mingw-users Mailing List for MinGW - Minimalist GNU for Windows

Erwin Waterlander wrote:
> Or dowload vim from www.vim.org
> (ftp://ftp.vim.org/pub/vim/pc/gvim72.exe). Install it, add "/c/Program
> Files/Vim/vim72" to PATH, and then you have a full-fledged 'vim' and
> 'gvim', which is also usable outside MSYS.
>    

My choice.

I then add an alias to my ~/.profile like so:
alias vim='start /path/to/vim/vim72/gvim'

This allows the MSYS console to become free.

Earnie

re-re-send to mingw list after filtering:

TO: Rowan Sylvester-Bradley (rowan at last-name.org)

> I'm trying to get AutoGen working under Windows Vista Business. I'm hampered
> by the fact that I know almost nothing about Unix/Linux. I've downloaded and
> installed Cygwin v1.7.1. It appears to run, although I don't know enough to
> test it properly. Are there some test scripts or commands that I should try
> to check that it's working properly?
> 
> It also seems to have Guile installed - at least, I can type "guile" and get
> a "guile>" prompt. I know nothing at all about Guile, so I don't know how to
> test this either. Are there some scripts I should run or commands to try
> that would check it out?
> 
> To install AutoGen, what file(s) do I have to download, and what do I have
> to do with them?

Hi Rowan,

You have to have libguile installed (you do) with the development headers
as well (you may not).  I am forwarding this to the Minimalist GNU for Windows
list.  They have something that works on Windows.  Unfortunately, Windows
does not properly support fork() and guile uses some dynamic linking
options that make its library problematic on Windows.  Both of these issues
make Windows support challenging.  I don't use Windows myself.  So, hopefully,
someone on the MinGW list can help point you in the right direction.  Sorry
I could not be more helpful.

Regards, Bruce

You can get pkg-config prebuilt here:
http://ftp.gnome.org/pub/GNOME/binaries/win32/dependencies/pkg-config_0.23-3_win32.zip

Also  get http://ftp.gnome.org/pub/GNOME/binaries/win32/dependencies/pkg-config-dev_0.23-3_win32.zip
if you need to re-autogen something that needs pkg.m4.

pkg-config.exe depends on libglib-2.0-0.dll, in
http://ftp.gnome.org/pub/gnome/binaries/win32/glib/2.22/glib_2.22.4-1_win32.zip
.

--tml

Hallo,

I get the following warning if I link my app against the
htmllib.lib library using MinGW gcc (3.4.5/4.4.0)
and WinAPI 3.14

---snip---
Warning: .drectve `-defaultlib:uuid.lib ' unrecognized
Warning: .drectve `-defaultlib:uuid.lib ' unrecognized
---snap---

What is going wrong and how can I fix this warning?

The problem ist, that my lecturer said "_without_ any
warning - or the work ist not valid".

Is there an alternate htmllib.lib with the same functions
 provided by the MinGW Project?

Thomas

[1] http://msdn.microsoft.com/en-us/library/ms669985(VS.85).aspx

On 26.02.2010 13:54, Philippe Berthault wrote:
> Hello,
>
> I've written a small C (not C++) GUI Windows library with the MinGW 
> compiler.
>
> This library is named "radGui", where "rad" is the abbreviation of 
> "Rapid Application Development".
>
> This library is a wrapper of the Win32 API and runs perfectly on Windows 
> XP and Vista.
>
> With radGui, I'm able to build very quickly simple Windows applications 
> without needing any resource editor.
>
> This library is free without any restriction. I can give it (full 
> sources and complete documentation) to anyone who is interested by radGui.
>
> Philippe Berthault.
>
>   
Can i take a look at the API?

Hi!

I would like to give the library intended for Unix libgphoto a try 
with mingw/msys.
When running configure, it asks for pkg-config.
I downloaded pkg-config Version 023 but I am unable to make it.
configure complains about missing
ac_nonexistent.h
dlfcn.h
alloca.h
sys/wait.h

I am not expert on this but I understand that dlfcn has something to 
do with posix.
So I tried to run configure under msys with LIBS=-lposix but then I 
receive the error message
C compiler cannot create executables.

