Unify emulator and main VDP codebases

video/agon_ps2.h

@@ -23,11 +23,17 @@ uint32_t		mWheelAcc = MOUSE_DEFAULT_WHEELACC;	// Mouse wheel acceleration
 
 // Forward declarations
 //
-bool resetMousePositioner(uint16_t width, uint16_t height, fabgl::VGABaseController * display);
+#ifdef USERSPACE
+bool zdi_mode () { return false; }
+void zdi_enter () {}
+void zdi_process_cmd (uint8_t key) {}
+#else
+bool zdi_mode ();
+void zdi_enter ();
+void zdi_process_cmd (uint8_t key);
+#endif
 
-extern bool zdi_mode();
-extern void zdi_enter();
-extern void zdi_process_cmd(uint8_t key);
+bool resetMousePositioner(uint16_t width, uint16_t height, fabgl::VGABaseController * display);
 extern bool consoleMode;
 extern HardwareSerial DBGSerial;
 

video/context.h

@@ -96,7 +96,6 @@ class Context {
 		uint8_t			linePatternLength = 8;			// Dotted line pattern length
 		std::vector<uint16_t>	charToBitmap = std::vector<uint16_t>(256, 65535);	// character to bitmap mapping
 		bool			plottingText = false;			// Are we currently plotting text?
-
 		bool			logicalCoords = true;			// Use BBC BASIC logical coordinates
 
 		Point			origin;							// Screen origin
@@ -240,7 +239,7 @@ class Context {
 		bool usingSystemFont();
 		char getScreenChar(uint8_t x, uint8_t y);
 		char getScreenCharAt(uint16_t px, uint16_t py);
-		void mapCharToBitmap(char c, uint16_t bitmapId);
+		void mapCharToBitmap(uint8_t c, uint16_t bitmapId);
 		void unmapBitmapFromChars(uint16_t bitmapId);
 		void resetCharToBitmap();
 

video/context/fonts.h

@@ -195,7 +195,7 @@ char Context::getScreenCharAt(uint16_t px, uint16_t py) {
 	return getScreenChar(toScreenCoordinates(px, py));
 }
 
-void Context::mapCharToBitmap(char c, uint16_t bitmapId) {
+void Context::mapCharToBitmap(uint8_t c, uint16_t bitmapId) {
 	auto bitmap = getBitmap(bitmapId);
 	if (bitmap) {
 		charToBitmap[c] = bitmapId;

video/context/graphics.h

@@ -14,6 +14,7 @@
 #include "buffers.h"
 #include "sprites.h"
 #include "types.h"
+#include "mat.h"
 
 // Definitions for the functions we're implementing here
 #include "context.h"

video/hexload.h

@@ -2,14 +2,18 @@
 #define HEXLOAD_H
 
 #include <stdbool.h>
+#ifndef USERSPACE
 #include "CRC16.h"
 #include "CRC32.h"
+#endif /* !USERSPACE */
 
 extern void printFmt(const char *format, ...);
 extern HardwareSerial DBGSerial;
 
+#ifndef USERSPACE
 CRC16 linecrc16(0x8005, 0x0, 0x0, false, false);
 CRC32 crc32,crc32tmp;
+#endif /* !USERSPACE */
 bool aborted;
 
 #define DEF_LOAD_ADDRESS			0x040000
@@ -57,13 +61,15 @@ void consumeHexMarker(void) {
 
 // Receive a single iHex Nibble from the external Debug serial interface
 uint8_t getIHexNibble(bool addcrc) {
+#ifndef USERSPACE
 	uint8_t nibble, input;
 	input = toupper(serialRx_t());
 	if(addcrc) linecrc16.add(input);
 	if((input >= '0') && input <='9') nibble = input - '0';
 	else nibble = input - 'A' + 10;
 	// illegal characters will be dealt with by checksum later
 	return nibble;
+#endif /* !USERSPACE */
 }
 
