From djberg96 at gmail.com  Sat Apr 12 10:22:59 2008
From: djberg96 at gmail.com (Daniel Berger)
Date: Sat, 12 Apr 2008 08:22:59 -0600
Subject: [Win32utils-devel] file-temp failure
Message-ID: <4800C5C3.9020604@gmail.com>

Hi all,

Anyone know why this code would work with VC++ 6, but fail with VC++ 8? 
Attached is the source.

The sample code I tried is:

require 'file/temp'
fh = FileTemp.new
fh.puts "hello"
fh.close

But it chokes with an Errno::EBADF error in the call to FileTemp.new.

I tried removing the call to tmpfile_s() and just using tmpfile(), but 
that didn't seem to matter.

Thanks,

Dan

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: test.rb
Url: http://rubyforge.org/pipermail/win32utils-devel/attachments/20080412/c0e1e4f3/attachment.pl 
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: extconf.rb
Url: http://rubyforge.org/pipermail/win32utils-devel/attachments/20080412/c0e1e4f3/attachment-0001.pl 
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: temp.c
Url: http://rubyforge.org/pipermail/win32utils-devel/attachments/20080412/c0e1e4f3/attachment.c 
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: temp.h
Url: http://rubyforge.org/pipermail/win32utils-devel/attachments/20080412/c0e1e4f3/attachment.h 

From djberg96 at gmail.com  Sat Apr 12 10:42:59 2008
From: djberg96 at gmail.com (Daniel Berger)
Date: Sat, 12 Apr 2008 08:42:59 -0600
Subject: [Win32utils-devel] file-temp failure
In-Reply-To: <4800C5C3.9020604@gmail.com>
References: <4800C5C3.9020604@gmail.com>
Message-ID: <6037b70c0804120742n1de9d96fwd9a08265f7a536f1@mail.gmail.com>

It's actually "wb+", but that doesn't help.

On Sat, Apr 12, 2008 at 8:22 AM, Daniel Berger <djberg96 at gmail.com> wrote:
> Hi all,
>
>  Anyone know why this code would work with VC++ 6, but fail with VC++ 8?
> Attached is the source.
>
>  The sample code I tried is:
>
>  require 'file/temp'
>  fh = FileTemp.new
>  fh.puts "hello"
>  fh.close
>
>  But it chokes with an Errno::EBADF error in the call to FileTemp.new.
>
>  I tried removing the call to tmpfile_s() and just using tmpfile(), but that
> didn't seem to matter.
>
>  Thanks,
>
>  Dan
>
>
> $:.unshift Dir.pwd
>
>  require 'temp'
>
>  fh = FileTemp.new
>  fh.puts "hello"
>  fh.close
>
> require 'mkmf'
>  dir_config('tmpfile')
>
>  have_func('tmpfile_s')
>  have_func('_sopen_s')
>
>  unless have_func('tmpfile')
>    raise 'This library is not supported on your platform. Aborting'
>  end
>
>  have_func('mkstemp')
>
>  create_makefile('file/temp')
>
> #include <ruby.h>
>  #include <stdio.h>
>
>  #ifdef HAVE__SOPEN_S
>  #include <share.h>
>  #endif
>
>  #ifndef HAVE_MKSTEMP
>  #include <fcntl.h>
>  #include <temp.h>
>  #endif
>
>  #ifndef P_tmpdir
>  #define P_tmpdir "/tmp"
>  #endif
>
>  #define VERSION "0.1.3"
>
>  VALUE cFileTemp;
>
>  /* call-seq:
>   *    FileTemp.new(delete = true, template = 'rb_file_temp_XXXXXX') => file
>   *
>   * Creates a new, anonymous temporary file in your FileTemp::TMPDIR
> directory,
>   * or /tmp if that cannot be accessed. If your $TMPDIR environment variable
> is
>   * set, it will be used instead. If $TMPDIR is not writable by the process,
> it
>   * will resort back to FileTemp::TMPDIR or /tmp.
>   *
>   * If the +delete+ option is set to true (the default) then the temporary
> file
>   * will be deleted automatically as soon as all references to it are
> closed.
>   * Otherwise, the file will live on in your $TMPDIR.
>   *
>   * If the +delete+ option is set to false, then the file is *not* deleted.
> In
>   * addition, you can supply a string +template+ that the system replaces
> with
>   * a unique filename. This template should end with 3 to 6 'X' characters.
>   * The default template is 'rb_file_temp_XXXXXX'. In this case the
> temporary
>   * file lives in the directory where it was created.
>   *
>   * The +template+ argument is ignored if the +delete+ argument is true.
>   */
>  static VALUE tempfile_init(int argc, VALUE* argv, VALUE self){
>    VALUE v_args[2];
>    VALUE v_delete;
>    VALUE v_template;
>
>    rb_scan_args(argc, argv, "02", &v_delete, &v_template);
>
>    if(NIL_P(v_delete))
>       v_delete = Qtrue;
>
>    if(RTEST(v_delete)){
>       /*
>  #ifdef HAVE_TMPFILE_S
>       FILE* fptr;
>
>       if(tmpfile_s(&fptr))
>          rb_sys_fail("tmpfile_s()");
>  #else
>  */
>       FILE* fptr = tmpfile();
>       if(!fptr)
>          rb_sys_fail("tmpfile()");
>  //#endif
>       v_args[0] = INT2FIX(fileno(fptr));
>    }
>    else{
>       int fd;
>
>       if(NIL_P(v_template))
>          v_template = rb_str_new2("rb_file_temp_XXXXXX");
>
>       fd = mkstemp(StringValuePtr(v_template));
>
>       if(fd < 0)
>          rb_sys_fail("mkstemp()");
>
>       v_args[0] = INT2FIX(fd);
>    }
>
>    // This bit of explicitness is necessary for MS Windows
>    v_args[1] = rb_str_new2("wb");
>
>    return rb_call_super(2, v_args);
>  }
>
>  /*
>   * Generates a unique file name, prefixed with the value of
> FileTemp::TMPDIR.
>   * Note that a file is not actually generated on the filesystem.
>   */
>  static VALUE tempfile_tmpnam(VALUE klass){
>    char buf[L_tmpnam];
>    return rb_str_new2(tmpnam(buf));
>  }
>
>  void Init_temp(){
>
>    /* The FileTemp class creates managed temporary files. Unlike Ruby's
>     * Tempfile class from this standard library, this is a subclass of File.
>     * In addition, the temporary file is automatically deleted when all
>     * references to it are closed (instead of waiting until the Ruby process
>     * is complete).
>     */
>    cFileTemp = rb_define_class("FileTemp", rb_cFile);
>
>    /* Instance Methods */
>    rb_define_method(cFileTemp, "initialize", tempfile_init, -1);
>
>    /* Singleton Methods */
>    rb_define_singleton_method(cFileTemp, "temp_name", tempfile_tmpnam, 0);
>
>    /* Constants */
>
>    /* ENV['P_tmpdir']: Your system's tmpdir */
>    rb_define_const(cFileTemp, "TMPDIR", rb_str_new2(P_tmpdir));
>
>    /* 0.1.3: The version of this library */
>    rb_define_const(cFileTemp, "VERSION", rb_str_new2(VERSION));
>  }
>
> /* This is a custom definition of the mkstemp() function for those
>   * platforms that don't support it natively.
>   */
>  int mkstemp(char* mtemplate){
>    int fd;
>
>  #ifdef HAVE__SOPEN_S
>    if(_sopen_s(
>       &fd,
>       mtemplate,
>       O_RDWR | O_BINARY | O_CREAT | O_EXCL | _O_SHORT_LIVED,
>       _SH_DENYNO,
>       _S_IREAD |_S_IWRITE
>    ))
>       rb_sys_fail("_sopen_s()");
>  #else
>    fd = _open(
>       mtemplate,
>       O_RDWR | O_BINARY | O_CREAT | O_EXCL | _O_SHORT_LIVED,
>       _S_IREAD |_S_IWRITE
>    );
>
>    if(fd < 0)
>       rb_sys_fail("open()");
>  #endif
>    return fd;
>  }
>

From djberg96 at gmail.com  Sat Apr 12 10:49:42 2008
From: djberg96 at gmail.com (Daniel Berger)
Date: Sat, 12 Apr 2008 08:49:42 -0600
Subject: [Win32utils-devel] file-temp failure
In-Reply-To: <6037b70c0804120742n1de9d96fwd9a08265f7a536f1@mail.gmail.com>
References: <4800C5C3.9020604@gmail.com>
	<6037b70c0804120742n1de9d96fwd9a08265f7a536f1@mail.gmail.com>
Message-ID: <6037b70c0804120749o207f675ardbeca36f7b339de7@mail.gmail.com>

Upon further review, it seems I had a botched Ruby build. It works
with p114 anyway.

