From b0a021c07aa38905090058bd57b9304425594557 Mon Sep 17 00:00:00 2001 From: Ajax Date: Tue, 23 Oct 2018 15:44:48 -0400 Subject: Made a userspace that allows you to use your keyboard as an order of operations based calculator (#2864) * Made DZ60 layout with calculator * Cleaned up and commented, preparing to fix bug with negative in front of open parenthesis as first character * Fixed bug where negative sign infront of parenthesis as first character was parsed incorrectly * Made a better solution for the bug from the previous commit * Modularized and added a userfile so that this code can be used on various keyboards, found in doogle999 * Removed commented code from keymap * Made the layer that is used for calculations a define so that it can be changed per keyboard * Made the readme * Made the readme in the correct place * Revert "Made the readme in the correct place" This reverts commit 7f8b59ed9e59c77401a48be3a7ac1e8fd8e84e32. * Manually synced with qmk upstream * Stopped repeat, made keys print character that they are defined as rather than what the keyboard wants them to do * Added support for numpad, might make all keycodes custom so that there is no need to change doogle999.c if you want to change the keycode that is associated with a function, also made numpad automatically activating an option * Fixed some bugs with backspacing, updated the readme * Fixed some bugs with numlock turning on at the wrong times when a shift key was down * Made the return to layer work automatically instead of just forcing it to layer 0 * fixes and style changes, 20% decreased binary size * Fixed some bugs with double printing and compilation errors on my side * Fixed bug with exceeding the buffer size * Removed changes that added const-ness * Made changes so that backspace does not repeat to remove backspace bugs, still some bugs with recalculating without having typed anything * Fixed obo error with calc main loop * Made includes more accurate in keymap for dz60 * Moved flags to user makefile --- users/doogle999/doogle999.c | 460 ++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 460 insertions(+) create mode 100644 users/doogle999/doogle999.c (limited to 'users/doogle999/doogle999.c') diff --git a/users/doogle999/doogle999.c b/users/doogle999/doogle999.c new file mode 100644 index 000000000..320de7cff --- /dev/null +++ b/users/doogle999/doogle999.c @@ -0,0 +1,460 @@ +#include "doogle999.h" + +static unsigned char inputLocation = 0; // Current index in text input + +static double calc(const char input[CALC_BUFFER_SIZE +1]) // Finds value of input char array, relatively small and fast I think +{ + char inputToken[CALC_BUFFER_SIZE + 1]; // Input buffer, used when a single token (generally a number) takes up more + unsigned char inputTokenLocation = 0, inputLocation = 0; // Keep track of indices + + struct Token tokens[CALC_BUFFER_SIZE + 1]; // Input, converted to tokens, one extra large to accomodate for possible negative sign then open parenthesis as first character + unsigned char tokenCount = 0; // Keep track of index + + bool dashAsMinus = false; // Kind of a hacky solution to determining whether to treat a dash as a minus sign or a negative sign + + while(inputLocation < CALC_BUFFER_SIZE + 1) + { + char digit = input[inputLocation]; + + if(inputLocation == 0 && input[inputLocation] == CALC_CHAR_SUB && input[inputLocation + 1] == CALC_CHAR_BEG) + { + tokens[tokenCount].raw.num = 0; + tokens[tokenCount].isNum = true; + + tokenCount++; + dashAsMinus = true; + } + + if ((digit >= '0' && digit <= '9') || /* valid digit */ + (inputTokenLocation != 0 && input[inputLocation] == CALC_CHAR_DEC) || /* valid floating point */ + (!dashAsMinus && inputTokenLocation == 0 && input[inputLocation] == CALC_CHAR_SUB)) /* - is negative sign */ + { + inputToken[inputTokenLocation] = input[inputLocation]; + inputTokenLocation++; + inputLocation++; + continue; + } + + if(inputTokenLocation != 0) + { + // sscanf(inputToken, "%lf", &tokens[tokenCount].raw.num); // I would like to use sscanf here, but the small version of stdio.h on the chip doesn't allow sscanf or its sister functions to be used to process floats + tokens[tokenCount].raw.num = atof(inputToken); + tokens[tokenCount].isNum = true; + for(unsigned char i = 0; i < inputTokenLocation + 1; i++) + { + inputToken[i] = '\0'; + } + inputTokenLocation = 0; + tokenCount++; + dashAsMinus = true; + continue; + } + + /* inputTokenLocation == 0 */ + tokens[tokenCount].isNum = false; + tokens[tokenCount].raw.op.c = input[inputLocation]; + tokens[tokenCount].raw.op.priority = 0; + tokens[tokenCount].raw.op.ltr = true; + dashAsMinus = false; + + switch(input[inputLocation]) + { + case CALC_CHAR_BEG: + break; + case CALC_CHAR_END: + dashAsMinus = true; + break; + case CALC_CHAR_ADD: + tokens[tokenCount].raw.op.priority = CALC_PRIO_ADD; + break; + case CALC_CHAR_SUB: + tokens[tokenCount].raw.op.priority = CALC_PRIO_SUB; + break; + case CALC_CHAR_MUL: + tokens[tokenCount].raw.op.priority = CALC_PRIO_MUL; + break; + case CALC_CHAR_DIV: + tokens[tokenCount].raw.op.priority = CALC_PRIO_DIV; + break; + case CALC_CHAR_EXP: + tokens[tokenCount].raw.op.priority = CALC_PRIO_EXP; + tokens[tokenCount].raw.op.ltr = false; + break; + case CALC_CHAR_SIN: + case CALC_CHAR_COS: + case CALC_CHAR_TAN: + case CALC_CHAR_ASN: + case CALC_CHAR_ACS: + case CALC_CHAR_ATN: + case CALC_CHAR_LGE: + case CALC_CHAR_LOG: + case CALC_CHAR_SQT: + break; + case CALC_CHAR_EUL: + tokens[tokenCount].isNum = true; + tokens[tokenCount].raw.num = CALC_VALU_EUL; + dashAsMinus = true; + break; + case CALC_CHAR_PI: + tokens[tokenCount].isNum = true; + tokens[tokenCount].raw.num = CALC_VALU_PI; + dashAsMinus = true; + break; + case '\0': + tokenCount--; + inputLocation = CALC_BUFFER_SIZE; + break; + default: + tokenCount--; + break; + } + tokenCount++; + inputLocation++; + } + + struct Token output[CALC_BUFFER_SIZE + 1]; // Final output tokens before evaluation + struct Token opstack[CALC_BUFFER_SIZE + 1]; // Stack of operators + unsigned char outputLocation = 0, opstackLocation = 0; // Keep track of indices + + unsigned char numBrackets = 0; // The number of parenthesis + + for(unsigned char i = 0; i < tokenCount; i++) + { + if(tokens[i].isNum) + { + output[outputLocation] = tokens[i]; + outputLocation++; + } + else if(tokens[i].raw.op.c == CALC_CHAR_BEG) + { + opstack[opstackLocation] = tokens[i]; + opstackLocation++; + } + else if(tokens[i].raw.op.c == CALC_CHAR_END) + { + while(opstack[opstackLocation - 1].raw.op.c != CALC_CHAR_BEG) + { + output[outputLocation] = opstack[opstackLocation - 1]; + outputLocation++; + opstackLocation--; + } + opstackLocation--; + + numBrackets += 2; + } + else if(tokens[i].raw.op.priority == 0) + { + opstack[opstackLocation] = tokens[i]; + opstackLocation++; + } + else + { + while(opstackLocation != 0 + && (opstack[opstackLocation - 1].raw.op.priority == 0 + || tokens[i].raw.op.priority < opstack[opstackLocation - 