I am using mingw 4.4.0 with the updated packages for both mingw and msys.
msysDTK etc. has been added and compiling e.g. project mathGL or FLTK 
went fine.

Does anybody know how to solve this and has anybody tried to compile 
gphoto for win32?

Thanks!
Uwe



___________________________
PD Dr. Uwe Rossow
TU Braunschweig
Inst. f. Angewandte Physik
Mendelssohnstr. 2
38106 Braunschweig
Fax: 0531-391 8511
Email: u.r...@tu...
Tel.: 0531-391 8523 (good luck!) 

Hello,

I've written a small C (not C++) GUI Windows library with the MinGW 
compiler.

This library is named "radGui", where "rad" is the abbreviation of 
"Rapid Application Development".

This library is a wrapper of the Win32 API and runs perfectly on Windows 
XP and Vista.

With radGui, I'm able to build very quickly simple Windows applications 
without needing any resource editor.

This library is free without any restriction. I can give it (full 
sources and complete documentation) to anyone who is interested by radGui.

Philippe Berthault.

> Where are you getting that these are thread safe?

http://support.microsoft.com/default.aspx?scid=kb;en-us;104641 for instance.

Also, by looking at the C runtime sources as distributed already with
Visual C 6.0 in 1998 one can verify it. Thay have hardly made it less
MT-safe since.

> The microsoft description in http://msdn.microsoft.com/en-us/library/aa246456%28VS.60%29.aspx
> states they use a single statically stored space for the return

That is probably a quite old page, as it talks about Win95 and Windows
NT but doesn't mention newer OSes. It also mentions the non-MT-safe C
libraries with nobody uses any more.

Also, it is trivial to verify with a test program:

#include <time.h>
#include <stdio.h>
#include <process.h>

#include <windows.h>

static unsigned __stdcall
threadfun (void *p)
{
  time_t t = time (NULL);

  struct tm *lt = localtime (&t);

  printf ("locatime returns %p\n", lt);
  fflush (stdout);

  Sleep (1000);

  return 0;
}

int
main (int argc, char **argv)
{
  int i;

  for (i = 0; i < 100; i++)
    _beginthreadex (NULL, 0, threadfun, NULL, 0, NULL);

  Sleep (2000);

  return 0;
}

--tml

2010/2/25 Joe <lj...@gm...>:
> Hi all together
>
> How to compile the following code:
>
> case INDEX_op_addl_A0_EAX: {
>    extern void op_addl_A0_EAX();
> extern char @taintcheck_fn2regs@16;
>    memcpy(gen_code_ptr, (void *)((char *)&op_addl_A0_EAX+0), 43);
>    *(uint32_t *)(gen_code_ptr + 33) = (long)(&@taintcheck_fn2regs@16) -
> (long)(gen_code_ptr + 33) + 0 -4;
>    gen_code_ptr += 43;
> }
>
> ./op.h:14470: error: stray '@' in program
> ./op.h:14470: error: stray '@' in program
> ./op.h:14470: error: syntax error before numeric constant
>
> $ uname -a
> MINGW32_NT-5.1 HANUELE-BC60720 1.0.10(0.46/3/2) 2004-03-15 07:17 i686
> unknown
>
> I tried to demangle the extern char with c++filt but the output is the
> same as the input.
>
> Cheers
>
> Stefan
>
> ------------------------------------------------------------------------------
> Download Intel&#174; Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> MinGW-users mailing list
> Min...@li...
>
> This list observes the Etiquette found at
> http://www.mingw.org/Mailing_Lists.
> We ask that you be polite and do the same.  Disregard for the list etiquette may cause your account to be moderated.
>
> _______________________________________________
> You may change your MinGW Account Options or unsubscribe at:
> https://lists.sourceforge.net/lists/listinfo/mingw-users
>

Hello,

you are using here invalid symbol names for C. The symbol
'@taintcheck_fn2regs@16' seems to be a function with '__fastcall'
calling convention. So you have to declare it as function prototype.
Anyway the '@' character isn't valid in C/C++ symbol-names. They are
reserved and used (for x86) for symbol decorations.

