
-		vec->a = tmp;
-		vec->capacity = size + CVEC_float_SZ;
-	}
-	return 1;
-}
-
-int cvec_set_cap_float(cvector_float* vec, cvec_sz size)
-{
-	float* tmp;
-	if (size < vec->size) {
-		vec->size = size;
-	}
-
-	if (!(tmp = (float*)CVEC_REALLOC(vec->a, sizeof(float)*size))) {
-		CVEC_ASSERT(tmp != NULL);
-		return 0;
-	}
-	vec->a = tmp;
-	vec->capacity = size;
-	return 1;
-}
-
-void cvec_set_val_sz_float(cvector_float* vec, float val)
-{
-	cvec_sz i;
-	for (i=0; i<vec->size; i++) {
-		vec->a[i] = val;
-	}
-}
-
-void cvec_set_val_cap_float(cvector_float* vec, float val)
-{
-	cvec_sz i;
-	for (i=0; i<vec->capacity; i++) {
-		vec->a[i] = val;
-	}
-}
-
-void cvec_clear_float(cvector_float* vec) { vec->size = 0; }
-
-void cvec_free_float_heap(void* vec)
-{
-	cvector_float* tmp = (cvector_float*)vec;
-	if (!tmp) return;
-	CVEC_FREE(tmp->a);
-	CVEC_FREE(tmp);
-}
-
-void cvec_free_float(void* vec)
-{
-	cvector_float* tmp = (cvector_float*)vec;
-	CVEC_FREE(tmp->a);
-	tmp->size = 0;
-	tmp->capacity = 0;
-}
-
-
 static glContext* c;
 
 static Color blend_pixel(vec4 src, vec4 dst);
@@ -5972,7 +5622,7 @@ static void do_vertex(glVertex_Attrib* v, int* enabled, unsigned int num_enabled
 		c->vertex_attribs_vs[enabled[j]] = get_v_attrib(&v[enabled[j]], i);
 	}
 
-	float* vs_out = &c->vs_output.output_buf.a[vert*c->vs_output.size];
+	float* vs_out = &c->vs_output.output_buf[vert*c->vs_output.size];
 	c->programs.a[c->cur_program].vertex_shader(vs_out, c->vertex_attribs_vs, &c->builtins, c->programs.a[c->cur_program].uniform);
 
 	c->glverts.a[vert].vs_out = vs_out;
@@ -6217,7 +5867,7 @@ static int depthtest(float zval, float zbufval)
 
 static void setup_fs_input(float t, float* v1_out, float* v2_out, float wa, float wb, unsigned int provoke)
 {
-	float* vs_output = &c->vs_output.output_buf.a[0];
+	float* vs_output = &c->vs_output.output_buf[0];
 
 	float inv_wa = 1.0/wa;
 	float inv_wb = 1.0/wb;
@@ -7570,7 +7220,7 @@ static void draw_triangle_fill(glVertex* v0, glVertex* v1, glVertex* v2, unsigne
 	float alpha, beta, gamma, tmp, tmp2, z;
 	float fs_input[GL_MAX_VERTEX_OUTPUT_COMPONENTS];
 	float perspective[GL_MAX_VERTEX_OUTPUT_COMPONENTS*3];
-	float* vs_output = &c->vs_output.output_buf.a[0];
+	float* vs_output = &c->vs_output.output_buf[0];
 
 	for (int i=0; i<c->vs_output.size; ++i) {
 		perspective[i] = v0->vs_out[i]/p0.w;
@@ -8262,9 +7912,8 @@ int init_glContext(glContext* context, u32** back, int w, int h, int bitdepth, u
 	cvec_glTexture(&c->textures, 0, 1);
 	cvec_glVertex(&c->glverts, 0, 10);
 