1].raw.op.priority + || (tokens[i].raw.op.priority == opstack[opstackLocation - 1].raw.op.priority && opstack[opstackLocation - 1].raw.op.ltr)) + && opstack[opstackLocation - 1].raw.op.c != CALC_CHAR_BEG) + { + output[outputLocation] = opstack[opstackLocation - 1]; + outputLocation++; + opstackLocation--; + } + opstack[opstackLocation] = tokens[i]; + opstackLocation++; + } + } + + tokenCount -= numBrackets; + + for(signed char i = opstackLocation - 1; i >= 0; i--) + { + output[outputLocation] = opstack[i]; + outputLocation++; + opstackLocation--; + } + + double answer[CALC_BUFFER_SIZE]; + unsigned char answerLocation = 0; + + for(unsigned char i = 0; i < tokenCount; i++) + { + if(output[i].isNum) + { + answer[answerLocation] = output[i].raw.num; + answerLocation++; + continue; + } + + if(output[i].raw.op.priority == 0) + { + if (answerLocation < 1) { /* not handled here -- ERROR? */ } else + if(answerLocation >= 1) + { + double (*op)(double); + switch(output[i].raw.op.c) + { + case CALC_CHAR_SIN: + op = sin; + break; + case CALC_CHAR_COS: + op = cos; + break; + case CALC_CHAR_TAN: + op = tan; + break; + case CALC_CHAR_ASN: + op = asin; + break; + case CALC_CHAR_ACS: + op = acos; + break; + case CALC_CHAR_ATN: + op = atan; + break; + case CALC_CHAR_LGE: + op = log; + break; + case CALC_CHAR_LOG: + op = log10; + break; + case CALC_CHAR_SQT: + op = sqrt; + break; + default: + continue; /* invalid input */ + } + answer[answerLocation - 1] = op(answer[answerLocation - 1]); + } + } + /* priority != 0 */ + else if(answerLocation >= 2) + { + switch(output[i].raw.op.c) + { + case CALC_CHAR_ADD: + answer[answerLocation - 2] += answer[answerLocation - 1]; + break; + case CALC_CHAR_SUB: + answer[answerLocation - 2] -= answer[answerLocation - 1]; + break; + case CALC_CHAR_MUL: + answer[answerLocation - 2] *= answer[answerLocation - 1]; + break; + case CALC_CHAR_DIV: + answer[answerLocation - 2] /= answer[answerLocation - 1]; + break; + case CALC_CHAR_EXP: + answer[answerLocation - 2] = pow(answer[answerLocation - 2], answer[answerLocation - 1]); + break; + } + + answerLocation--; + } + } + + return answer[0]; +} + +/* + * @returns 0 when nothing should happen and QMK should work as usual + * @returns -1 when invalid input was given, QMK should ignore it + * @returns -2 when BSP should be done + * @returns -3 when CALC should be done + * @returns -4 when ENDCALC should be done + * @returns positive value of CALC_* when normal input was processed + */ +static int process_input(const uint16_t keycode, const uint8_t mods, const keyevent_t event) +{ + /* handle even when no key was pressed */ + if(!event.pressed) + { + switch(keycode) + { + /* QMK should handle those */ + case KC_RSFT: + case KC_LSFT: + return 0; + break; + } + /* ??? ignore */ + return -1; + } + + /* when shift key is pressed handle characters differently */ + char characterPressed; + if((get_mods() & MODS_SHIFT_MASK)) + { + switch(keycode) + { + case KC_9: + characterPressed = CALC_CHAR_BEG; + break; + case KC_0: + characterPressed = CALC_CHAR_END; + break; + case KC_EQUAL: + characterPressed = CALC_CHAR_ADD; + break; + case KC_KP_PLUS: + characterPressed = CALC_CHAR_ADD; + break; + case KC_6: + characterPressed = CALC_CHAR_EXP; + break; + case KC_8: + characterPressed = CALC_CHAR_MUL; + break; + case KC_KP_ASTERISK: + characterPressed = CALC_CHAR_MUL; + break; + case KC_S: + characterPressed = CALC_CHAR_ASN; + break; + case KC_C: + characterPressed = CALC_CHAR_ACS; + break; + case