Sorry for the noise.

Dan

On Sat, Apr 12, 2008 at 8:42 AM, Daniel Berger <djberg96 at gmail.com> wrote:
> It's actually "wb+", but that doesn't help.
>
>
>
>  On Sat, Apr 12, 2008 at 8:22 AM, Daniel Berger <djberg96 at gmail.com> wrote:
>  > Hi all,
>  >
>  >  Anyone know why this code would work with VC++ 6, but fail with VC++ 8?
>  > Attached is the source.
>  >
>  >  The sample code I tried is:
>  >
>  >  require 'file/temp'
>  >  fh = FileTemp.new
>  >  fh.puts "hello"
>  >  fh.close
>  >
>  >  But it chokes with an Errno::EBADF error in the call to FileTemp.new.
>  >
>  >  I tried removing the call to tmpfile_s() and just using tmpfile(), but that
>  > didn't seem to matter.
>  >
>  >  Thanks,
>  >
>  >  Dan
>  >
>  >
>  > $:.unshift Dir.pwd
>  >
>  >  require 'temp'
>
> >
>  >  fh = FileTemp.new
>  >  fh.puts "hello"
>  >  fh.close
>  >
>  > require 'mkmf'
>  >  dir_config('tmpfile')
>  >
>  >  have_func('tmpfile_s')
>  >  have_func('_sopen_s')
>  >
>  >  unless have_func('tmpfile')
>  >    raise 'This library is not supported on your platform. Aborting'
>  >  end
>  >
>  >  have_func('mkstemp')
>  >
>  >  create_makefile('file/temp')
>  >
>  > #include <ruby.h>
>  >  #include <stdio.h>
>  >
>  >  #ifdef HAVE__SOPEN_S
>  >  #include <share.h>
>  >  #endif
>  >
>  >  #ifndef HAVE_MKSTEMP
>  >  #include <fcntl.h>
>  >  #include <temp.h>
>  >  #endif
>  >
>  >  #ifndef P_tmpdir
>  >  #define P_tmpdir "/tmp"
>  >  #endif
>  >
>  >  #define VERSION "0.1.3"
>  >
>  >  VALUE cFileTemp;
>  >
>  >  /* call-seq:
>  >   *    FileTemp.new(delete = true, template = 'rb_file_temp_XXXXXX') => file
>  >   *
>  >   * Creates a new, anonymous temporary file in your FileTemp::TMPDIR
>  > directory,
>  >   * or /tmp if that cannot be accessed. If your $TMPDIR environment variable
>  > is
>  >   * set, it will be used instead. If $TMPDIR is not writable by the process,
>  > it
>  >   * will resort back to FileTemp::TMPDIR or /tmp.
>  >   *
>  >   * If the +delete+ option is set to true (the default) then the temporary
>  > file
>  >   * will be deleted automatically as soon as all references to it are
>  > closed.
>  >   * Otherwise, the file will live on in your $TMPDIR.
>  >   *
>  >   * If the +delete+ option is set to false, then the file is *not* deleted.
>  > In
>  >   * addition, you can supply a string +template+ that the system replaces
>  > with
>  >   * a unique filename. This template should end with 3 to 6 'X' characters.
>  >   * The default template is 'rb_file_temp_XXXXXX'. In this case the
>  > temporary
>  >   * file lives in the directory where it was created.
>  >   *
>  >   * The +template+ argument is ignored if the +delete+ argument is true.
>  >   */
>  >  static VALUE tempfile_init(int argc, VALUE* argv, VALUE self){
>  >    VALUE v_args[2];
>  >    VALUE v_delete;
>  >    VALUE v_template;
>  >
>  >    rb_scan_args(argc, argv, "02", &v_delete, &v_template);
>  >
>  >    if(NIL_P(v_delete))
>  >       v_delete = Qtrue;
>  >
>  >    if(RTEST(v_delete)){
>  >       /*
>  >  #ifdef HAVE_TMPFILE_S
>  >       FILE* fptr;
>  >
>  >       if(tmpfile_s(&fptr))
>  >          rb_sys_fail("tmpfile_s()");
>  >  #else
>  >  */
>  >       FILE* fptr = tmpfile();
>  >       if(!fptr)
>  >          rb_sys_fail("tmpfile()");
>  >  //#endif
>  >       v_args[0] = INT2FIX(fileno(fptr));
>  >    }
>  >    else{
>  >       int fd;
>  >
>  >       if(NIL_P(v_template))
>  >          v_template = rb_str_new2("rb_file_temp_XXXXXX");
>  >
>  >       fd = mkstemp(StringValuePtr(v_template));
>  >
>  >       if(fd < 0)
>  >          rb_sys_fail("mkstemp()");
>  >
>  >       v_args[0] = INT2FIX(fd);
>  >    }
>  >
>  >    // This bit of explicitness is necessary for MS Windows
>  >    v_args[1] = rb_str_new2("wb");
>  >
>  >    return rb_call_super(2, v_args);
>  >  }
>  >
>  >  /*
>  >   * Generates a unique file name, prefixed with the value of
>  > FileTemp::TMPDIR.
>  >   * Note that a file is not actually generated on the filesystem.
>  >   */
>  >  static VALUE tempfile_tmpnam(VALUE klass){
>  >    char buf[L_tmpnam];
>  >    return rb_str_new2(tmpnam(buf));
>  >  }
>  >
>  >  void Init_temp(){
>  >
>  >    /* The FileTemp class creates managed temporary files. Unlike Ruby's
>  >     * Tempfile class from this standard library, this is a subclass of File.
>  >     * In addition, the temporary file is automatically deleted when all
>  >     * references to it are closed (instead of waiting until the Ruby process
>  >     * is complete).
>  >     */
>  >    cFileTemp = rb_define_class("FileTemp", rb_cFile);
>  >
>  >    /* Instance Methods */
>  >    rb_define_method(cFileTemp, "initialize", tempfile_init, -1);
>  >
>  >    /* Singleton Methods */
>  >    rb_define_singleton_method(cFileTemp, "temp_name", tempfile_tmpnam, 0);
>  >
>  >    /* Constants */
>  >
>  >    /* ENV['P_tmpdir']: Your system's tmpdir */
>  >    rb_define_const(cFileTemp, "TMPDIR", rb_str_new2(P_tmpdir));
>  >
>  >    /* 0.1.3: The version of this library */
>  >    rb_define_const(cFileTemp, "VERSION", rb_str_new2(VERSION));
>  >  }
>  >
>  > /* This is a custom definition of the mkstemp() function for those
>  >   * platforms that don't support it natively.
>  >   */
>  >  int mkstemp(char* mtemplate){
>  >    int fd;
>  >
>  >  #ifdef HAVE__SOPEN_S
>  >    if(_sopen_s(
>  >       &fd,
>  >       mtemplate,
>  >       O_RDWR | O_BINARY | O_CREAT | O_EXCL | _O_SHORT_LIVED,
>  >       _SH_DENYNO,
>  >       _S_IREAD |_S_IWRITE
>  >    ))
>  >       rb_sys_fail("_sopen_s()");
>  >  #else
>  >    fd = _open(
>  >       mtemplate,
>  >       O_RDWR | O_BINARY | O_CREAT | O_EXCL | _O_SHORT_LIVED,
>  >       _S_IREAD |_S_IWRITE
>  >    );
>  >
>  >    if(fd < 0)
>  >       rb_sys_fail("open()");
>  >  #endif
>  >    return fd;
>  >  }
>  >
>

