Compiling microsoft code. Where to get fopen_s?

[tkzv]

I am trying to compile a console program I downloaded from MSDN: http://xbox.create.msdn.com/en-US/sample/xnb_format

There seem to be no references to any external libraries beside a single header file:

Code: Select all

#pragma once

#include <stdint.h>
#include <stdio.h>
#include <stdarg.h>
#include <assert.h>

#include <algorithm>
#include <exception>
#include <type_traits>
#include <string>
#include <vector>

using namespace std;

What libraries and headers do I need to specify?

GCC recognizes the code as C++11 and suggests the switch -std=c++11 . When I set it, it still doesn't find the declaration for fopen_s. Other sites suggest specifying -std=c11, but GCC allows it only for C or ObjC.

What do I need to build it?

Alternatively, how else can I unpack *.XNB files from XNA Game Studio?

sys-devel/gcc versions 4.9.3 and 5.3.0.

[Akkara]

First time I've heard of that one. Seems to be a Microsoftism. This link and the followup comments seem to explain the situation clearly: http://stackoverflow.com/questions/2575 ... g-to-files

It doesn't seem it would be too difficult to define your own as a static inline that calls fopen, checks for errors, and returns the results expected of fopen_s. Put it in the .h file and compile away. I'd offer to take a stab at providing a sample, but instead of risking advising in error, I'll wait see if steveL visits and offers his signature POSIX-ly correct advice right off the top of his head

[tkzv]

First time I've heard of that one. Seems to be a Microsoftism. This link and the followup comments seem to explain the situation clearly: http://stackoverflow.com/questions/2575 ... g-to-files

It doesn't seem it would be too difficult to define your own as a static inline that calls fopen, checks for errors, and returns the results expected of fopen_s. Put it in the .h file and compile away. I'd offer to take a stab at providing a sample, but instead of risking advising in error, I'll wait see if steveL visits and offers his signature POSIX-ly correct advice right off the top of his head

If you think stackoverflow.com is a reliable source of information, here's another page: http://stackoverflow.com/questions/1513 ... e-about-it

1. fopen_s is defined in the 2011 standard:
http://en.cppreference.com/w/c/io/fopen
http://www.open-std.org/jtc1/sc22/wg14/ ... /n1570.pdf
And it is available in GCC for *.c files.

2. There are examples of simple hacks that use macros to give fopen_s, but kind of defeat its purpose being as unsafe as fopen.

What I'm asking about, is the way to do it right. Maybe define some GCC-specific variable. Maybe use some MinGW header file. Maybe rebuild some library. There will be other troubles like this one, I'm searching for the universal solution.

[Genone]

Quick research doesn't give any hints that gcc (or rather: glibc) implements fopen_s at all (it is only an optional part of C11). And assuming they do, the spec says:

As all bounds-checked functions, fopen_s is only guaranteed to be available if __STDC_LIB_EXT1__ is defined by the implementation and if the user defines __STDC_WANT_LIB_EXT1__ to the integer constant 1 before including <stdio.h>.

MS has a long history of making their own special _s versions of standard (string) functions for "security" reasons, so this is not exactly a new problem.

EDIT: You'll have to live with one of the usual hacks: https://gcc.gnu.org/wiki/C11Status / https://sourceware.org/ml/libc-alpha/20 ... 00297.html

[tkzv]

EDIT: You'll have to live with one of the usual hacks: https://gcc.gnu.org/wiki/C11Status / https://sourceware.org/ml/libc-alpha/20 ... 00297.html

Thanks.

In other words, instead of compiling readily available Microsoft code, I'm better off rewriting it in a language I actually know.