Regards,
Kai

-- 
|  (\_/) This is Bunny. Copy and paste
| (='.'=) Bunny into your signature to help
| (")_(") him gain world domination

Hi all together

How to compile the following code:

case INDEX_op_addl_A0_EAX: {
    extern void op_addl_A0_EAX();
extern char @taintcheck_fn2regs@16;
    memcpy(gen_code_ptr, (void *)((char *)&op_addl_A0_EAX+0), 43);
    *(uint32_t *)(gen_code_ptr + 33) = (long)(&@taintcheck_fn2regs@16) -
(long)(gen_code_ptr + 33) + 0 -4;
    gen_code_ptr += 43;
}

./op.h:14470: error: stray '@' in program
./op.h:14470: error: stray '@' in program
./op.h:14470: error: syntax error before numeric constant

$ uname -a
MINGW32_NT-5.1 HANUELE-BC60720 1.0.10(0.46/3/2) 2004-03-15 07:17 i686
unknown

I tried to demangle the extern char with c++filt but the output is the
same as the input.

Cheers

Stefan

With such a vague description and without specific minimal (but
complete) sample code to look at (and possibly compile and debug) it's
hard to say what exactly is the problem. My first guess is that what
causes your problem is different C libraries being used in different
modules (the program's exe and dlls loaded).

The C "file descriptors" (the small integers as returned and accepted
by functions like open(), write(), fileno()) are specific to the C
library used by one (or several) modules. (This obviously is very
unlike POSIX where file descriptors are part of the system call API
and per-process.) If one module uses msvcrt.dll and opens a file with
open() or fopen(), passing such a file descriptor or FILE pointer to
code in another module that uses msvcr90.dll will not work as
intended.

On the other hand, you talk about SetStdHandle() which of course is at
the Win32 level, not the C library, and HANDLes are Windows kernel
concepts and not related to any C library, so I am not so sure about
the first guess.

Another possibility is that you are trying to access Win32 standard
input and output handles in a program whose exe header marks it as a
"windowing" application, and which is started without any IO
redirection . The standard handles are then invalid. This was
discussed very recently on this list, check the archives. Hint:
AttachConsole().

--tml

I've got a project building a .dll file with a few library functions.
Those library functions write to stdout and stderr. I'm cross
compiling to Windows x86 and trying to access this output
from a C# application. However, standard methods for grabbing
the output stream (SetStdHandle, that works for other windows
.dlls) don't do anything for cross-compiled libraries. What do I
need to do to access the output stream?

> On 2010-02-24 13:30Z, Fab...@se... wrote:
> > 
> > However when I open a MSYS console using the desktop icon and type vim 
at 
> > the prompt, I get a 'sh: vim: command not found'. Is vim not supposed 
to 
> > be part of the MSYS package? Is it possible to install it at a later 
stage 
> > (using the binary version of the MinGW download page, not by 
recompiling 
> > it)?
> 
> Sure: download the binary tarball here:
>   http://sourceforge.net/projects/mingw/files/
> then extract it, e.g. with this command:
>   bsdtar -xkf vim-7.2-1-msys-1.0.11-bin.tar.lzma
> to /usr as the release notes suggest.

Thanks for the hint. I downloaded all four lzma files and installed it 
with:

   lzma -d *.lzma
   cd /
   for f in $HOME/*.lzma; do tar xvf $f; done

Fabien.


DISCLAIMER: 
This e-mail contains proprietary information some or all of which may be legally privileged. It is for the intended recipient only. If an addressing or transmission error has misdirected this e-mail, please notify the author by replying to this e-mail. If you are not the intended recipient you must not use, disclose, distribute, copy, print, or rely on this e-mail.

