I'm right now looking for alternative ways to read text files on a line-by-line basis, for cross-platform compatibility (Windows / Linux).

My current solution is to go look for (or seek out) the newline (line break) in a text file using SDL2.

To demonstrate, I have a simple text file, with just 2 lines of gibberish text:

Hello world,
This is a new line with a line break.

On some text editors, you are able to display newlines or line breaks, and the above text would look somewhat like this:

Hello world,↵
This is a new line with a line break.

I want to use SDL_RWops, and if possible, use SDL_RWseek, or (SDL_RWops*)->seek(SDL_RWops*), to find the newline (line break), denoted by in the above example. That way, I can write a shorthanded version of SDL_RWreadToLine or some method that allows me to fetch text, line by line.

I believed the current method is very inefficient to check to see if I reached a newline, using psuedo-code:

//File reading/writing (We're only going to read)
//edict2 is a gigantic UTF-8 Japanese <-> English dictionary text file. 
//That text file is useful for text operations (read/write) testing purposes.

SDL_RWops* edict2 = SDL_RWFromFile("edict2u", "r");         //The "u" stands for UTF-8 compatible.
if (edict2) {
    Sint64 currentBytePointer = 0;              //Holds the current pointer.
    Sint64 currentReadPointer = 0;              //Holds the current read pointer.
    Sint64 fileSize = edict2->size(edict2);     //Grabs the file size of the file.
    char* buffer = new char[256];               //Some buffer to capture the read text characters.
    while (currentBytePointer < fileSize && currentReadPointer != 0) {
        Sint64 length = edict2->seek(edict2, 1, RW_SEEK_CUR);       //Reads 1 byte from current read position.
        Sint64 temp = edict2->read(edict2, buffer, length, length); //Implicit C conversion from void* to char*.
        if (buffer[0] != '\0' && buffer[0] == '\n') {
            //Found, do something about it
    delete buffer;
else {
    //File doesn't exist, do nothing.

I'm sure there are bugs that I may have missed out on, considering that I'm reading from a text file that may contain multiple bytes.

And that's why I wanted to look for SDL2 alternatives.

Does anyone know of a better way to read text line by line using SDL2?

  • \$\begingroup\$ Is there a specific reason you're looking for a SDL2 solution instead of using C++'s standard library? \$\endgroup\$ – Tyyppi_77 Dec 26 '17 at 0:20
  • \$\begingroup\$ @Tyyppi_77 Your comment made me realize that I could've just use std::ifstream and use std::getline() when the file is opened. Right now, I'm experimenting how to split the line string up into multiple substrings and tokens, so I can better manage the Japanese dictionary. I'll go ahead and make an answer about C++'s standard library solution. \$\endgroup\$ – tom_mai78101 Dec 26 '17 at 2:16

Instead of using SDL2's SDL_RWops, just use the C++ standard library.

A combination of std::ifstream and std::getline() should do the trick for reading strings from text files, line by line.

//File reading/writing (We're only going to read)
std::ifstream edict2("dict/edict2u");
if (edict2.is_open() && edict2.good()) {
    std::string buffer;
    while (std::getline(edict2, buffer)) {
        //Do something about the buffer.

        //Then clear the buffer once you're done with it.

Thanks to @Tyyppi_77 for pointing out the obvious.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.