Thread: Strange problem with notepad

Hey there.

One of my programs prints prime numbers to a file, and everything compiles and runs as expected. It's written in C.

However, when I open it in notepad, and ONLY notepad, I simply see the following:

″‵‷ㄱㄠ″㜱ㄠ‹㌲㈠‹ㄳ㌠‷ㄴ㐠″㜴㔠″㤵㘠‱㜶

And so on.

Yes, I know what you're thinking. "You're printing the ASCII equivalent of the prime numbers as opposed to the numbrers."

No, that was my first thought, too.

"Then you must be ........ing something up when opening the file. Have you tried using the binary and text flags of fopen() ?"

No to ........ing up and yes to flags.

Anyway, what pretty much proves that I'm doing it correctly is that the file is viewed correctly in wordpad. Now, I understand this might not necessarily be a programming problem, but I couldn't really find an appropriate forum for it.

Any ideas?

All suggestions greatly appreciated.

Originally Posted by Salem

Well without seeing the code, I'm guessing you're using one of the more recent compilers from Microsoft, and that you've managed to create a UNICODE text file.

Wordpad will understand this, and notepad (being simpler) does not.

Yes, this is what I thought too, as the chinese(Asian?) letters aren't a part of the ASCII set, but no, I'm simply using the Dev-C++ IDE which uses MinGW.

Originally Posted by Salem

So post a short program which demonstrates the problem so other people can try it to see if they observe the same thing.

I can put as little as "fprintf(out, "%d ", number);" and it get the same results.

Originally Posted by MacGyver

How about writing an entire program out as requested?

Then after that, check the flags your IDE has set for when it compiles and links.

This is the code in question. I have however narrowed it to a problem with notepad, using the > operator in cmd to redirect stdout to primes.txt gives the same results.

I also removed optimization from the compiling process - No change.

Implementation of Sieve of Eratosthenes.

Code:

#include <stdio.h>
#include <math.h>

int main(int argc, char **argv)
{
    
    unsigned int limit = 100000000, *digits = malloc(sizeof(unsigned int)*limit), idx, idx1 = 1, max = sqrt(limit), store = 0;
    
   
    fprintf( stderr, "Generating primes.. ");
    
    for(idx = 0; idx < limit; idx++)
            digits[idx] = idx1+=2;
            
    for(idx = 0; idx < max; idx++)
    {
            if(digits[idx] == 0) continue;
            
            store = digits[idx];
            for(idx1 = idx; idx1 < limit; idx1 += store)
            {
                     if(digits[idx1] == store) continue;
                     
                     digits[idx1] = 0;
            }
                     
    }
    
    
    fprintf( stderr, "Done!\n");
    fprintf( stderr, "Writing primes to file.. ");
    
    
    FILE *out = fopen("primes.txt", "w+");
    

    
    for(idx = 0; idx < 1000; idx++) {
            
            if(digits[idx] != 0) { fprintf( stdout,  "&#37;d ", digits[idx]); }
            
    }
         
    fclose(out);
    
    fprintf( stderr, "Done!\n");
    
    
    
    
    free(digits);
            
            
    
    
}

Originally Posted by iMalc

This is nothing to do with your code. Notepad is buggy in how it determines whether a file is unicode or not and in this case has gotten it wrong.

I came across this myself one day by opening Notepad and typing in an innocent short English sentence, closed and reopened the file, and got this problem. It is reported in various places all over the net. I think one of the more well known examples is something like "Bush hid the facts", which comes up in chinese.

Notepad ultimately sux!

Yes, this I had discovered already, as you can see in my previous post.

Anyway, is there any way to prevent this? This has never happened before, and tbh, changing file extension associations to make wordpad default is an annoying thing to do when I only need a text application to view simple, unformatted text.

or use a hex editor and you can see all the characters (or their hex values) even if Notepad doesn't show them. Then you can see pretty fast whether it's in Unicode or not.

Originally Posted by @nthony

I wouldn't bash Notepad just yet guys, I recently used it to scan through 75 MB of binary and text, both in unicode and ASCII (and some UTF-8) and it perfomed pretty well.
What I had to do when encountering parts that were Unicode, etc was:
1) Open notepad.exe,
2) File->Open
3) Set File type to All Files and Encoding to the encoding that you expect the file to be in.
It should work fine now.

Ah, thank you. It seems it was defaulting to Unicode encoding when I opened the file(Using "notepad textfile" from command line), but when I opened it with UTF-8, it shows what it's supposed to show.

thanks =)

If there is no BOM, Notepad tries to figure the encoding of the text, it does that by creating a histogram of the text and comparing it to histograms of different encodings. not a stupid program after all.
because the file you created is short and not typical text it is type casted with the wrong encoding.