On Tue, 23 Feb 2010 22:57:30 +0200
 Tor Lillqvist <tm...@ik...> wrote:
> >> Run the program from cmd.exe and you will see a more helpful
> message.
> 
> > I did and the program worked correctly!
> 
> Oh? Even if it was called "patcher.exe"? I expected you to get a UAC
> (User Account Control) dialog in that case. ("Do you want to let this
> program from an unknown publisher make changes to this computer" or
> something like that.)

I just tried it and that's what happened.  When you answer yes/allow,
the program works correctly.


> That is what happens for me, if I have a program called
> "patcher.exe":
> When run from MSYS I just get "Bad file number", from cmd.exe the UAC
> dialog. Possibly some caching is involved, if you have been renaming
> files back and forth Windows might be confused by now;)
> 
> What a program *actually* does, what code it contains, is irrelevant
> for the triggering of the UAC dialog as far as I know, it is just the
> *name* that matters.

I didn't think the problem had anything to do with my code
(particularly, when a do-nothing program has the same problem).


> Note that if you have a program with some random
> "safe" name and it tries to do some "administrative" operations to
> the
> computer, you won't get this dialog when you run it, but if you run
> it
> without administrative privileges such operations won't succeed, of
> course.
> 
> I.e. my point was that the name "patcher" makes Windows assume it is
> a
> program that needs administrative privileges and thus Windows
> helpfully asks for permission to provide them to the program; for
> some
> reason this doesn't happen when the program is run from the MSYS
> shell. Anyway, the way around this "useful" feature is to provide a
> .manifest file for the program that tells Windows that no privilege
> elevation is necessary (i.e. even if the program name contains
> "patch"
> it doesn't actually do anything that would require administrative
> privileges).
> 
> The .manifest file should look like this:
> 
> <?xml version="1.0" encoding="UTF-8" standalone="yes"?>
> <assembly xmlns="urn:schemas-microsoft-com:asm.v1"
> manifestVersion="1.0">
>   <v3:trustInfo xmlns:v3="urn:schemas-microsoft-com:asm.v3">
>     <v3:security>
>       <v3:requestedPrivileges>
>         <v3:requestedExecutionLevel level="asInvoker" />
>       </v3:requestedPrivileges>
>     </v3:security>
>   </v3:trustInfo>
> </assembly>
> 
> and for patcher.exe, it should be called patcher.exe.manifest . I had
> to create such files for MSYS's install.exe and patch.exe back when I
> switched to Windows 7 from XP. (By now presumably MSYS presumably
> comes with manifests for these exes?) And also for a program called
> gtk-update-icon-info.exe for instance.
> 
> --tml

I never suspected the file name had anything to do with it.  If that's
the case, then it seems easiest just to rename the file to something
that doesn't contain "patch".

-- Glenn

Op 02/24/2010 04:56 PM, Greg Chicares schreef:
> On 2010-02-24 13:30Z, Fab...@se... wrote:
>   
>> However when I open a MSYS console using the desktop icon and type vim at 
>> the prompt, I get a 'sh: vim: command not found'. Is vim not supposed to 
>> be part of the MSYS package? Is it possible to install it at a later stage 
>> (using the binary version of the MinGW download page, not by recompiling 
>> it)?
>>     
>
> Sure: download the binary tarball here:
>   http://sourceforge.net/projects/mingw/files/
> then extract it, e.g. with this command:
>   bsdtar -xkf vim-7.2-1-msys-1.0.11-bin.tar.lzma
> to /usr as the release notes suggest.
>
>   

Or dowload vim from www.vim.org 
(ftp://ftp.vim.org/pub/vim/pc/gvim72.exe). Install it, add "/c/Program 
Files/Vim/vim72" to PATH, and then you have a full-fledged 'vim' and 
'gvim', which is also usable outside MSYS.

