Unfortunately, just as with the 64-bit installation above, the hardest part is
configuring Windows to use the compiler. If you are having problems,
it is likely because you haven't set up things properly after the installation.
The rest of this guide assumes that you installed the compiler to c:\mingw32
during the installation.
Currently, when you type g++ from the command line, you will run the 64-bit
version of the MinGW g++ compiler. You want to continue to do that. To run the 32-bit
version, you will need to create a batch file called g++32.bat and put it
in your PATH somewhere. (You should have created a C:\utils directory
on your computer to put your add-on programs. This directory should be the first
directory in your PATH.) The contents of the batch
file should look like this:
::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
: This batch file only handles 20 options for the compiler. This should
: be sufficient. You can figure out how to add more, if necessary.
::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
@echo off
set /a count=1
:BEGIN
if "%1"=="" goto END
set arg%count%=%1
shift
set /a count=%count%+1
goto BEGIN
:END
c:\mingw32\mingw32\bin\g++.exe %arg1% %arg2% %arg3% %arg4% %arg5% %arg6% %arg7% %arg8% ^
%arg9% %arg10% %arg11% %arg12% %arg13% %arg14% %arg15% %arg16% ^
%arg17% %arg18% %arg19% %arg20%
set arg1=
set arg2=
set arg3=
set arg4=
set arg5=
set arg6=
set arg7=
set arg8=
set arg9=
set arg10=
set arg11=
set arg12=
set arg13=
set arg14=
set arg15=
set arg16=
set arg17=
set arg18=
set arg19=
set arg20=
You can create the batch file with any text editor (e.g. notepad++) or, you can download
the batch file here: g++32.bat (recommended). You need to download this file, not click
on it and run it (the default in some versions of Windows). There is also a batch file
for gcc here: gcc32.bat. Use this to compile 32-bit C programs.
Once you've downloaded the batch to (or created it in) some directory in your PATH, you
can check that it's correct by typing this from any directory:
g++32 --version
If everything was setup properly, you should see this message in the console:
g++.exe (i686-posix-dwarf-rev1, Built by MinGW-W64 project) 7.1.0
Copyright (C) 2017 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
Now, when building 32-bit programs, you type g++32 instead of g++.
For example, to build a 64-bit program (assume all of the usual command line options)
g++ file1.cpp file2.cpp file3.cpp -o prog64.exe
To build a 32-bit program:
g++32 file1.cpp file2.cpp file3.cpp -o prog32.exe
IMPORTANT NOTE #1:
When running the executable program, if you get an error saying something like:
"The program can't start because <some_dll_file>" is missing from your
computer..."
or
"The application was unable to start correctly (0xc000007b)..."
then you will need to put the bin directory of the 32-bit MinGW
compiler at the end of your PATH. If you installed it to the location
recommended in these notes, that directory is
C:\mingw32\mingw32\bin.
The name of the missing DLL file may
be something like libwinpthread-1.dll, or libgcc_s_dw2-1.dll,
or some other file.
Note: If you put the bin directory somewhere other than
the end of your PATH, it may interfere with your 64-bit compiler causing it
to no longer work.
IMPORTANT NOTE #2:
When running the executable program, if you get an error message that looks
like this:
Then you need to statically link the executable. Add this option (in bold) to the command
line:
g++32 file1.cpp file2.cpp file3.cpp -o prog32.exe -static
This will make the executable much bigger, but it should avoid the error.
This is also trivial to override the compiler in your makefile. If your
makefile has a macro like this (which all of your makefiles should have)
CC=g++
You can simply run make like this:
make CC=g++32.bat
and it will build a 32-bit executable instead of a 64-bit executable.
Note that you must provide the .bat file extension for the makefile.
To verify that your program is really a 32-bit executable, use the
file
command (part of the Cygwin utilities). For example, this command:
file prog32.exe prog64.exe
will display this in the console:
prog32.exe: PE32 executable (console) Intel 80386, for MS Windows
prog64.exe: PE32+ executable (console) x86-64, for MS Windows
PE32 is a 32-bit (Intel 80386) executable.
PE32+ is a 64-bit (x86-64) executable.
Note to Linux and Mac users: Building a 32-bit program under Mac or
Linux is trivial. You simply add a -m32 option to the command line. For example:
g++ -Wall -Wextra -ansi -pedantic -m32 foo.c
That's it!
Incidentally, to create a 64-bit executable, you can use -m64, but since
this is the default, you don't have to provide it to the compiler. If you run
the file command on Linux, you would see something like this
(with possibly more information displayed)
prog32: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV)
prog64: ELF 64-bit LSB executable, x86-64, version 1 (SYSV)
And on macOS you would see this:
prog32: Mach-O executable i386
prog64: Mach-O 64-bit executable x86_64
