Re: binary file and vc
Ludo wrote:
I'm working on an open source project and I have a problem with a
binary file.
All files on a computer are binary, the question is how the content is
formatted or what each bit or byte means. This might seem nitpicky, but see
below...
When I read the binary file with the program downloaded
from sourceforge, I have no problem. However, I'm unable to read this
binary file with the program compiled with Microsoft Visual C++
Express Edition SP1. The program read wrong values from the file and
crash. I don't understand because:
*the program downloaded from sourceforge was compiled with Microsoft
Visual C++ (in 2004)
That will probably be VC7.1 then, or maybe even an older version.
*it is not a problem of big/little endian.
You mean it's not a problem of the format of the file vs the format that the
program expects? Endianess is just one of the issues concerned here and I'm
pretty confident that indeed it is that the fileformats differ.
*C I/O functions and C++ I/O class read the same wrong values from
the files.
Wait: there are two way to do IO in C/C++. Using fread/fwrite does
unformatted IO, i.e. it transfers bytes from the memory to the file just
like they are (barring CR/LF translation on win32 though). Using
fprintf/fscanf or C++ IOStreams uses formatted IO though and both generally
expect a textual format in the file. That and the nit-pick above mean that
maybe your testing method and/or understanding of the issue is flawed
(well, you wouldn't ask here if you did understand everything completely,
but let's try to tackle the problem at hand...).
Anyhow, there are two things I would try:
1. If you compile with a recent version, can you save and reload one of the
files? If so, indeed the format used by the two programs differ. This of
course can have several reasons.
2. It could be that that the problem is locale-dependant. The locale affects
e.g. how printf/ostream formats numbers. Example: 1.000,95 (German) and
1,000.95 (USA) are different culture-specific representations of the same
number. It might already work if you set your OS-language to e.g.
English/USA.
With the second case, the problem is that the newer version of VC handles
locales differently, in the former versions in defaulted to the "C" locale,
which is supposed to be more or less neutral. It might also work if you
added a call to 'setlocale("C");' at the beginning of the program or if you
set the locale explicitly on every stream used for IO.
hope that helps...
Uli