-	//TODO might as well just set it to PGL_MAX_VERTICES * MAX_OUTPUT_COMPONENTS
-	cvec_float(&c->vs_output.output_buf, 0, 0);
-
+	// If not pre-allocating max, need to track size and edit glUseProgram and pglSetInterp
+	c->vs_output.output_buf = (float*)PGL_MALLOC(PGL_MAX_VERTICES * GL_MAX_VERTEX_OUTPUT_COMPONENTS * sizeof(float));
 
 	c->clear_stencil = 0;
 	c->clear_color = make_Color(0, 0, 0, 0);
@@ -8276,7 +7925,6 @@ int init_glContext(glContext* context, u32** back, int w, int h, int bitdepth, u
 	c->depth_range_far = 1.0f;
 	make_viewport_matrix(c->vp_mat, 0, 0, w, h, 1);
 
-
 	//set flags
 	//TODO match order in structure definition
 	c->provoking_vert = GL_LAST_VERTEX_CONVENTION;
@@ -8400,7 +8048,7 @@ void free_glContext(glContext* ctx)
 	cvec_free_glTexture(&ctx->textures);
 	cvec_free_glVertex(&ctx->glverts);
 
-	cvec_free_float(&ctx->vs_output.output_buf);
+	PGL_FREE(ctx->vs_output.output_buf);
 
 	if (c == ctx) {
 		c = NULL;
@@ -9662,7 +9310,9 @@ void glDeleteProgram(GLuint program)
 void glUseProgram(GLuint program)
 {
 	c->vs_output.size = c->programs.a[program].vs_output_size;
-	cvec_reserve_float(&c->vs_output.output_buf, c->vs_output.size * PGL_MAX_VERTICES);
+	// c->vs_output.output_buf was pre-allocated to max size needed in init_glContext
+	// otherwise would need to assure it's at least
+	// c->vs_output_size * PGL_MAX_VERTS * sizeof(float) right here
 	c->vs_output.interpolation = c->programs.a[program].interpolation;
 	c->fragdepth_or_discard = c->programs.a[program].fragdepth_or_discard;
 
@@ -10547,7 +10197,10 @@ void pglSetInterp(GLsizei n, GLenum* interpolation)
 	c->vs_output.size = n;
 
 	memcpy(c->programs.a[c->cur_program].interpolation, interpolation, n*sizeof(GLenum));
-	cvec_reserve_float(&c->vs_output.output_buf, n * PGL_MAX_VERTICES);
+
+	// c->vs_output.output_buf was pre-allocated to max size needed in init_glContext
+	// otherwise would need to assure it's at least
+	// c->vs_output_size * PGL_MAX_VERTS * sizeof(float) right here
 
 	//vs_output.interpolation would be already pointing at current program's array
 	//unless the programs array was realloced since the last glUseProgram because
diff --git a/src/generate_gl_h.py b/src/generate_gl_h.py
index 5afcc5d..dacb33c 100755
--- a/src/generate_gl_h.py
+++ b/src/generate_gl_h.py
@@ -219,7 +219,7 @@ def get_gl_funcs():
     gl_h.write(open("crsw_math.h").read())
 
     # we actually use this for output_buf in gl_types..for now
-    gl_h.write(open("cvector_float.h").read())
+    #gl_h.write(open("cvector_float.h").read())
 
     gl_h.write(open("gl_types.h").read())
 
@@ -247,7 +247,8 @@ def get_gl_funcs():
     # maybe I should stick to using cvector_macro.h and use the macros
     # maybe an option to add vectors for commonly used types
     gl_h.write(open("cvector_combined.c").read())
-    gl_h.write(open("cvector_float.c").read())
+
+    #gl_h.write(open("cvector_float.c").read())
 
     # maybe this should go last? does it matter beyond aesthetics?
     gl_h.write(open("gl_internal.c").read())
diff --git a/src/gl_impl.c b/src/gl_impl.c
index bcdd6a8..8421d57 100644
--- a/src/gl_impl.c
+++ b/src/gl_impl.c
@@ -207,9 +207,8 @@ int init_glContext(glContext* context, u32** back, int w, int h, int bitdepth, u
 	cvec_glTexture(&c->textures, 0, 1);
 	cvec_glVertex(&c->glverts, 0, 10);
 