From djberg96 at gmail.com  Wed Apr 16 00:16:09 2008
From: djberg96 at gmail.com (Daniel Berger)
Date: Tue, 15 Apr 2008 22:16:09 -0600
Subject: [Win32utils-devel] Rethinking stat
Message-ID: <48057D89.6030901@gmail.com>

Here's an interesting link on stat'ing a file on Windows:

http://www.mail-archive.com/factor-talk at lists.sourceforge.net/msg00816.html

Regards,

Dan

From djberg96 at gmail.com  Thu Apr 17 19:18:38 2008
From: djberg96 at gmail.com (Daniel Berger)
Date: Thu, 17 Apr 2008 17:18:38 -0600
Subject: [Win32utils-devel] Process.create patch - please review
Message-ID: <4807DACE.9060002@gmail.com>

Hi all,

There's a patch on the tracker I'd like some feedback on. It's for 
Process.create to handle command_line and app_name separately. In 
practice this only seems to affect how command line parameters are seen 
by something like GetCommandLine().

Please take a look at the command-line.patch at:

http://rubyforge.org/tracker/index.php?func=detail&aid=18280&group_id=85&atid=411

And let me know if you think it's reasonable.

Thanks,

Dan

From gthiesfeld at gmail.com  Thu Apr 17 19:45:11 2008
From: gthiesfeld at gmail.com (Gordon Thiesfeld)
Date: Thu, 17 Apr 2008 18:45:11 -0500
Subject: [Win32utils-devel] win32-api (1.0.5) on Ruby 1.8.6 Mingw
Message-ID: <e9ece6840804171645x14b945f1tdfaa9f64480487c1@mail.gmail.com>

When trying to compile the win32-api gem on ruby 1.8.6 (2008-03-03
patchlevel 114) [i386-mingw32], I get the following error:

gcc -shared -s -o api.so api.o -L"." -L"C:/ruby/ruby_mingw/lib" -L.
-Wl,--enable-auto-image-base,--enable-auto-import,--export-all
-lmsvcrt-ruby18  -lws2_32
api.o: In function `api_init':
C:/ruby/ruby_mingw/lib/ruby/gems/1.8/gems/win32-api-1.0.5/ext/win32/api.c:201:
undefined reference to `_tcscmp'
collect2: ld returned 1 exit status
make: *** [api.so] Error 1

Here's the relevant code from api.c (lines 200 - 204)

  /* The order of 'A' and 'W' is reversed if $KCODE is set to 'UTF8'. */
  if(!_tcscmp(rb_get_kcode(), "UTF8")){
     first  = "W";
     second = "A";
  }

Any ideas?

Thanks,

Gordon

From djberg96 at gmail.com  Thu Apr 17 20:05:20 2008
From: djberg96 at gmail.com (Daniel Berger)
Date: Thu, 17 Apr 2008 18:05:20 -0600
Subject: [Win32utils-devel] win32-api (1.0.5) on Ruby 1.8.6 Mingw
In-Reply-To: <e9ece6840804171645x14b945f1tdfaa9f64480487c1@mail.gmail.com>
References: <e9ece6840804171645x14b945f1tdfaa9f64480487c1@mail.gmail.com>
Message-ID: <4807E5C0.3080107@gmail.com>

Gordon Thiesfeld wrote:
> When trying to compile the win32-api gem on ruby 1.8.6 (2008-03-03
> patchlevel 114) [i386-mingw32], I get the following error:
> 
> gcc -shared -s -o api.so api.o -L"." -L"C:/ruby/ruby_mingw/lib" -L.
> -Wl,--enable-auto-image-base,--enable-auto-import,--export-all
> -lmsvcrt-ruby18  -lws2_32
> api.o: In function `api_init':
> C:/ruby/ruby_mingw/lib/ruby/gems/1.8/gems/win32-api-1.0.5/ext/win32/api.c:201:
> undefined reference to `_tcscmp'
> collect2: ld returned 1 exit status
> make: *** [api.so] Error 1
> 
> Here's the relevant code from api.c (lines 200 - 204)
> 
>   /* The order of 'A' and 'W' is reversed if $KCODE is set to 'UTF8'. */
>   if(!_tcscmp(rb_get_kcode(), "UTF8")){
>      first  = "W";
>      second = "A";
>   }
> 
> Any ideas?

Yeah, I got a little overzealous with the TCHAR stuff there. Just change 
"_tcscmp" to "strcmp" and rebuild. I think I've already made this change 
in CVS, since we know $KCODE is always ASCII.

Regards,

Dan


From djberg96 at gmail.com  Fri Apr 18 07:16:59 2008
From: djberg96 at gmail.com (Daniel Berger)
Date: Fri, 18 Apr 2008 05:16:59 -0600
Subject: [Win32utils-devel] win32-api segfault
Message-ID: <4808832B.1030103@gmail.com>

It looks like win32-api, both 1.0.5 and the latest in CVS, are 
segfaulting when I build with VC++ 8 or later. Both work fine with VC++ 
6, though.

Thinking it might be strncpy_s, I removed that, but it didn't help. I 
installed a fresh Ruby built with VC++ 8 as well, but that didn't help 
either.

Loaded suite C:/rubyvc8/lib/ruby/site_ruby/1.8/rake/rake_test_loader
Started
............/test/tc_win32_api_callback.rb:45: [BUG] Segmentation fault
ruby 1.8.6 (2008-03-03) [i386-mswin32_80]

Any ideas?

Thanks,

Dan

From phasis at gmail.com  Fri Apr 18 08:39:59 2008
From: phasis at gmail.com (Heesob Park)
Date: Fri, 18 Apr 2008 21:39:59 +0900
Subject: [Win32utils-devel] win32-api segfault
In-Reply-To: <4808832B.1030103@gmail.com>
References: <4808832B.1030103@gmail.com>
Message-ID: <aea95bc50804180539l1dc38c70k20c06a8694308be0@mail.gmail.com>

Hi,

2008/4/18, Daniel Berger <djberg96 at gmail.com>:

> It looks like win32-api, both 1.0.5 and the latest in CVS, are
> segfaulting when I build with VC++ 8 or later. Both work fine with VC++
> 6, though.
>
> Thinking it might be strncpy_s, I removed that, but it didn't help. I
> installed a fresh Ruby built with VC++ 8 as well, but that didn't help
> either.
>
> Loaded suite C:/rubyvc8/lib/ruby/site_ruby/1.8/rake/rake_test_loader
> Started
> ............/test/tc_win32_api_callback.rb:45: [BUG] Segmentation fault
> ruby 1.8.6 (2008-03-03) [i386-mswin32_80]
>
> Any ideas?


I have no idea.

In my test with ruby 1.8.6 (2007-09-24 patchlevel 111) [i386-mswin32_90]
the latest in CVS works fine.