Erwin

Tor Lillqvist schrieb:
>
> Hmm, a quite strange thing to do, especially for somebody claiming to
> be new to C. Hopefully you aren't trying to do something naughty? I
> can't really think of any reason to use this function that wouldn't be
> annoying for the user, or even evil, especially if doing something
> without the user noticing.
>
>   
My goal is to write a C extension to the Ruby programming language that 
allows programmers
to access the input simulation functions from within Ruby (similar to 
the DLL interface of AutoIt),
but before I make a C extension I thought I have to get it work as C only.
I said, I am "quite new to the C world" which means at least for me that 
I had played around
with C before, but am now going into deep. After I mainly looked on Ruby 
C extensions before
and learned from what I saw there, I followed this website: 
http://publications.gbdirect.co.uk/c_book/ .
> Is this on Vista or newer? Read the documentation for the SendInput()
> function. It says: "This function is subject to UIPI. Applications are
> permitted to inject input only into applications that are at an equal
> or lesser integrity level." So presumably you happen to have the mouse
> on top of an application that is at a higher integrity level. If your
> mouse cursor is inside the cmd.exe window, apparently cmd.exe is that.
>
>   
It runs on Vista. And yeah, I think I should read the docs better... I 
always run MSYS as
admin, because for the Ruby C extensions I need the "make install" 
command which requires
beacuse of it's name administrative privileges, and therefore I made the 
MSYS startup script asking me
for administrative privileges every time I execute it. When I run CMD as 
admin, it works.
Thank you.
> It works for me also from cmd as long as the cursor is positioned on
> for instance this Firefox text input window where I am typing now when
> I start the program (by typing its name and pressing enter). (I.e. the
> mouse cursor is on the Firefox text input window, but input focus is
> in the MSYS console window, no text caret is visible in the Firefox
> text input window.) Then indeed what happens is that a mouse click at
> the cursor position is simulated, i.e. the Firefox window is activated
> and the blinking caret becomes visible at the mouse cursor location .
>
>   
If I don't run it as admin, it doesn't work.
> It works for me also if I start it from Explorer. Obviously just
> double-clicking the program name it will not really be easy to see
> whether it wors, as that essentially means just a third mouse click on
> it...
OK, that one was silly, I didn't thought about it. But even when using 
the enter technique, it
doesn't work.

Quintus

On 2010-02-24 13:30Z, Fab...@se... wrote:
> 
> However when I open a MSYS console using the desktop icon and type vim at 
> the prompt, I get a 'sh: vim: command not found'. Is vim not supposed to 
> be part of the MSYS package? Is it possible to install it at a later stage 
> (using the binary version of the MinGW download page, not by recompiling 
> it)?

Sure: download the binary tarball here:
  http://sourceforge.net/projects/mingw/files/
then extract it, e.g. with this command:
  bsdtar -xkf vim-7.2-1-msys-1.0.11-bin.tar.lzma
to /usr as the release notes suggest.

Test case:
I created a file Test.java as follows:
 public class Test
 {
  public static void main(String args[])
  {
    System.out.println("This is a test.");
  }
 }
and compiled it by running:
   gcj --main=Test -o test.exe Test.java
and file test.exe appeared but when I ran it, it produced no output -
  it just immediately quits.
I uploaded it to my site and so
you can get this file at http://www.gameofbeads.com/files/test.exe (3.4 Mb)

Additional tests:
  I made a simple program in test.cpp to do the same in cpp
#include <stdio.h>
int main(int argc,char *argv[])
{
  printf("This is a cpp test!");
  return 0;
}
and compiled it with gcc -o test.exe test.cpp and it produced output
   This is a cpp test!
So, the main part of mingw works correctly.
I also made a more complicated program in java:
import java.io.*;
public class c  {
  public static void main (String args[]) {
       try {
          FileOutputStream fout = new FileOutputStream("vov.txt");
          fout.write(45);
          fout.close();
          System.out.println("!!Completed!!");
       }
       catch (Exception e) {
         e.printStackTrace();
       }
  }
}
and compiled it: gcj --main=c -o c.exe c.java
and the compiler made c.exe (3.4 Mb)
when I ran it, it immediately quitted, and it neither created vov.txt
nor produced output.