 // Receive a byte from the external Debug serial interface as two iHex nibbles
@@ -106,6 +112,10 @@ void serialTx(uint32_t crc) {
 }
 
 void VDUStreamProcessor::vdu_sys_hexload(void) {
+#ifdef USERSPACE
+	// no hexload for emulators :)
+	return;
+#else /* !USERSPACE */
 	uint32_t 	segment_address;
 	uint32_t 	crc32target;
 	uint8_t 	u,h,l,tmp;
@@ -290,6 +300,7 @@ void VDUStreamProcessor::vdu_sys_hexload(void) {
 	}
 	if(rom_area) printFmt("\r\nHEX data overlapping ROM area, transfer unsuccessful\r\nERROR\r\n");
 	printFmt("VDP done\r\n");   
+#endif /* !USERSPACE */
 }
 
-#endif // HEXLOAD_H
+#endif // HEXLOAD_H

video/types.h

@@ -36,6 +36,7 @@
 
 #include <cstddef>
 #include <cstdint>
+#include <cmath>
 #include <optional>
 #include <memory>
 
@@ -114,6 +115,12 @@ class psram_allocator
 	}
 };
 
+template <typename T, typename U>
+bool operator==(const psram_allocator<T>&, const psram_allocator<U>&) { return true; }
+
+template <typename T, typename U>
+bool operator!=(const psram_allocator<T>&a, const psram_allocator<U>&b) { return !(a == b); }
+
 // Typically we do not need a deleter because the regular one can handle PSRAM deallocations just fine,
 // but for completeness, here it is.
 

video/vdu_buffered.h

@@ -253,7 +253,7 @@ void IRAM_ATTR VDUStreamProcessor::vdu_sys_buffered() {
 			}
 			auto matrixSize = getMatrixSize(bufferId);
 			if (matrixSize.value != 0) {
-				float transform[matrixSize.size()] = {0.0f};
+				float transform[matrixSize.size()];
 				if (getMatrixFromBuffer(bufferId, transform, matrixSize)) {
 					force_debug_log("buffer contains a %d x %d matrix with contents:\n\r", matrixSize.rows, matrixSize.columns);
 					for (int i = 0; i < matrixSize.rows; i++) {
@@ -1580,7 +1580,8 @@ void VDUStreamProcessor::bufferAffineTransform(uint16_t bufferId, uint8_t comman
 	MatrixSize size;
 	size.rows = dimensions + 1;
 	size.columns = size.rows;
-	float transform[size.size()] = {0.0f};
+	float transform[size.size()];
+	memset(transform, 0, sizeof(transform));
 	for (int i = 0; i < size.rows; i++) {
 		transform[i * size.rows + i] = 1.0f;
 	}
@@ -1657,7 +1658,8 @@ void VDUStreamProcessor::bufferAffineTransform(uint16_t bufferId, uint8_t comman
 		}	break;
 		case AFFINE_SCALE: {
 			// scale by a given factor in each dimension
-			float scales[dimensions] = {0.0f};
+			float scales[dimensions];
+			memset(scales, 0, sizeof(scales));
 			if (!readFloatArguments(scales, dimensions, useBufferValue, useAdvancedOffsets, useMultiFormat)) {
 				return;
 			}
@@ -1667,7 +1669,8 @@ void VDUStreamProcessor::bufferAffineTransform(uint16_t bufferId, uint8_t comman
 		}	break;
 		case AFFINE_TRANSLATE: {
 			// translate by a given amount