BTW, why not use Visual C++ 2008 Express Edition(VC++ 9.0) ? It is free
also.
http://www.microsoft.com/express/vc/

Regards,

Park Heesob
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/win32utils-devel/attachments/20080418/764dfcb1/attachment.html 

From gthiesfeld at gmail.com  Fri Apr 18 10:46:25 2008
From: gthiesfeld at gmail.com (Gordon Thiesfeld)
Date: Fri, 18 Apr 2008 09:46:25 -0500
Subject: [Win32utils-devel] win32-api (1.0.5) on Ruby 1.8.6 Mingw
In-Reply-To: <4807E5C0.3080107@gmail.com>
References: <e9ece6840804171645x14b945f1tdfaa9f64480487c1@mail.gmail.com>
	<4807E5C0.3080107@gmail.com>
Message-ID: <e9ece6840804180746i7e8f1a66mdf8095bca9d66e37@mail.gmail.com>

On Thu, Apr 17, 2008 at 7:05 PM, Daniel Berger <djberg96 at gmail.com> wrote:
>  Yeah, I got a little overzealous with the TCHAR stuff there. Just change
>  "_tcscmp" to "strcmp" and rebuild. I think I've already made this change
>  in CVS, since we know $KCODE is always ASCII.

That did the trick.  Thanks.

From djberg96 at gmail.com  Sat Apr 19 13:42:49 2008
From: djberg96 at gmail.com (Daniel Berger)
Date: Sat, 19 Apr 2008 11:42:49 -0600
Subject: [Win32utils-devel] [Fwd: win32-service-0.6.1 using a Rails Model]
Message-ID: <480A2F19.5040607@gmail.com>

Any ideas here? Line 911-921 indicates a failure in OpenService:

handle_scs = OpenService(
    handle_scm,
    service,
    SERVICE_QUERY_STATUS
)

if handle_scs == 0
    error = get_last_error
    CloseServiceHandle(handle_scm)
    raise Error, error
end

Something to do with WEBrick perhaps?

Regards,

Dan
-------------- next part --------------
An embedded message was scrubbed...
From: david <davidhooey at gmail.com>
Subject: win32-service-0.6.1 using a Rails Model
Date: Sun, 20 Apr 2008 01:30:07 +0900
Size: 10165
Url: http://rubyforge.org/pipermail/win32utils-devel/attachments/20080419/bd238074/attachment-0001.eml 

From noreply at rubyforge.org  Sun Apr 20 16:05:47 2008
From: noreply at rubyforge.org (noreply at rubyforge.org)
Date: Sun, 20 Apr 2008 16:05:47 -0400 (EDT)
Subject: [Win32utils-devel] [ win32utils-Feature Requests-19668 ] play 2
	tones ones on sometime
Message-ID: <20080420200547.1501618585DE@rubyforge.org>

Feature Requests item #19668, was opened at 2008-04-20 20:05
You can respond by visiting: 
http://rubyforge.org/tracker/?func=detail&atid=414&aid=19668&group_id=85

Category: win32-sound
Group: Code
Status: Open
Resolution: None
Priority: 3
Submitted By: Christof Spies (wingfire)
Assigned to: Nobody (None)
Summary: play 2 tones ones on sometime

Initial Comment:
Hi,

i need to play a dualtone.

Instead of playing behind
Sound.beep(100, 1000)
Sound.beep(200, 1000)

Play both frequences on the same time
Sound.beep([100, 200], 1000)

Christof

----------------------------------------------------------------------

You can respond by visiting: 
http://rubyforge.org/tracker/?func=detail&atid=414&aid=19668&group_id=85

From djberg96 at gmail.com  Mon Apr 21 23:16:21 2008
From: djberg96 at gmail.com (Daniel Berger)
Date: Mon, 21 Apr 2008 21:16:21 -0600
Subject: [Win32utils-devel] Global substitution with win32-mmap
Message-ID: <480D5885.9090600@gmail.com>

Hi all,

I'm trying to figure out how to do an inline file substitute/replace 
using the address. In short, I'm trying to replace "assert_raises" with 
"assert_raise" throughout a file.

I tried this:

require 'win32/mmap'
require 'windows/msvcrt/string'
require 'windows/msvcrt/buffer'
include Win32
include Windows::MSVCRT::String
include Windows::MSVCRT::Buffer

StrlenL = API.new('strlen', 'L', 'L', 'msvcrt')

MMap.new(:file => 'test.rb') do |addr|
    old_str = 'assert_raises'
    old_len = old_str.length
    new_str = 'assert_raise'
    new_len = new_str.length

    ptr1 = ptr2 = ptr3 = [strstr(addr,old_str)].pack('p*').unpack('l')[0]

    while ptr1 && ptr1 != 0
       ptr2 += new_len
       ptr3 += old_len
       memmove(ptr2, ptr3, 1 + StrlenL.call(ptr3))
       memcpy(ptr1, new_str, new_len)
       ptr1 = ptr2 = ptr3 = [strstr(ptr2, 
old_str)].pack('p*').unpack('l')[0]
    end
end

I based it off of this thread:

http://tinyurl.com/5qvbnw

But, that doesn't seem to be working. Any ideas?

Thanks,

Dan

# test.rb
class Foo < Test::Unit::TestCase
    def setup
       @foo = Foo.new
    end

    def test_bar_expected_errors
       assert_raises(ArgumentError){ 1 == 2 }
       assert_raise(StandardError){ 2 == 1 }
    end

    def test_foo_expected_errors
       assert_raises(ArgumentError){ 1 == 2 }
       assert_raise(StandardError){ 2 == 1 }
    end

    def teardown
       @foo = nil
    end
en

From phasis at gmail.com  Tue Apr 22 01:51:44 2008
From: phasis at gmail.com (Heesob Park)
Date: Tue, 22 Apr 2008 14:51:44 +0900
Subject: [Win32utils-devel] Global substitution with win32-mmap
In-Reply-To: <480D5885.9090600@gmail.com>
References: <480D5885.9090600@gmail.com>
Message-ID: <aea95bc50804212251w2ef3aa17wf601c3e54534af28@mail.gmail.com>

Hi,

2008/4/22, Daniel Berger <djberg96 at gmail.com>:
>
> Hi all,
>
> I'm trying to figure out how to do an inline file substitute/replace
> using the address. In short, I'm trying to replace "assert_raises" with
> "assert_raise" throughout a file.
>
> I tried this:
>
> require 'win32/mmap'
> require 'windows/msvcrt/string'
> require 'windows/msvcrt/buffer'
> include Win32
> include Windows::MSVCRT::String
> include Windows::MSVCRT::Buffer
>
> StrlenL = API.new('strlen', 'L', 'L', 'msvcrt')
>
> MMap.new(:file => 'test.rb') do |addr|
>    old_str = 'assert_raises'
>    old_len = old_str.length
>    new_str = 'assert_raise'
>    new_len = new_str.length
>
>    ptr1 = ptr2 = ptr3 = [strstr(addr,old_str)].pack('p*').unpack('l')[0]
>
>    while ptr1 && ptr1 != 0
>       ptr2 += new_len
>       ptr3 += old_len
>       memmove(ptr2, ptr3, 1 + StrlenL.call(ptr3))
>       memcpy(ptr1, new_str, new_len)
>       ptr1 = ptr2 = ptr3 = [strstr(ptr2,
> old_str)].pack('p*').unpack('l')[0]
>    end
> end
>
> I based it off of this thread:
>
> http://tinyurl.com/5qvbnw
>
> But, that doesn't seem to be working. Any ideas?