My parameters:
Windows XP
GCC: Reading specs from E:/MinGW/bin/../lib/gcc/mingw32/3.4.5/specs
Configured with: ../gcc-3.4.5-20060117-3/configure --with-gcc --with-gnu-ld
--wi
th-gnu-as --host=mingw32 --target=mingw32 --prefix=/mingw --enable-threads
--dis
able-nls --enable-languages=c,c++,f77,ada,objc,java --disable-win32-registry
--d
isable-shared --enable-sjlj-exceptions --enable-libgcj --disable-java-awt
--with
out-x --enable-java-gc=boehm --disable-libgcj-debug --enable-interpreter
--enabl
e-hash-synchronization --enable-libstdcxx-debug
Thread model: win32
gcc version 3.4.5 (mingw-vista special r3)
BINUTILS: GNU ld (GNU Binutils) 2.20
MINGW: 5.1.6
Environment to run programs: FAR

> First off, I'm quite new to the C world, but I try to learn what I can
> get...So, here's my problem: I've a small C program that simulates some mouse
> input to Windows

Hmm, a quite strange thing to do, especially for somebody claiming to
be new to C. Hopefully you aren't trying to do something naughty? I
can't really think of any reason to use this function that wouldn't be
annoying for the user, or even evil, especially if doing something
without the user noticing.

> and if I run this from the MSYS shell it works quite well.

Is this on Vista or newer? Read the documentation for the SendInput()
function. It says: "This function is subject to UIPI. Applications are
permitted to inject input only into applications that are at an equal
or lesser integrity level." So presumably you happen to have the mouse
on top of an application that is at a higher integrity level. If your
mouse cursor is inside the cmd.exe window, apparently cmd.exe is that.

> Now, if I try to run it from CMD or by double-clicking on the executable
> file, nothing happens.

It works for me also from cmd as long as the cursor is positioned on
for instance this Firefox text input window where I am typing now when
I start the program (by typing its name and pressing enter). (I.e. the
mouse cursor is on the Firefox text input window, but input focus is
in the MSYS console window, no text caret is visible in the Firefox
text input window.) Then indeed what happens is that a mouse click at
the cursor position is simulated, i.e. the Firefox window is activated
and the blinking caret becomes visible at the mouse cursor location .

It works for me also if I start it from Explorer. Obviously just
double-clicking the program name it will not really be easy to see
whether it wors, as that essentially means just a third mouse click on
it... But if I put select the program in Explorer (so that it is
highlighed), then move the mouse cursor on some other file, and press
enter (i.e. run the highlighted program), a mouse click is simulated,
i.e. that other file becomes selected.

 --tml

On Wed, 2010-02-24 at 15:30 +0100, Quintus wrote:
> Hi there,
> 
> First off, I'm quite new to the C world, but I try to learn what I can 
> get...
> So, here's my problem: I've a small C program that simulates some mouse
> input to Windows and if I run this from the MSYS shell it works quite well.
> Now, if I try to run it from CMD or by double-clicking on the executable 
> file,
> nothing happens. I searched the web for that problem, but couldn't find any
> reasonable help, so now I'm posting here...
> 
> That's a small sample program:
> ----------------------------------------
> #define _WIN32_WINNT 0x0500
> #include <windows.h>
> 
> int main(int argc, char *argv[])
> {
>   INPUT inps[2];
>   MOUSEINPUT mis[2];
>   inps[0].type = INPUT_MOUSE;
>   inps[1].type = INPUT_MOUSE;
>  
>   mis[0].mouseData = 0;
>   mis[0].dwFlags = MOUSEEVENTF_LEFTDOWN;
>   mis[1].mouseData = 0;
>   mis[1].dwFlags = MOUSEEVENTF_LEFTUP;
>  
>   inps[0].mi = mis[0];
>   inps[1].mi = mis[1];
>  
>   SendInput(2, inps, sizeof(INPUT));
>   return 0;
> }
> 
> ----------------------------------------
> I'm running on Windows Vista and didn't pass any compilation
> options to gcc, I just did "gcc abc.c" and than ran "a.exe".
> 
> Can anyone explain this to me?
> 

My best guess is that is *some* cases windows has a stale mouse move
event... Your program moves the pointer once, but if windows has another
mouse move in the queue then you wont even see yours.