-			float translateXY[dimensions] = {0.0f};
+			float translateXY[dimensions];
+			memset(translateXY, 0, sizeof(translateXY));
 			if (!readFloatArguments(translateXY, dimensions, useBufferValue, useAdvancedOffsets, useMultiFormat)) {
 				return;
 			}
@@ -1677,7 +1680,8 @@ void VDUStreamProcessor::bufferAffineTransform(uint16_t bufferId, uint8_t comman
 		}	break;
 		case AFFINE_TRANSLATE_OS_COORDS: {
 			// translate by a given amount of pixels where x and y match current coordinate system scaling
-			float translateXY[dimensions] = {0.0f};
+			float translateXY[dimensions];
+			memset(translateXY, 0, sizeof(translateXY));
 			if (!readFloatArguments(translateXY, dimensions, useBufferValue, useAdvancedOffsets, useMultiFormat)) {
 				return;
 			}
@@ -1690,7 +1694,8 @@ void VDUStreamProcessor::bufferAffineTransform(uint16_t bufferId, uint8_t comman
 		}	break;
 		case AFFINE_SHEAR: {
 			// shear by a given amount
-			float shearXY[dimensions] = {0.0f};
+			float shearXY[dimensions];
+			memset(shearXY, 0, sizeof(shearXY));
 			if (!readFloatArguments(shearXY, dimensions, useBufferValue, useAdvancedOffsets, useMultiFormat)) {
 				return;
 			}
@@ -1702,7 +1707,8 @@ void VDUStreamProcessor::bufferAffineTransform(uint16_t bufferId, uint8_t comman
 		case AFFINE_SKEW_RAD: {
 			// skew by a given amount (angle)
 			bool conversion = op == AFFINE_SKEW ? DEG_TO_RAD : 1.0f;
-			float skewXY[dimensions] = {0.0f};
+			float skewXY[dimensions];
+			memset(skewXY, 0, sizeof(skewXY));
 			if (!readFloatArguments(skewXY, dimensions, useBufferValue, useAdvancedOffsets, useMultiFormat)) {
 				return;
 			}
@@ -1744,10 +1750,11 @@ void VDUStreamProcessor::bufferAffineTransform(uint16_t bufferId, uint8_t comman
 	if (!replace) {
 		// we are combining - for now, only if the existing matrix is the same size
 		// TODO consider handling different size matrices - could combine at larger size, and then truncate
-		float existing[size.size()] = {0.0f};
+		float existing[size.size()];
+		memset(existing, 0, sizeof(existing));
 		if (getMatrixFromBuffer(bufferId, existing, size, false)) {
 			// combine the two matrices together
-			float newTransform[size.size()] = {0.0f};
+			float newTransform[size.size()];
 			dspm_mult_f32(transform, existing, newTransform, size.rows, size.columns, size.columns);
 			// copy data from matrix back to our working transform matrix
 			memcpy(transform, newTransform, size.sizeBytes());
@@ -1775,7 +1782,8 @@ void VDUStreamProcessor::bufferMatrixManipulate(uint16_t bufferId, uint8_t comma
 	const bool useAdvancedOffsets = command & MATRIX_OP_ADVANCED_OFFSETS;
 	const bool useBufferValue = command & MATRIX_OP_BUFFER_VALUE;
 
