Что я здесь делаю не так (пытаясь запрограммировать парсер * .srt на C)?

Question

Я только что нашел файл *.srt, который mpv не может быть загружен.Итак, я подумал, давайте создадим свой собственный анализатор субтитров, который принимает путь субтитров в качестве аргумента командной строки.Вот что я попробовал:

/* Intended to be a program for parsing *.srt subtitles as an alternative to video players' */

#include <ncurses.h>
#include <unistd.h>

#define SEC_IN_MIN 60
#define MIN_IN_HR 60

long get_duration(FILE *fp); // to get the duration of a dialogue in seconds
long turn_to_sec(int hours, int minutes, int seconds); // returns sum of hours and minutes, all in seconds

int main(int argc, char **argv)
{
    FILE *fp;
    long sec;
    char ch;

    if(argv[1] == NULL)
    {
        printf("Please enter a filename!\n");
        return 1;
    }
    printf("Trying to open specified file %s\n",argv[1]);

    fp = fopen(argv[1],"r");
    if(fp == NULL)
    {
        printf("Error while opening file %s\n",argv[1]);
        return 1;
    }

    initscr(); // initialise nCurses window

    ch = getc(fp);
    while(ch != EOF)
    {
        clear();
        sec = get_duration(fp);
        while(1)
        {
            if((ch = getc(fp)) == '\n')
            {
                if((ch = getc(fp)) == '\n' || ch == EOF)
                    break;
                else
                    addch(ch);
            }
            addch(ch);
        }
        refresh();
        sleep(sec);
    }

    endwin(); // close nCurses
    fclose(fp); // close the file
    return 0;
}

long get_duration(FILE *fp)
{
    long duration = 0;
    char ch;
    short hour_start = 0, hour_end = 0, minute_start = 0, minute_end = 0, second_start = 0, second_end = 0;
    short count=0;

    /* just to get to the point where time-specs of the dialogue start */
    while((ch = getc(fp)) != '\n');

    /* extract characters until ':' to get hour_start */
    while((ch = getc(fp)) != 58)
    {
        hour_start += ch;
        count++;
    }
    hour_start -= (hour_start/(49*count));

    /* extract characters until ':' to get minute_start */
    count = 0;
    while((ch = getc(fp)) != 58)
    {
        minute_start += ch;
        count++;
    }
    minute_start -= (minute_start/(49*count));

    /* extract characters until ',' to get second_start */
    count = 0;
    while((ch = getc(fp)) != 44)
    {
        second_start += ch;
        count++;
    }
    second_start -= (second_start/(49*count));

    /* now, see if you can find a '>' */
    while((ch = getc(fp)) != 62);
    ch = getc(fp); // to get rid of that space after "-->"

    /* extract characters until ':' to get hour_end */
    while((ch = getc(fp)) != 58)
    {
        hour_end += ch;
        count++;
    }
    hour_end -= (hour_end/(49*count));

    /* extract characters until ':' to get minute_end */
    count = 0;
    while((ch = getc(fp)) != 58)
    {
        minute_end += ch;
        count++;
    }
    minute_end -= (minute_end/(49*count));

    /* extract characters until ',' to get second_end */;
    count = 0;
    while((ch = getc(fp)) != 44)
    {
        second_end += ch;
        count++;
    }
    second_end -= (second_end/(49*count));

    /* finally, gonna get those values */
    second_end -= second_start;
    minute_end -= minute_start;
    hour_end -= hour_start;

    duration += (turn_to_sec(hour_end, minute_end, second_end));

    /* positioning the fp to the right position just to keep the 'main()' :) */
    while((ch = getc(fp)) != '\n' || ch != EOF);

    return duration;
}

long turn_to_sec(int hours, int minutes, int seconds)
{
    long temp;

    /* manipulating hours */
    temp = hours;
    temp *= MIN_IN_HR;
    temp *= SEC_IN_MIN;

    seconds += temp;