Try putting the move in a loop, or add a time delay before exit.

Jon

On 24.02.2010 16:30, Fab...@se... wrote:
> Hello,
>
> I am a fresh user freshly installing MinGW-5.1.6.exe + MSYS-1.0.11.exe + 
> msysDTK-1.0.1.exe + msysCORE-1.0.11-bin.tar.gz on my Windows XP without 
> any error or complaint from the installers. MinGW is installed in C:\MinGW 
> and I answered C:/MinGW to the MSYS installer when prompted. 
>
> However when I open a MSYS console using the desktop icon and type vim at 
> the prompt, I get a 'sh: vim: command not found'. Is vim not supposed to 
> be part of the MSYS package? Is it possible to install it at a later stage 
> (using the binary version of the MinGW download page, not by recompiling 
> it)?
>
> Fabien.
>
>   
MSysGit comes with vim, you might want to check it out.
> DISCLAIMER: 
> This e-mail contains proprietary information some or all of which may be legally privileged. It is for the intended recipient only. If an addressing or transmission error has misdirected this e-mail, please notify the author by replying to this e-mail. If you are not the intended recipient you must not use, disclose, distribute, copy, print, or rely on this e-mail.
>   
DISCLAIMER:
This e-mail contains no proprietary information and there is no legally
privileged information in it. It is for intended recipient(s), and for
everyone who cares to read it. If an addressing or transmission error
has misdirected this e-mail, please don't do anything - just forget
about it. If you are not the intended recipient you can do whatever you
want - i don't care.

Hi there,

First off, I'm quite new to the C world, but I try to learn what I can 
get...
So, here's my problem: I've a small C program that simulates some mouse
input to Windows and if I run this from the MSYS shell it works quite well.
Now, if I try to run it from CMD or by double-clicking on the executable 
file,
nothing happens. I searched the web for that problem, but couldn't find any
reasonable help, so now I'm posting here...

That's a small sample program:
----------------------------------------
#define _WIN32_WINNT 0x0500
#include <windows.h>

int main(int argc, char *argv[])
{
  INPUT inps[2];
  MOUSEINPUT mis[2];
  inps[0].type = INPUT_MOUSE;
  inps[1].type = INPUT_MOUSE;
 
  mis[0].mouseData = 0;
  mis[0].dwFlags = MOUSEEVENTF_LEFTDOWN;
  mis[1].mouseData = 0;
  mis[1].dwFlags = MOUSEEVENTF_LEFTUP;
 
  inps[0].mi = mis[0];
  inps[1].mi = mis[1];
 
  SendInput(2, inps, sizeof(INPUT));
  return 0;
}

----------------------------------------
I'm running on Windows Vista and didn't pass any compilation
options to gcc, I just did "gcc abc.c" and than ran "a.exe".

Can anyone explain this to me?

Quintus

Hello,

I am a fresh user freshly installing MinGW-5.1.6.exe + MSYS-1.0.11.exe + 
msysDTK-1.0.1.exe + msysCORE-1.0.11-bin.tar.gz on my Windows XP without 
any error or complaint from the installers. MinGW is installed in C:\MinGW 
and I answered C:/MinGW to the MSYS installer when prompted. 

However when I open a MSYS console using the desktop icon and type vim at 
the prompt, I get a 'sh: vim: command not found'. Is vim not supposed to 
be part of the MSYS package? Is it possible to install it at a later stage 
(using the binary version of the MinGW download page, not by recompiling 
it)?

Fabien.


DISCLAIMER: 
This e-mail contains proprietary information some or all of which may be legally privileged. It is for the intended recipient only. If an addressing or transmission error has misdirected this e-mail, please notify the author by replying to this e-mail. If you are not the intended recipient you must not use, disclose, distribute, copy, print, or rely on this e-mail.