You must define strstr like this:
         Strstr   = API.new('strstr', 'LP', 'L', 'msvcrt')

and try this:

StrlenL = API.new('strlen', 'L', 'L', 'msvcrt')

MMap.new(:file => 'test.rb') do |addr|
   old_str = 'assert_raises'
   old_len = old_str.length
   new_str = 'assert_raise'
   new_len = new_str.length

   ptr1 = ptr2 = ptr3 = strstr(addr,old_str)
   while ptr1 && ptr1 != 0
      ptr2 += new_len
      ptr3 += old_len
      memmove(ptr2, ptr3, 1 + StrlenL.call(ptr3))
      memcpy(ptr1, new_str, new_len)
      ptr1 = ptr2 = ptr3 = strstr(ptr2,old_str)
   end
end
Regads,

Park Heesob
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/win32utils-devel/attachments/20080422/f9b876bb/attachment.html 

From Daniel.Berger at qwest.com  Thu Apr 24 10:13:37 2008
From: Daniel.Berger at qwest.com (Berger, Daniel)
Date: Thu, 24 Apr 2008 09:13:37 -0500
Subject: [Win32utils-devel] A generalized struct implementation for
	windows-api
Message-ID: <7524A45A1A5B264FA4809E2156496CFB023D31F4@ITOMAE2KM01.AD.QINTRA.COM>

Hi,
 
After yet another win32-process bug, I've decided that we need some form
of generalized struct handling for use with the windows-api library
instead of packing and unpacking data structures manually.
 
We currently do stuff like this:
 
proc_info = 0.chr * 16 # sizeof(PROCESS_INFORMATION)
Process.create(... proc_info ...)
CloseHandle(proc_info[0,4])
  
I would like a way to generalize the creation of these structures.
Something like this:
 
require 'win32/struct'
 
Win32::Struct.new(
   'PROCESS_INFORMATION',
   :hProcess    => HANDLE,
   :hThread     => HANDLE,
   :dwProcessId => DWORD,
   :dwThreadId  => DWORD
)

And, with that in place, we could then write:
 
proc_info = PROCESS_INFORMATION.new
Process.create(... proc_info ...)
CloseHandle(proc_info.hProcess)

The struct would automatically handle the packing/unpacking behind the
scenes based on the data type that was declared in the constructor.
 
It's easier to read and easier to maintain, IMO. And, we could predefine
a bunch of the most common structures in advance so that users wouldn't
have to do this themselves. They would be immediately available.

This could probably be an addon to the windows-api library already out
there.

What do you think?

Thanks,

Dan


This communication is the property of Qwest and may contain confidential or
privileged information. Unauthorized use of this communication is strictly 
prohibited and may be unlawful.  If you have received this communication 
in error, please immediately notify the sender by reply e-mail and destroy 
all copies of the communication and any attachments.

From Daniel.Berger at qwest.com  Fri Apr 25 11:30:30 2008
From: Daniel.Berger at qwest.com (Berger, Daniel)
Date: Fri, 25 Apr 2008 10:30:30 -0500
Subject: [Win32utils-devel] win32-mmap test failures
Message-ID: <7524A45A1A5B264FA4809E2156496CFB023D31FF@ITOMAE2KM01.AD.QINTRA.COM>

Hi all,

This is odd. It looks to me like we have all 4 variations of memcpy set
in Windows::MSVCRT::Buffer but somehow this one creeps up. I actually
noticed it in a few cases with that earlier mmap sub/replace example,
but I wasn't sure what was happening.

Note that windows-api 0.2.3 and windows-pr-0.8.3 are the latest versions
in CVS, but this happens with windows-api-0.2.2 and windows-pr-0.8.2 as
well.

Any ideas?

Thanks,

Dan

C:\Documents and Settings\djberge\workspace\win32-mmap>rake test
(in C:/Documents and Settings/djberge/workspace/win32-mmap)
c:/ruby/bin/ruby -w -Ilib;test
"c:/ruby/lib/ruby/gems/1.8/gems/rake-0.8.1/lib/rake/rake_test_loader.rb"
"test/
tc_mmap.rb"
Loaded suite
c:/ruby/lib/ruby/gems/1.8/gems/rake-0.8.1/lib/rake/rake_test_loader
Started
....FFFF.......
Finished in 0.031 seconds.
 
  1) Failure:
test_dynamic_array_value(TC_Mmap) [./test/tc_mmap.rb:42]:
Exception raised:
Class: <TypeError>
Message: <"can't convert String into Integer">
---Backtrace---
c:/ruby/lib/ruby/gems/1.8/gems/windows-api-0.2.3/lib/windows/api.rb:479:
in `call'
c:/ruby/lib/ruby/gems/1.8/gems/windows-api-0.2.3/lib/windows/api.rb:479:
in `call'
c:/ruby/lib/ruby/gems/1.8/gems/windows-pr-0.8.3/lib/windows/msvcrt/buffe
r.rb:34:in `memcpy'
./lib/win32/mmap.rb:321:in `method_missing'
./test/tc_mmap.rb:42:in `test_dynamic_array_value'
./test/tc_mmap.rb:42:in `test_dynamic_array_value'
---------------
 
  2) Failure:
test_dynamic_hash_value(TC_Mmap) [./test/tc_mmap.rb:35]:
Exception raised:
Class: <TypeError>
Message: <"can't convert String into Integer">
---Backtrace---
c:/ruby/lib/ruby/gems/1.8/gems/windows-api-0.2.3/lib/windows/api.rb:479:
in `call'
c:/ruby/lib/ruby/gems/1.8/gems/windows-api-0.2.3/lib/windows/api.rb:479:
in `call'
c:/ruby/lib/ruby/gems/1.8/gems/windows-pr-0.8.3/lib/windows/msvcrt/buffe
r.rb:34:in `memcpy'
./lib/win32/mmap.rb:321:in `method_missing'
./test/tc_mmap.rb:35:in `test_dynamic_hash_value'
./test/tc_mmap.rb:35:in `test_dynamic_hash_value'
---------------
 
  3) Failure:
test_dynamic_integer_value(TC_Mmap) [./test/tc_mmap.rb:28]:
Exception raised:
Class: <TypeError>
Message: <"can't convert String into Integer">
---Backtrace---
c:/ruby/lib/ruby/gems/1.8/gems/windows-api-0.2.3/lib/windows/api.rb:479:
in `call'
c:/ruby/lib/ruby/gems/1.8/gems/windows-api-0.2.3/lib/windows/api.rb:479:
in `call'
c:/ruby/lib/ruby/gems/1.8/gems/windows-pr-0.8.3/lib/windows/msvcrt/buffe
r.rb:34:in `memcpy'
./lib/win32/mmap.rb:321:in `method_missing'
./test/tc_mmap.rb:28:in `test_dynamic_integer_value'
./test/tc_mmap.rb:28:in `test_dynamic_integer_value'
---------------
 
  4) Failure:
test_dynamic_string_value(TC_Mmap) [./test/tc_mmap.rb:21]:
Exception raised:
Class: <TypeError>
Message: <"can't convert String into Integer">
---Backtrace---
c:/ruby/lib/ruby/gems/1.8/gems/windows-api-0.2.3/lib/windows/api.rb:479:
in `call'
c:/ruby/lib/ruby/gems/1.8/gems/windows-api-0.2.3/lib/windows/api.rb:479:
in `call'
c:/ruby/lib/ruby/gems/1.8/gems/windows-pr-0.8.3/lib/windows/msvcrt/buffe
r.rb:34:in `memcpy'
./lib/win32/mmap.rb:321:in `method_missing'
./test/tc_mmap.rb:21:in `test_dynamic_string_value'
./test/tc_mmap.rb:21:in `test_dynamic_string_value'
---------------
 
15 tests, 31 assertions, 4 failures, 0 errors
rake aborted!
Command failed with status (1): [c:/ruby/bin/ruby -w -Ilib;test
"c:/ruby/li...]
 