-	//TODO might as well just set it to PGL_MAX_VERTICES * MAX_OUTPUT_COMPONENTS
-	cvec_float(&c->vs_output.output_buf, 0, 0);
-
+	// If not pre-allocating max, need to track size and edit glUseProgram and pglSetInterp
+	c->vs_output.output_buf = (float*)PGL_MALLOC(PGL_MAX_VERTICES * GL_MAX_VERTEX_OUTPUT_COMPONENTS * sizeof(float));
 
 	c->clear_stencil = 0;
 	c->clear_color = make_Color(0, 0, 0, 0);
@@ -221,7 +220,6 @@ int init_glContext(glContext* context, u32** back, int w, int h, int bitdepth, u
 	c->depth_range_far = 1.0f;
 	make_viewport_matrix(c->vp_mat, 0, 0, w, h, 1);
 
-
 	//set flags
 	//TODO match order in structure definition
 	c->provoking_vert = GL_LAST_VERTEX_CONVENTION;
@@ -345,7 +343,7 @@ void free_glContext(glContext* ctx)
 	cvec_free_glTexture(&ctx->textures);
 	cvec_free_glVertex(&ctx->glverts);
 
-	cvec_free_float(&ctx->vs_output.output_buf);
+	PGL_FREE(ctx->vs_output.output_buf);
 
 	if (c == ctx) {
 		c = NULL;
@@ -2210,7 +2208,7 @@ void glDeleteProgram(GLuint program)
 
 void glUseProgram(GLuint program)
 {
-	// Not a problem is program is marked "deleted" already
+	// Not a problem if program is marked "deleted" already
 	if (program >= c->programs.size) {
 		if (!c->error)
 			c->error = GL_INVALID_VALUE;
@@ -2218,7 +2216,9 @@ void glUseProgram(GLuint program)
 	}
 
 	c->vs_output.size = c->programs.a[program].vs_output_size;
-	cvec_reserve_float(&c->vs_output.output_buf, c->vs_output.size * PGL_MAX_VERTICES);
+	// c->vs_output.output_buf was pre-allocated to max size needed in init_glContext
+	// otherwise would need to assure it's at least
+	// c->vs_output_size * PGL_MAX_VERTS * sizeof(float) right here
 	c->vs_output.interpolation = c->programs.a[program].interpolation;
 	c->fragdepth_or_discard = c->programs.a[program].fragdepth_or_discard;
 
diff --git a/src/gl_impl_unsafe.c b/src/gl_impl_unsafe.c
index b0b8f4f..2f397d9 100644
--- a/src/gl_impl_unsafe.c
+++ b/src/gl_impl_unsafe.c
@@ -207,9 +207,8 @@ int init_glContext(glContext* context, u32** back, int w, int h, int bitdepth, u
 	cvec_glTexture(&c->textures, 0, 1);
 	cvec_glVertex(&c->glverts, 0, 10);
 
-	//TODO might as well just set it to PGL_MAX_VERTICES * MAX_OUTPUT_COMPONENTS
-	cvec_float(&c->vs_output.output_buf, 0, 0);
-
+	// If not pre-allocating max, need to track size and edit glUseProgram and pglSetInterp
+	c->vs_output.output_buf = (float*)PGL_MALLOC(PGL_MAX_VERTICES * GL_MAX_VERTEX_OUTPUT_COMPONENTS * sizeof(float));
 
 	c->clear_stencil = 0;
 	c->clear_color = make_Color(0, 0, 0, 0);
@@ -221,7 +220,6 @@ int init_glContext(glContext* context, u32** back, int w, int h, int bitdepth, u
 	c->depth_range_far = 1.0f;
 	make_viewport_matrix(c->vp_mat, 0, 0, w, h, 1);
 
-
 	//set flags
 	//TODO match order in structure definition
 	c->provoking_vert = GL_LAST_VERTEX_CONVENTION;
@@ -345,7 +343,7 @@ void free_glContext(glContext* ctx)
 	cvec_free_glTexture(&ctx->textures);
 	cvec_free_glVertex(&ctx->glverts);
 