    /* manipulating minutes */
    temp = minutes;
    temp *= SEC_IN_MIN;

    seconds += temp;

    return seconds;
}

При первой попытке я использовал время начала диалога в качестве длительности диалога , то есть end_time - start_time и вот почему эта часть отсутствовала:

/* extract characters until ':' to get hour_end */
    while((ch = getc(fp)) != 58)
    {
        hour_end += ch;
        count++;
    }
    hour_end = (hour_end/(49*count));

    /* extract characters until ':' to get minute_end */
    count = 0;
    while((ch = getc(fp)) != 58)
    {
        minute_end += ch;
        count++;
    }
    minute_end = (minute_end/(49*count));

    /* extract characters until ',' to get second_end */
    count = 0;
    while((ch = getc(fp)) != 44)
    {
        second_end += ch;
        count++;
    }
    second_end = (second_end/(49*count));

и имена переменных были немного другими, и затем я понял, что был неправ, но это все не имеет значения.Я просто говорю это, потому что до того времени код работал просто отлично (результаты были неожиданными, хотя был какой-то мусор), но теперь он просто зависает и ничего не делает.Это почему?Большое спасибо за ваше время!

Вот файл, который я пробую: https://gist.github.com/gaurav712/6646ad7dfd3c487536dce9b0712471e7

Answer 1

Одна из ваших проблем, возможно, та, которую вы испытываете сейчас, заключается в том, что getc() не возвращает значение ASCII прочитанного символа.Он также может возвращать EOF, если больше нечего читать.

И так как ваши циклы заканчиваются, когда найден определенный символ (например, ":"), и EOF наиболее определенно не этот символв этих случаях ваша программа будет цикл навсегда .

Я бы предложил инкапсулировать эту логику в функцию чтения цифр:

/**
 * Reads a positive number (hopefully less than INT_MAX) from a stream
 * Returns -1 if the stream is at end-of-file
 * 
 * @param FILE *fp
 * @return int            the number read, or -1 if EOF
 */

int readDigits(FILE *fp) {
    int value = 0, c;
    if (feof(fp)) {
        return -1;
    }
    for (;;) {
        c = fgetc(fp);
        // EOF is not a digit so we catch it in the digit check
        // if (EOF == c) {
        //    break;
        // }
        if ((c < '0') || (c > '9')) {
            break;
        }
        value *= 10;
        value += (c - '0');
    }
    return value;
}

int seconds(int h, int m, int s) {
    return (h*60 + m) * 60 + s;
}

Теперь вы можете сделать:

hour_start = readDigits(fp);
if (hour_start < 0) {
    // Error, do something
}
min_start = readDigits(fp);
// check if -1
sec_start = readDigits(fp);
// check if -1

sec_start = seconds(hour_start, min_start, sec_start);

...

Answer 2

Я решил это:

Я переписал его с нуля с кучей комментариев

/* Third attempt to create a subtitle parser
 * 29 March, 2019
 * 12:55
 */

#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <ncurses.h>

/* to convert strings of format hh:mm:ss,m_s to seconds (ignoring milliseconds for now) */
int strToTime(char start[13], char end[13]);

int main(int argc, char **args){

    short count, shouldExit = 0;
    int duration, timeBwDialogues;
    char ch, startTimeStr[13], endTimeStr[13], prevTimeStr[13];
    FILE *fp;

    endTimeStr[0] = 0;

    /* Check if argument is present */
    if(args[1] == NULL){
        printf("No file specified!\n");
        exit(1);
    }

    /* Opening file for reading */
    printf("Trying to open file \"%s\"\n", args[1]);
    fp = fopen(args[1], "r");
    /* Checking if file was opened properly */
    if(fp == NULL){
        printf("Failed to open file \"%s\"!\n", args[1]);
        exit(2);
    }

    /* Checking if file has contents or not */
    if((ch = getc(fp)) == EOF){
        printf("File has no contents!\n");
        exit(3);
    }
    ungetc(ch, fp); /* putting ch back as file isn't empty */

    /* Section for setting up dialogue-number details
     * If I ever want to do that
     */

    /* initialising screen for nCurses */
    initscr();

