fwscanf not doing the same thing as fscanf

here's the working code

Code:

char a[16], b[16];
while (fscanf(file, "%s", a) != EOF && fscanf(file, "%s", b) != EOF)
{
  // do stuff
}

here's the unicode version, which is pretty much exactly the same.

Code:

wchar_t a[16], b[16];
while (fwscanf(file, L"%s", a) != EOF && fwscanf(file, L"%s", b) != EOF)
{
  // do stuff
}

the text file i'm reading is very simple, something like this:
AAA A
BBB B
CCC C

the non-unicode will wrap around the parse the next line. the unicode version loops indefinitely the first line.

what am i doing wrong? thanks.

Not sure why it would not work right in fwscanf(), but in fwscanf(), if you use a wchar_t a[...] as input, you probably want "%ls" or "%S" to indicate that you have a wchar_t string, rather than char string.

At least that's what it says here:
http://www.manpagez.com/man/3/wscanf/

--
Mats

thanks for the quick reply. yeah i'm not sure why it didn't work either. i'm guessing since fwscanf is a microsoft implementation it might not be 100% the same as they claim...

changing the comparison to > 0 (instead of != EOF), works.

You do realize, I guess, that you could do all of your f(w)scanf in one line, something like this:

Code:

while (fscanf(file, "%s%s", a, b) == 2)
...

Where 2 is the number of "%" format specifiers. If it is not 2, then that particular line is malformed, assuming you expect ALL lines to have two things on them.

--
Mats

thanks
i realized it when i was going back to my code after posting and thinking...wait a minute...!!!

The *scanf() functions return the number of items read. You shouldn't compare the return value with EOF.

--
Computer Programming: An Introduction for the Scientifically Inclined

Originally Posted by Sander

The *scanf() functions return the number of items read. You shouldn't compare the return value with EOF.

The *scanf() functions do ALSO return EOF when they reach end of file - I do agree that checking for EOF on it's own is insufficient to ensure that correct input was received under all conditions, it is perhaps acceptable to check for EOF only - particularly if the input is character/string data where the content of the file is acceptable whatever it is. If the input is numeric, then it's critical to check the return value (and correctly handle the situation) to avoid a lock-up when the input is invalid.

--
Mats