-	cvec_free_float(&ctx->vs_output.output_buf);
+	PGL_FREE(ctx->vs_output.output_buf);
 
 	if (c == ctx) {
 		c = NULL;
@@ -1607,7 +1605,9 @@ void glDeleteProgram(GLuint program)
 void glUseProgram(GLuint program)
 {
 	c->vs_output.size = c->programs.a[program].vs_output_size;
-	cvec_reserve_float(&c->vs_output.output_buf, c->vs_output.size * PGL_MAX_VERTICES);
+	// c->vs_output.output_buf was pre-allocated to max size needed in init_glContext
+	// otherwise would need to assure it's at least
+	// c->vs_output_size * PGL_MAX_VERTS * sizeof(float) right here
 	c->vs_output.interpolation = c->programs.a[program].interpolation;
 	c->fragdepth_or_discard = c->programs.a[program].fragdepth_or_discard;
 
diff --git a/src/gl_internal.c b/src/gl_internal.c
index c7819d8..f0af8db 100644
--- a/src/gl_internal.c
+++ b/src/gl_internal.c
@@ -148,7 +148,7 @@ static void do_vertex(glVertex_Attrib* v, int* enabled, unsigned int num_enabled
 		c->vertex_attribs_vs[enabled[j]] = get_v_attrib(&v[enabled[j]], i);
 	}
 
-	float* vs_out = &c->vs_output.output_buf.a[vert*c->vs_output.size];
+	float* vs_out = &c->vs_output.output_buf[vert*c->vs_output.size];
 	c->programs.a[c->cur_program].vertex_shader(vs_out, c->vertex_attribs_vs, &c->builtins, c->programs.a[c->cur_program].uniform);
 
 	c->glverts.a[vert].vs_out = vs_out;
@@ -393,7 +393,7 @@ static int depthtest(float zval, float zbufval)
 
 static void setup_fs_input(float t, float* v1_out, float* v2_out, float wa, float wb, unsigned int provoke)
 {
-	float* vs_output = &c->vs_output.output_buf.a[0];
+	float* vs_output = &c->vs_output.output_buf[0];
 
 	float inv_wa = 1.0/wa;
 	float inv_wb = 1.0/wb;
@@ -1746,7 +1746,7 @@ static void draw_triangle_fill(glVertex* v0, glVertex* v1, glVertex* v2, unsigne
 	float alpha, beta, gamma, tmp, tmp2, z;
 	float fs_input[GL_MAX_VERTEX_OUTPUT_COMPONENTS];
 	float perspective[GL_MAX_VERTEX_OUTPUT_COMPONENTS*3];
-	float* vs_output = &c->vs_output.output_buf.a[0];
+	float* vs_output = &c->vs_output.output_buf[0];
 
 	for (int i=0; i<c->vs_output.size; ++i) {
 		perspective[i] = v0->vs_out[i]/p0.w;
diff --git a/src/gl_types.h b/src/gl_types.h
index 0e24804..7fd8d50 100644
--- a/src/gl_types.h
+++ b/src/gl_types.h
@@ -617,7 +617,7 @@ typedef struct Vertex_Shader_output
 	// but still easily add back functions as needed...
 	//
 	// or like comment in init_glContext says just allocate to the max size and be done
-	cvector_float output_buf;
+	float* output_buf;
 } Vertex_Shader_output;
 
 
diff --git a/src/pgl_ext.c b/src/pgl_ext.c
index 600768a..f992fed 100644
--- a/src/pgl_ext.c
+++ b/src/pgl_ext.c
@@ -23,7 +23,10 @@ void pglSetInterp(GLsizei n, GLenum* interpolation)
 	c->vs_output.size = n;
 
 	memcpy(c->programs.a[c->cur_program].interpolation, interpolation, n*sizeof(GLenum));
-	cvec_reserve_float(&c->vs_output.output_buf, n * PGL_MAX_VERTICES);
+
+	// c->vs_output.output_buf was pre-allocated to max size needed in init_glContext
+	// otherwise would need to assure it's at least
+	// c->vs_output_size * PGL_MAX_VERTS * sizeof(float) right here
 
 	//vs_output.interpolation would be already pointing at current program's array
 	//unless the programs array was realloced since the last glUseProgram because