    /* Here comes the actual stuff */
    while(1){
        /* As we aren't considering dialogue number, let's just skip that*/
        while((getc(fp)) != '\n');

        /* Reading dialogue's starting time :
         * It says keep reading until you find a character other that 0-9, ','
         * or ':' and store it to startTimeStr[]. Finally pass startTimeStr[]
         * to strToTime() to convert it to seconds(for now)
         */
        count = 0; /* Setting-up counter for the loop */
        while(1){
            ch = getc(fp);
            if(ch == ' ' || ch == '-')
                break;

            startTimeStr[count] = ch;
            count++;
        }
        startTimeStr[count] = '\0'; /* to terminate the string */

        /* Making a copy of endTimeStr[] in prevTimeStr[] to get timeBwDialogues */
        if(endTimeStr[0]){
            strcpy(prevTimeStr, endTimeStr);

            /* Calculating timeBwDialogues */
            timeBwDialogues = strToTime(prevTimeStr, startTimeStr);
        } else
            timeBwDialogues = strToTime("00:00:00,000", startTimeStr);

        /* For better readability */
        clear();
        refresh();
        /* Sleeping when there's no voice for synchronisation */
        sleep(timeBwDialogues);

        /* Placing the pointer to right position for reading ending time.
         * Using do-while to make sure at least one character is read before checking the condition
         */
        while((getc(fp)) != '>');
        if((ch = getc(fp)) == ' ');
        else
            ungetc(ch, fp);

        /* Just similar to what was done above, reading ending time */
        count = 0; /* Setting-up counter for the loop */
        while(1){
            ch = getc(fp);
            if(ch == '\n' || ch == ' ')
                break;

            endTimeStr[count] = ch;
            count++;
        }
        endTimeStr[count] = '\0'; /* to terminate the string */

        /* Calculating duration for individual dialogues */
        duration = strToTime(startTimeStr, endTimeStr); /* passing startTimeStr[] to strToTime */

        /* displaying the dialogue */
        while(1){
            ch = getc(fp);

            /* If ch is newline, the next character maybe EOF. So let's check */
            if(ch == EOF){
                shouldExit = 1;
                break;
            } else if(ch == '<'){
                while((ch = getc(fp)) != '>');
                continue;
            } else if(ch == '\n'){
                if((ch = getc(fp)) == EOF){
                    shouldExit  = 1;
                    break;
                } else if(ch == '\n') /* if the next character is newline, it's the end of the dialogue */
                    break;
                else{
                    ungetc(ch, fp);
                    addch('\n');
                    continue;
                }
            }

            /* As the next character to ch is not EOF, dialogue still exists(a dialogue might take multiple lines)
             * and it should be put on the screen
             */
            addch(ch);
        }
        refresh();
        sleep(duration);
        if(shouldExit)
            break;
    }
    /* Closing nCurses' window */
    endwin();

    /* Closing the file */
    fclose(fp);

    return 0;
}

/* Defining the function */
int strToTime(char start[13], char end[13]){
    int hour_start, hour_end, minute_start, minute_end, second_start, second_end;

    /* Separating hh, mm and ss for starting time. As I said above, I'll ignore milliseconds */

    /* getting hour_start */
    hour_start = ((start[0] - '0')*10)+(start[1] - '0');
    /* getting minute_start */
    minute_start = ((start[3] - '0')*10)+(start[4] - '0');
    /* getting second_start */
    second_start = ((start[6] - '0')*10)+(start[7] - '0');

    /* Separating hh, mm and ss for ending time. As I said above, I'll ignore milliseconds */

    /* getting hour_end */
    hour_end = ((end[0] - '0')*10)+(end[1] - '0');
    /* getting minute_end */
    minute_end = ((end[3] - '0')*10)+(end[4] - '0');
    /* getting second_end */
    second_end = ((end[6] - '0')*10)+(end[7] - '0');

    return ( ( ( ( (hour_end - hour_start) * 60) + (minute_end - minute_start) ) * 60) + (second_end - second_start) );
}