--- El mié, 24/2/10, Greg Chicares <gch...@sb...> escribió:

De: Greg Chicares <gch...@sb...>
Asunto: Re: [Mingw-users] GMAKE: Unexpected 'test' at this moment
Para: "MinGW Users List" <min...@li...>
Fecha: miércoles, 24 de febrero, 2010 00:09

On 2010-02-23 20:13Z, Sebastian Ledesma wrote:
> 
> I'm updating gc.mak file for MingW for the OWLNext project. 
> I want to have the same .mak  for both Linux and Windows.
> I have a section to check dirs and create them if necessary.
> It works in Linux, but fails in Windows when using GMAKE.
> 
> The section is:
>
> CreateDirs:
>     if ! test -e $(TARGETDIR);then mkdir $(TARGETDIR);fi
>     if ! test -e $(OBJROOT);then mkdir $(OBJROOT);fi
>     if ! test -e $(OBJDIR);then > mkdir $(OBJDIR);fi
>     if ! test -e $(LIBDIR);then mkdir $(LIBDIR);fi
> 
> When running gmake, I've receive:
> if ! test -e C:\OWL\\bin;then mkdir C:\OWL\\bin;fi
> No se esperaba test en este momento.
> gmake: *** 
> [CreateDirs] Error 255
> 
> "No se esperaba test en este momento" is 
> spanish for "Unexpected 'test' at this moment"

The problem isn't with 'make'; it's with the shell. You're using
*nix shell syntax, but I believe you're running CMD.EXE because
this is what happens if I paste the command above into CMD.EXE:

  C:\home>if ! test -e C:\OWL\\bin;then mkdir C:\OWL\\bin;fi
  test was unexpected at this time.

If CMD.EXE is the shell you intend to use, then you'll need to
use its commands and syntax (which I've forgotten, but it's not
as powerful as a *nix shell).

However, since you say:
> I want to have the same .mak  for both Linux and Windows.
you'll need a msw port of a *nix shell...so get MSYS here:
  http://sourceforge.net/projects/mingw/files/
Then you should probably write something like this:
  SHELL:=/bin/sh
in your makefile--see:
  http://www.gnu.org/software/make/manual/html_node/Choosing-the-Shell.html

------------------------------------------------------------------------------
_______________________________________________


Thank's Greg!!
It's working now.

Sebas.



.



      

Thanks Tor!

That was the clue I was looking for.

Using AllocConsole()  the following code functions:
m_nCRTOut = _open_osfhandle((intptr_t) handle, _O_TEXT);
if (-1 == m_nCRTOut) return;
m_fpCRTOut = _fdopen(m_nCRTOut, "w");
if (! m_fpCRTOut) return;
setvbuf(m_fpCRTOut, 0, _IONBF, 0);
m_fOldStdOut = *stdout;
*stdout = *m_fpCRTOut;
...
std::ios::sync_with_stdio();

However the above code doesn't function with 
AttachConsole(ATTACH_PARENT_PROCESS).

With AttachConsole(ATTACH_PARENT_PROCESS) (and AllocConsole) the 
following code functions:
m_conOut.open("CONOUT$", std::ios::out);
if (m_conOut.good())
{
   m_oldCout = std::cout.rdbuf();
   std::cout.rdbuf(m_conOut.rdbuf());
}

Tor Lillqvist wrote:
>> I´ve been searching for this to attach the std c++ streams to the parent
>> console but so far without luck. If anyone knows how to, please...
>>     
>
> AttachConsole(ATTACH_PARENT_PROCESS). After that, re-open the streams
> to/from pseudo-filenames "CONOUT$" and "CONIN$".
>
> (AttachConsole is present only on XP or later, so you need to define
> _WIN32_WINNT appropriately to pick up its declaration from
> <windows.h>. And if you need your compiled code to run also on Windows
> 2000 or even earlier, you need to look up AttachConsole() dynamically
> at run-time, of course.)
>
> (We went through this same issue just some weeks ago on this very list.)
>
> --tml
>