KC_T: + characterPressed = CALC_CHAR_ATN; + break; + case KC_L: + characterPressed = CALC_CHAR_LOG; + break; + default: + return -1; + break; + } + return characterPressed; + } + + /* normal key handling: shift not pressed */ + + /* digits */ + if (keycode == KC_KP_0 || keycode == KC_0) { + return '0'; + } else if (keycode >= KC_KP_1 && keycode <= KC_KP_9) { + return keycode - KC_KP_1 +1 + '0'; + } else if (keycode >= KC_1 && keycode <= KC_9) { + return keycode - KC_1 +1 + '0'; + } + + /* other tokens */ + switch (keycode) { + case KC_MINUS: + case KC_KP_MINUS: + return characterPressed = CALC_CHAR_SUB; + case KC_SLASH: + case KC_KP_SLASH: + return characterPressed = CALC_CHAR_DIV; + case KC_S: + return characterPressed = CALC_CHAR_SIN; + case KC_C: + return characterPressed = CALC_CHAR_COS; + case KC_T: + return characterPressed = CALC_CHAR_TAN; + case KC_Q: + return characterPressed = CALC_CHAR_SQT; + case KC_L: + return characterPressed = CALC_CHAR_LGE; + case KC_DOT: + case KC_KP_DOT: + return characterPressed = CALC_CHAR_DEC; + case KC_P: + return characterPressed = CALC_CHAR_PI; + case KC_E: + return characterPressed = CALC_CHAR_EUL; + case KC_BSPC: + return -2; + case KC_RSFT: + return 0; + case KC_LSFT: + return 0; + case CALC: + return -3; + case ENDCALC: + return -4; + default: + return -1; + } +} + +bool process_record_user(uint16_t keycode, keyrecord_t* record) +{ + static char text[CALC_BUFFER_SIZE + 1]; // Used to store input and then output when ready to print + static char backspaceText[CALC_BUFFER_SIZE + 1]; // Pretty dumb waste of memory because only backspace characters, used with send_string to backspace and remove input + + if((biton32(layer_state) == CALC_LAYER && CALC_FORCE_NUM_LOCK_INSIDE_CALC) || (biton32(layer_state) != CALC_LAYER && CALC_FORCE_NUM_LOCK_OUTSIDE_CALC)) + { + bool numpadKeyPressed = record->event.pressed && + !(get_mods() & MODS_SHIFT_MASK) && + /* KC_KP_1, KC_KP_2, ..., KC_KP_0, KC_KP_DOT */ + (keycode >= KC_KP_1 && keycode <= KC_KP_DOT); + + if(numpadKeyPressed && !(host_keyboard_leds() & (1 << USB_LED_NUM_LOCK))) + { + add_key(KC_NLCK); + send_keyboard_report(); + del_key(KC_NLCK); + } + } + + if(biton32(layer_state) != CALC_LAYER) { return true; } + + int action = process_input(keycode, get_mods(), record->event); + switch(action) + { + case 0: + return true; + case -1: + return false; + case -2: + if(inputLocation > 0) + { + inputLocation--; + text[inputLocation] = '\0'; + backspaceText[0] = (char)8; + backspaceText[1] = '\0'; + send_string(backspaceText); + } + return false; + case -3: + for(int i = 0; i < inputLocation; i++) + { + backspaceText[i] = (char)8; + } + send_string(backspaceText); + dtostrf(calc(text), CALC_PRINT_SIZE, CALC_PRINT_SIZE, text); + send_string(text); + for(unsigned char i = 0; i < CALC_BUFFER_SIZE; i++) + { + text[i] = '\0'; + backspaceText[i] = '\0'; + } + inputLocation = 0; + return false; + case -4: + for(unsigned char i = 0; i < CALC_BUFFER_SIZE; i++) + { + text[i] = '\0'; + backspaceText[i] = '\0'; + } + inputLocation = 0; + layer_off(CALC_LAYER); + return false; + default: + break; + } + char characterPressed = (char)action; + + if(inputLocation < CALC_BUFFER_SIZE) + { + text[inputLocation] = characterPressed; + inputLocation++; + + char characterToSend[2]; + characterToSend[0] = characterPressed; + characterToSend[1] = '\0'; + + send_string(characterToSend); + } + return false; +} \ No newline at end of file -- cgit v1.2.3-24-g4f1b