(See full trace by running task with --trace)


This communication is the property of Qwest and may contain confidential or
privileged information. Unauthorized use of this communication is strictly 
prohibited and may be unlawful.  If you have received this communication 
in error, please immediately notify the sender by reply e-mail and destroy 
all copies of the communication and any attachments.

From Daniel.Berger at qwest.com  Fri Apr 25 12:05:45 2008
From: Daniel.Berger at qwest.com (Berger, Daniel)
Date: Fri, 25 Apr 2008 11:05:45 -0500
Subject: [Win32utils-devel] win32-mmap test failures
In-Reply-To: <7524A45A1A5B264FA4809E2156496CFB023D31FF@ITOMAE2KM01.AD.QINTRA.COM>
References: <7524A45A1A5B264FA4809E2156496CFB023D31FF@ITOMAE2KM01.AD.QINTRA.COM>
Message-ID: <7524A45A1A5B264FA4809E2156496CFB023D3201@ITOMAE2KM01.AD.QINTRA.COM>

 
> -----Original Message-----
> From: win32utils-devel-bounces at rubyforge.org 
> [mailto:win32utils-devel-bounces at rubyforge.org] On Behalf Of 
> Berger, Daniel
> Sent: Friday, April 25, 2008 9:31 AM
> To: Development and ideas for win32utils projects
> Subject: [Win32utils-devel] win32-mmap test failures
> 
> Hi all,
> 
> This is odd. It looks to me like we have all 4 variations of 
> memcpy set in Windows::MSVCRT::Buffer but somehow this one 
> creeps up. I actually noticed it in a few cases with that 
> earlier mmap sub/replace example, but I wasn't sure what was 
> happening.

<snip>

This is baffling. Somehow it's failing the src.is_a?(String) check on
the marshalled source on line 31 of buffer.rb, even though it definitely
_is_ a String. I cannot duplicate this in a standalone script.

Ideas?

Thanks,

Dan


This communication is the property of Qwest and may contain confidential or
privileged information. Unauthorized use of this communication is strictly 
prohibited and may be unlawful.  If you have received this communication 
in error, please immediately notify the sender by reply e-mail and destroy 
all copies of the communication and any attachments.

From phasis at gmail.com  Fri Apr 25 12:20:20 2008
From: phasis at gmail.com (Heesob Park)
Date: Sat, 26 Apr 2008 01:20:20 +0900
Subject: [Win32utils-devel] win32-mmap test failures
In-Reply-To: <7524A45A1A5B264FA4809E2156496CFB023D3201@ITOMAE2KM01.AD.QINTRA.COM>
References: <7524A45A1A5B264FA4809E2156496CFB023D31FF@ITOMAE2KM01.AD.QINTRA.COM>
	<7524A45A1A5B264FA4809E2156496CFB023D3201@ITOMAE2KM01.AD.QINTRA.COM>
Message-ID: <aea95bc50804250920v6776e1dcme00956641910a58@mail.gmail.com>

Hi,

2008/4/26, Berger, Daniel <Daniel.Berger at qwest.com>:
>
>
> > -----Original Message-----
> > From: win32utils-devel-bounces at rubyforge.org
> > [mailto:win32utils-devel-bounces at rubyforge.org] On Behalf Of
> > Berger, Daniel
> > Sent: Friday, April 25, 2008 9:31 AM
> > To: Development and ideas for win32utils projects
> > Subject: [Win32utils-devel] win32-mmap test failures
> >
> > Hi all,
> >
> > This is odd. It looks to me like we have all 4 variations of
> > memcpy set in Windows::MSVCRT::Buffer but somehow this one
> > creeps up. I actually noticed it in a few cases with that
> > earlier mmap sub/replace example, but I wasn't sure what was
> > happening.
>
> <snip>
>
> This is baffling. Somehow it's failing the src.is_a?(String) check on
> the marshalled source on line 31 of buffer.rb, even though it definitely
> _is_ a String. I cannot duplicate this in a standalone script.
>
> Ideas?
>
> Thanks,
>
> Dan


If you modify src.is_a?(String)  to src.class.to_s=="String", it will work.
But I don't know what it means.

Regards,

Park Heesob
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://rubyforge.org/pipermail/win32utils-devel/attachments/20080426/f3c6fadf/attachment-0001.html>

From Daniel.Berger at qwest.com  Fri Apr 25 12:36:51 2008
From: Daniel.Berger at qwest.com (Berger, Daniel)
Date: Fri, 25 Apr 2008 11:36:51 -0500
Subject: [Win32utils-devel] win32-mmap test failures
In-Reply-To: <aea95bc50804250920v6776e1dcme00956641910a58@mail.gmail.com>
References: <7524A45A1A5B264FA4809E2156496CFB023D31FF@ITOMAE2KM01.AD.QINTRA.COM><7524A45A1A5B264FA4809E2156496CFB023D3201@ITOMAE2KM01.AD.QINTRA.COM>
	<aea95bc50804250920v6776e1dcme00956641910a58@mail.gmail.com>
Message-ID: <7524A45A1A5B264FA4809E2156496CFB023D3202@ITOMAE2KM01.AD.QINTRA.COM>

Thanks.
 
The only thing I can figure is that it's some sort of name collision
with Windows::MSVCRT::String, but I can't see how right off the bat.
 
Dan


________________________________

	From: win32utils-devel-bounces at rubyforge.org
[mailto:win32utils-devel-bounces at rubyforge.org] On Behalf Of Heesob Park
	Sent: Friday, April 25, 2008 10:20 AM
	To: Development and ideas for win32utils projects
	Subject: Re: [Win32utils-devel] win32-mmap test failures
	
	
	Hi,
	
	
	2008/4/26, Berger, Daniel <Daniel.Berger at qwest.com>: 


		> -----Original Message-----
		> From: win32utils-devel-bounces at rubyforge.org
		> [mailto:win32utils-devel-bounces at rubyforge.org] On
Behalf Of
		> Berger, Daniel
		> Sent: Friday, April 25, 2008 9:31 AM
		> To: Development and ideas for win32utils projects
		> Subject: [Win32utils-devel] win32-mmap test failures
		>
		> Hi all,
		>
		> This is odd. It looks to me like we have all 4
variations of
		> memcpy set in Windows::MSVCRT::Buffer but somehow this
one
		> creeps up. I actually noticed it in a few cases with
that
		> earlier mmap sub/replace example, but I wasn't sure
what was
		> happening.
		
		<snip>
		
		This is baffling. Somehow it's failing the
src.is_a?(String) check on
		the marshalled source on line 31 of buffer.rb, even
though it definitely
		_is_ a String. I cannot duplicate this in a standalone
script.
		
		Ideas?
		
		Thanks,
		
		Dan


	If you modify src.is_a?(String)  to src.class.to_s=="String", it
will work.
	But I don't know what it means.
	 
	Regards,

	Park Heesob
	 

	 


This communication is the property of Qwest and may contain confidential or
privileged information. Unauthorized use of this communication is strictly 
prohibited and may be unlawful.  If you have received this communication 
in error, please immediately notify the sender by reply e-mail and destroy 
all copies of the communication and any attachments.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://rubyforge.org/pipermail/win32utils-devel/attachments/20080425/813ef391/attachment.html>

From Daniel.Berger at qwest.com  Fri Apr 25 13:04:13 2008
From: Daniel.Berger at qwest.com (Berger, Daniel)
Date: Fri, 25 Apr 2008 12:04:13 -0500
Subject: [Win32utils-devel] win32-mmap test failures
In-Reply-To: <7524A45A1A5B264FA4809E2156496CFB023D3202@ITOMAE2KM01.AD.QINTRA.COM>
References: <7524A45A1A5B264FA4809E2156496CFB023D31FF@ITOMAE2KM01.AD.QINTRA.COM><7524A45A1A5B264FA4809E2156496CFB023D3201@ITOMAE2KM01.AD.QINTRA.COM><aea95bc50804250920v6776e1dcme00956641910a58@mail.gmail.com>
	<7524A45A1A5B264FA4809E2156496CFB023D3202@ITOMAE2KM01.AD.QINTRA.COM>
Message-ID: <7524A45A1A5B264FA4809E2156496CFB023D3203@ITOMAE2KM01.AD.QINTRA.COM>

Yeah, it's a scope resolution bug of some sort. Using "::String" works.
I'll fix that in the next windows-pr release.
 
