Question

I program a program to split file in C in Ubuntu. I have error when get buffer in readfile.

Here is my code:

    int split(char *filename, unsigned long part) {
        FILE *fp;
        char *buffer;
        size_t result; // bytes read
        off_t fileSize;
    
        fp = fopen(filename, "rb");
        if (fp == NULL) {
            fprintf(stderr, "Cannot Open %s", filename);
            exit(2);
        }
    // Get Size
        fileSize = get_file_size(filename);
    // Buffer
        buffer = (char*) malloc(sizeof(char) * (fileSize + 1));
        if (buffer == NULL) {
            fputs("Memory error", stderr);
            fclose(fp);
            return 1;
        }
    // Copy file into buffer
    //char buffers[11];
        result = fread(buffer, 1, fileSize, fp);
        buffer[fileSize] = '\0';
    
        if (result != fileSize) {
            fputs("Reading error", stderr);
            return 1;
        }
    
    // Split file
        off_t partSize = fileSize / part;
    // Last Part
        off_t lastPartSize = fileSize - partSize * part;
        unsigned long i;
        unsigned long j;
        // create part 1 to n-1
        for (j = 0; j < part; j++) {
            char partName[255];
            char *content;
            char partNumber[3];
            // Content of file part
    //      for (i = j; i < partSize * (j + 1); i++) {
    //
    //      }
            content = (char*) malloc(sizeof(char) * partSize);
            content = copychar(buffer, j + i, partSize + i);
            i += partSize;
            //copy name
            strcpy(partName, filename);
            // part Number
            sprintf(partNumber, "%d", j);
            // file name with .part1 2 3 4 ....
            strcat(partName, ".part");
            strcat(partName, partNumber);
            // Write to file
            writeFile(partName, content);
            free(content);
        }
    // last part
    char *content;
    content = (char*) malloc(sizeof(char) * (fileSize - partSize * (part - 1)));
    content = copychar(buffer, (part - 1) * partSize + 1, fileSize);
    char lastPartNumber[3];
    char lastPartName[255];
    sprintf(lastPartNumber, "%d", part);
    strcpy(lastPartName, filename);
    strcat(lastPartName, ".part");
    strcat(lastPartName, lastPartNumber);
    writeFile(lastPartName, content);
    free(content);

    free(buffer);
    fclose(fp);
    return 0;    
}

Here is function copychar from start to end:

char *copychar(char* buffer, unsigned long start, unsigned long end) {
    if (start >= end)
        return NULL;
    char *result;
    result = (char*) malloc(sizeof(char) * (end - start) + 1);
    unsigned long i;
    for (i = start; i <= end; i++)
        result[i] = buffer[i];
    result[end] = '\0';
    return result;
}

Here is function to get filesize:

off_t get_file_size(char *filename) {
    struct stat st;
    if (stat(filename, &st) == 0)
        return st.st_size;
    fprintf(stderr, "Cannot determine size of %s: %s\n", filename);
    return -1;
}

Here is function to write file:

    int writeFile(char* filename, char*buffer) {
    if (buffer == NULL || filename == NULL)
        return 1;
    FILE *file;
    file = fopen(filename, "wb");
    fwrite(buffer, sizeof(char), sizeof(buffer) + 1, file);
    fclose(file);
    return 0;
}

When I test I use file test 29MB and it dumped. I debug and it returned fileSize true but when readfile in buffer get from file it only return 135 characters and when use copychar it error.

    Breakpoint 1, 0x0000000000400a0b in copychar (buffer=0x7ffff5e3a010 "!<arch>\ndebian-binary   1342169369  0     0     100644  4         `\n2.0\ncontrol.tar.gz  1342169369  0     0     100644  4557      `\n\037\213\b", start=4154703576, end=4164450461) at final.c:43

Program received signal SIGSEGV, Segmentation fault.
0x0000000000400a0b in copychar (buffer=0x7ffff5e3a010 "!<arch>\ndebian-binary   1342169369  0     0     100644  4         `\n2.0\ncontrol.tar.gz  1342169369  0     0     100644  4557      `\n\037\213\b", start=4154703576, end=4164450461) at final.c:43

Program terminated with signal SIGSEGV, Segmentation fault.
The program no longer exists.

How can I divide buffer into part to write into part when split?

Answer 1

It’s highly impractical to copy files in 1 big block as you may have noticed. And it’s not needed.

At the simplest level you could copy the file byte by byte, like this

while( ( ch = fgetc(source) ) != EOF ) {
   fputc(ch, target);
}

Which will work, but it will be quite slow. Better to copy in blocks, like this:

 unsigned char buf[4096];
 size_t size;
 while( (size = fread(buf, 1, sizeof(buf), fpRead) ) > 0) {
     fwrite(buf, 1, size, fpWrite);
 }

Notice that the resulting code is way simpler and contains no dynamic memory allocation.

You still need to add the splitting logic of course, but that can be done by tracking the number of bytes written and opening a new write-file before actually writing it.

EDIT: how to handle the multipart facet – schematically, you still need to implement extra checks for some special cases and test results of the different system calls of course

 unsigned char buf[4096];
 size_t size;
 size_t partsize = 100000; // asssuming you want to write 100k parts.
 size_t stilltobewritten = partsize; // bytes remaining to be written in current part
 size_t chunksize = sizeof(buf); // first time around we read full buffersize
 while( (size = fread(buf, 1, chunksize, fpRead) ) > 0) {
     fwrite(buf, 1, size, fpWrite);
     stilltobewritten -= size; // subtract bytes written from saldo
     if (stilltobewritten == 0) {
         // part is complete, close this part and open next
         fclose(fpWrite);
         fpWrite = fopen(nextpart,"wb");
         // and reinit variables
         stilltobewritten = partsize;
         chunksize = sizeof(buf);
     } else {
         // prep next round on present file - just the special case of the last block
         // to handle
         chunksize = (stilltobewritten > sizeof(buf)) ? sizeof(buf) : stilltobewritten;
     }
 }

and EDIT 2: the file part name can be made a LOT simpler as well:

 sprintf(partName, "%s.part%d",file, j);