-	float matrix[size.size()] = {0.0f};
+	float matrix[size.size()];
+	memset(matrix, 0, sizeof(matrix));
 
 	switch (op) {
 		case MATRIX_SET: {
@@ -1817,7 +1825,8 @@ void VDUStreamProcessor::bufferMatrixManipulate(uint16_t bufferId, uint8_t comma
 		case MATRIX_DIAGONAL: {
 			// diagonal matrix with given values
 			auto argCount = fabgl::imin(size.rows, size.columns);
-			float args[argCount] = {0.0f};
+			float args[argCount];
+			memset(args, 0, sizeof(args));
 			if (!readFloatArguments(args, argCount, useBufferValue, useAdvancedOffsets, false)) {
 				return;
 			}
@@ -1830,7 +1839,8 @@ void VDUStreamProcessor::bufferMatrixManipulate(uint16_t bufferId, uint8_t comma
 			auto sourceId1 = readWord_t(); if (sourceId1 == -1) return;
 			auto sourceId2 = readWord_t(); if (sourceId2 == -1) return;
 			// Get the matrixes, for our target size, padding or truncating as necessary
-			float source[size.size()] = {0.0f};
+			float source[size.size()];
+			memset(source, 0, sizeof(source));
 			if (!getMatrixFromBuffer(sourceId1, matrix, size) || !getMatrixFromBuffer(sourceId2, source, size)) {
 				debug_log("bufferMatrixManipulate: failed to read matrix from buffer %d or %d\n\r", sourceId1, sourceId2);
 				return;
@@ -1845,7 +1855,8 @@ void VDUStreamProcessor::bufferMatrixManipulate(uint16_t bufferId, uint8_t comma
 			auto sourceId1 = readWord_t(); if (sourceId1 == -1) return;
 			auto sourceId2 = readWord_t(); if (sourceId2 == -1) return;
 			// Get the matrixes, for our target size, padding or truncating as necessary
-			float source[size.size()] = {0.0f};
+			float source[size.size()];
+			memset(source, 0, sizeof(source));
 			if (!getMatrixFromBuffer(sourceId1, matrix, size) || !getMatrixFromBuffer(sourceId2, source, size)) {
 				debug_log("bufferMatrixManipulate: failed to read matrix from buffer %d or %d\n\r", sourceId1, sourceId2);
 				return;
@@ -1871,14 +1882,17 @@ void VDUStreamProcessor::bufferMatrixManipulate(uint16_t bufferId, uint8_t comma
 			MatrixSize resultSize;
 			resultSize.rows = dimensions;
 			resultSize.columns = dimensions;
-			float source1[resultSize.size()] = {0.0f};
-			float source2[resultSize.size()] = {0.0f};
+			float source1[resultSize.size()];
+			float source2[resultSize.size()];
+			memset(source1, 0, sizeof(source1));
+			memset(source2, 0, sizeof(source2));
 			if (!getMatrixFromBuffer(sourceId1, source1, resultSize) || !getMatrixFromBuffer(sourceId2, source2, resultSize)) {
 				debug_log("bufferMatrixManipulate: failed to read matrix from buffer %d or %d\n\r", sourceId1, sourceId2);
 				return;
 			}
 			// multiply values in source1 and source2
-			float result[resultSize.size()] = {0.0f};
+			float result[resultSize.size()];
+			memset(result, 0, sizeof(result));
 			dspm_mult_f32(source1, source2, result, resultSize.rows, resultSize.columns, resultSize.columns);
 			for (int row = 0; row < size.rows; row++) {
 				for (int column = 0; column < size.columns; column++) {
@@ -1913,7 +1927,8 @@ void VDUStreamProcessor::bufferMatrixManipulate(uint16_t bufferId, uint8_t comma
 				debug_log("bufferMatrixManipulate: source matrix %d not found\n\r", sourceId);
 				return;
 			}
-			float source[sourceSize.size()] = {0.0f};
+			float source[sourceSize.size()];
+			memset(source, 0, sizeof(source));
 			if (!getMatrixFromBuffer(sourceId, source, sourceSize)) {
 				debug_log("bufferMatrixManipulate: failed to read matrix from buffer %d\n\r", sourceId);
 				return;
@@ -1940,7 +1955,8 @@ void VDUStreamProcessor::bufferMatrixManipulate(uint16_t bufferId, uint8_t comma
 				return;
 			}
 			// read source matrix
-			float source[sourceSize.size()] = {0.0f};
+			float source[sourceSize.size()];
+			memset(source, 0, sizeof(source));
 			if (!getMatrixFromBuffer(sourceId, source, sourceSize)) {
 				debug_log("bufferMatrixManipulate: failed to read matrix from buffer %d\n\r", sourceId);
 				return;
@@ -2101,8 +2117,8 @@ void VDUStreamProcessor::bufferTransformBitmap(uint16_t bufferId, uint8_t option
 
 	debug_log("bufferTransformBitmap: width %d, height %d, xOffset %d, yOffset %d\n\r", width, height, xOffset, yOffset);
 
-    for (int y = 0; y < height; y++) {
-    	for (int x = 0; x < width; x++) {
+	for (int y = 0; y < height; y++) {
+		for (int x = 0; x < width; x++) {
 			// calculate the source pixel
 			// NB we will need to adjust x,y here if we are auto-translating
 			pos[0] = (float)x + xOffset;
@@ -2114,7 +2130,7 @@ void VDUStreamProcessor::bufferTransformBitmap(uint16_t bufferId, uint8_t option
 			}
 			destination[(int)y * width + (int)x] = srcPixel;
 		}
-    }
+	}
 
 	// save new bitmap data to target buffer
 	bufferClear(bufferId);

video/vdu_sys.h

@@ -390,7 +390,7 @@ void VDUStreamProcessor::sendCursorPosition() {
 //
 void VDUStreamProcessor::sendScreenChar(char c) {
 	uint8_t packet[] = {
-		c,
+		(uint8_t)c,
 	};
 	send_packet(PACKET_SCRCHAR, sizeof packet, packet);
 }