Dan
 
PS - Sorry for the top posting (Outlook is being a pain)


________________________________

	From: win32utils-devel-bounces at rubyforge.org
[mailto:win32utils-devel-bounces at rubyforge.org] On Behalf Of Berger,
Daniel
	Sent: Friday, April 25, 2008 10:37 AM
	To: Development and ideas for win32utils projects
	Subject: Re: [Win32utils-devel] win32-mmap test failures
	
	
	Thanks.
	 
	The only thing I can figure is that it's some sort of name
collision with Windows::MSVCRT::String, but I can't see how right off
the bat.
	 
	Dan


________________________________

		From: win32utils-devel-bounces at rubyforge.org
[mailto:win32utils-devel-bounces at rubyforge.org] On Behalf Of Heesob Park
		Sent: Friday, April 25, 2008 10:20 AM
		To: Development and ideas for win32utils projects
		Subject: Re: [Win32utils-devel] win32-mmap test failures
		
		
		Hi,
		
		
		2008/4/26, Berger, Daniel <Daniel.Berger at qwest.com>: 


			> -----Original Message-----
			> From: win32utils-devel-bounces at rubyforge.org
			>
[mailto:win32utils-devel-bounces at rubyforge.org] On Behalf Of
			> Berger, Daniel
			> Sent: Friday, April 25, 2008 9:31 AM
			> To: Development and ideas for win32utils
projects
			> Subject: [Win32utils-devel] win32-mmap test
failures
			>
			> Hi all,
			>
			> This is odd. It looks to me like we have all 4
variations of
			> memcpy set in Windows::MSVCRT::Buffer but
somehow this one
			> creeps up. I actually noticed it in a few
cases with that
			> earlier mmap sub/replace example, but I wasn't
sure what was
			> happening.
			
			<snip>
			
			This is baffling. Somehow it's failing the
src.is_a?(String) check on
			the marshalled source on line 31 of buffer.rb,
even though it definitely
			_is_ a String. I cannot duplicate this in a
standalone script.
			
			Ideas?
			
			Thanks,
			
			Dan


		If you modify src.is_a?(String)  to
src.class.to_s=="String", it will work.
		But I don't know what it means.
		 
		Regards,

		Park Heesob
		 

		 



	This communication is the property of Qwest and may contain
confidential or privileged information. Unauthorized use of this
communication is strictly prohibited and may be unlawful. If you have
received this communication in error, please immediately notify the
sender by reply e-mail and destroy all copies of the communication and
any attachments. 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://rubyforge.org/pipermail/win32utils-devel/attachments/20080425/0fdaf13a/attachment.html>

From phasis at gmail.com  Mon Apr 28 06:53:26 2008
From: phasis at gmail.com (Heesob Park)
Date: Mon, 28 Apr 2008 19:53:26 +0900
Subject: [Win32utils-devel] rb_w32_select function patch
Message-ID: <aea95bc50804280353j66be9217ne96343322e60fbad@mail.gmail.com>

Hi, all

As you know, rb_w32_select don't work well with standard input.
Here is test code:

t = Thread.new {
   while true
     puts "printing a line"
     sleep 2
   end
 }

gets
t.exit
puts "exiting"

Following is the first version of patch code of rb_w32_select  function in
win32.c
It is inspired by cygwin select implementation code and adopted some code
from it.
The basic idea is creating thread for each file descriptor and wait events.

What's your thought about it?

 Regards,

Park Heesob


/*========================================================================================*/

static HANDLE main_select_event=NULL;
static HANDLE *select_thread_list=NULL;
static struct {
  int vk;
  int val[4];
} keytable[]  = {
       /* N S C A */
  {VK_LEFT, {1,1,1,1}},
  {VK_RIGHT, {1,1,1,1}},
  {VK_UP, {1,1,1,1}},
  {VK_DOWN, {1,1,1,1}},
  {VK_PRIOR, {1,1,1,1}},
  {VK_NEXT, {1,1,1,1}},
  {VK_HOME, {1,1,1,1}},
  {VK_END, {1,1,1,1}},
  {VK_INSERT, {1,1,1,1}},
  {VK_DELETE, {1,1,1,1}},
  {VK_F1, {1,1,0,0}},
  {VK_F2, {1,1,0,0}},
  {VK_F3, {1,1,0,0}},
  {VK_F4, {1,1,0,0}},
  {VK_F5, {1,1,0,0}},
  {VK_F6, {1,1,1,0}},
  {VK_F7, {1,1,0,0}},
  {VK_F8, {1,1,0,0}},
  {VK_F9, {1,1,0,0}},
  {VK_F10, {1,1,0,0}},
  {VK_F11, {1,0,0,0}},
  {VK_F12, {1,0,0,0}},
  {VK_NUMPAD5, {1,0,0,0}},
  {VK_CLEAR, {1,0,0,0}},
  {'6',  {0,0,1,0}},
  {0,  {0,0,0,0}}
};
static int
is_non_ascii_key(INPUT_RECORD* irec)
{
#define NORMAL  0
#define SHIFT 1
#define CONTROL 2
#define ALT 3
  int modifier_index = NORMAL;
  int i;
  if (irec->Event.KeyEvent.dwControlKeyState & SHIFT_PRESSED)
    modifier_index = SHIFT;
  else if (irec->Event.KeyEvent.dwControlKeyState &
  (LEFT_CTRL_PRESSED | RIGHT_CTRL_PRESSED))
    modifier_index = CONTROL;
  else if (irec->Event.KeyEvent.dwControlKeyState &
  (LEFT_ALT_PRESSED | RIGHT_ALT_PRESSED))
    modifier_index = ALT;
  for (i = 0; keytable[i].vk; i++)
    if (irec->Event.KeyEvent.wVirtualKeyCode == keytable[i].vk)
      return keytable[i].val[modifier_index];

  if (irec->Event.KeyEvent.uChar.AsciiChar) {
   return 1;
  }
  return 0;
}
static DWORD WINAPI
select_read_thread(PVOID argp)
{
    HANDLE fd_handle = (HANDLE)argp;
    DWORD ret = 1;
    DWORD mode = 0;

    mode = 0;
    GetConsoleMode(fd_handle,&mode);
    if(mode!=0) {
     // RealConsole Input
     INPUT_RECORD irec;
     DWORD events_read=0;
        for(;;) {
         ret = WaitForSingleObject(main_select_event,0);
         if(ret!=WAIT_TIMEOUT) break;
         if(!PeekConsoleInput(fd_handle,&irec,1,&events_read))
          break;
         if (irec.EventType == KEY_EVENT && irec.Event.KeyEvent.bKeyDown &&
   (irec.Event.KeyEvent.uChar.AsciiChar || is_non_ascii_key(&irec))) {
           return ret;
         }
         if(events_read>0)
            ReadConsoleInput(fd_handle, &irec, 1, &events_read);
        }
        return ret;
    }
    return ret;
}
static DWORD WINAPI
select_write_thread(PVOID argp)
{
    HANDLE fd_handle = (HANDLE)argp;
    DWORD ret = 1;
    DWORD mode = 0;

    GetConsoleMode(fd_handle,&mode);
    if(mode!=0) {
        // RealConsole Output
        return ret;
    }
    return ret;
}
long
rb_w32_select (int nfds, fd_set *rd, fd_set *wr, fd_set *ex,
        struct timeval *timeout)
{
    long r;
    fd_set file_rd;
    fd_set file_wr;
#ifdef USE_INTERRUPT_WINSOCK
    fd_set trap;
#endif /* USE_INTERRUPT_WINSOCK */
    int file_nfds;
    DWORD val;
    int fd;
    int i;
    if (!NtSocketsInitialized) {
 StartSockets();
    }
    r = 0;
    if (rd && rd->fd_count > r) r = rd->fd_count;
    if (wr && wr->fd_count > r) r = wr->fd_count;
    if (ex && ex->fd_count > r) r = ex->fd_count;
    if (nfds > r) nfds = r;
    if (nfds == 0 && timeout) {
 Sleep(timeout->tv_sec * 1000 + timeout->tv_usec / 1000);
 return 0;
    }
    file_nfds = extract_file_fd(rd, &file_rd);
    file_nfds += extract_file_fd(wr, &file_wr);
    if (file_nfds)
    {
        main_select_event =
CreateEvent(NULL,TRUE,FALSE,"main_select_event");
        if(main_select_event == NULL)
        {
   printf("CreateEvent failed (%d)\n", GetLastError());
   return -1;
        }
        select_thread_list = ALLOCA_N(HANDLE, file_nfds+1);
        for(i=0; i<file_nfds; i++)
        {
          if(i<file_rd.fd_count) {
            select_thread_list[i] = CreateThread(NULL,0,select_read_thread,
             (PVOID)file_rd.fd_array[i],0,&val);
          }
          else {
            select_thread_list[i] = CreateThread(NULL,0,select_write_thread,
             (PVOID)file_wr.fd_array[i-file_rd.fd_count],0,&val);
          }

          if (select_thread_list[i] == NULL)
          {
            printf("CreateThread failed (%d)\n", GetLastError());
            return -1;
          }
        }
        select_thread_list[file_nfds] = interrupted_event;
        r = WaitForMultipleObjects(file_nfds+1,select_thread_list,FALSE,
         timeout->tv_sec * 1000 + timeout->tv_usec / 1000);
        if (!SetEvent(main_select_event) )
        {
          printf("SetEvent failed (%d)\n", GetLastError());
          return -1;
        }
        CloseHandle(main_select_event);
        if(r==WAIT_TIMEOUT)
         return 0;
 // assume normal files are always readable/writable
 // fake read/write fd_set and return value
 if (rd) *rd = file_rd;
 if (wr) *wr = file_wr;
 return file_nfds;
    }
#if USE_INTERRUPT_WINSOCK
    if (ex)
 trap = *ex;
    else
 trap.fd_count = 0;
    if (trap.fd_count < FD_SETSIZE)
 trap.fd_array[trap.fd_count++] = (SOCKET)interrupted_event;
    // else unable to catch interrupt.
    ex = &trap;
#endif /* USE_INTERRUPT_WINSOCK */
    RUBY_CRITICAL({
 r = select(nfds, rd, wr, ex, timeout);
 if (r == SOCKET_ERROR) {
     errno = map_errno(WSAGetLastError());
 }
    });
    return r;
}
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://rubyforge.org/pipermail/win32utils-devel/attachments/20080428/a9adc2e0/attachment-0001.html>

From luislavena at gmail.com  Mon Apr 28 06:56:22 2008
From: luislavena at gmail.com (Luis Lavena)
Date: Mon, 28 Apr 2008 07:56:22 -0300
Subject: [Win32utils-devel] rb_w32_select function patch
In-Reply-To: <aea95bc50804280353j66be9217ne96343322e60fbad@mail.gmail.com>
References: <aea95bc50804280353j66be9217ne96343322e60fbad@mail.gmail.com>
Message-ID: <71166b3b0804280356k1d34d337x162b8e04944e8725@mail.gmail.com>

On Mon, Apr 28, 2008 at 7:53 AM, Heesob Park <phasis at gmail.com> wrote:
> Hi, all
>
> As you know, rb_w32_select don't work well with standard input.
> Here is test code:
>
> t = Thread.new {
>    while true
>      puts "printing a line"
>      sleep 2
>    end
>  }
>
> gets
> t.exit
> puts "exiting"
>
>
> Following is the first version of patch code of rb_w32_select  function in
> win32.c
> It is inspired by cygwin select implementation code and adopted some code
> from it.
> The basic idea is creating thread for each file descriptor and wait events.
>
> What's your thought about it?
>

Park,

This replacement is compatible with ruby_1_8 branch or just for
Sapphire? I'll love get this included in 1.8 to solve issues regarding
One-Click Installer for the time being.

>
> Regards,
>
> Park Heesob
>

Thank you for your work, tried to apply your older patch about this
and failed due changes introduced by years in ruby code.

Regards,
-- 
Luis Lavena
Multimedia systems
-
Human beings, who are almost unique in having the ability to learn from
the experience of others, are also remarkable for their apparent
disinclination to do so.
Douglas Adams

From phasis at gmail.com  Mon Apr 28 08:45:23 2008
From: phasis at gmail.com (Park Heesob)
Date: Mon, 28 Apr 2008 21:45:23 +0900
Subject: [Win32utils-devel] rb_w32_select function patch
References: <aea95bc50804280353j66be9217ne96343322e60fbad@mail.gmail.com>
	<71166b3b0804280356k1d34d337x162b8e04944e8725@mail.gmail.com>
Message-ID: <000d01c8a92d$bb37d500$9a7ba8c0@mycomputer>

Hi,
----- Original Message ----- 
From: "Luis Lavena" <luislavena at gmail.com>
To: "Development and ideas for win32utils projects" <win32utils-devel at rubyforge.org>
Sent: Monday, April 28, 2008 7:56 PM
Subject: Re: [Win32utils-devel] rb_w32_select function patch


> On Mon, Apr 28, 2008 at 7:53 AM, Heesob Park <phasis at gmail.com> wrote:
>> Hi, all
>>
>> As you know, rb_w32_select don't work well with standard input.
>> Here is test code:
>>
>> t = Thread.new {
>>    while true
>>      puts "printing a line"
>>      sleep 2
>>    end
>>  }
>>
>> gets
>> t.exit
>> puts "exiting"
>>
>>
>> Following is the first version of patch code of rb_w32_select  function in
>> win32.c
>> It is inspired by cygwin select implementation code and adopted some code
>> from it.
>> The basic idea is creating thread for each file descriptor and wait events.
>>
>> What's your thought about it?
>>
> 
> Park,
> 
> This replacement is compatible with ruby_1_8 branch or just for
> Sapphire? I'll love get this included in 1.8 to solve issues regarding
> One-Click Installer for the time being.
>
I tested code with ruby 1.8.6 p-114  compiled with VC++6.0 and VC++9.0.

Regards,

Park Heesob

From luislavena at gmail.com  Mon Apr 28 11:30:51 2008
From: luislavena at gmail.com (Luis Lavena)
Date: Mon, 28 Apr 2008 12:30:51 -0300
Subject: [Win32utils-devel] rb_w32_select function patch
In-Reply-To: <000d01c8a92d$bb37d500$9a7ba8c0@mycomputer>
References: <aea95bc50804280353j66be9217ne96343322e60fbad@mail.gmail.com>
	<71166b3b0804280356k1d34d337x162b8e04944e8725@mail.gmail.com>
	<000d01c8a92d$bb37d500$9a7ba8c0@mycomputer>
Message-ID: <71166b3b0804280830x38ddc02cxeb3a55e62f596dc4@mail.gmail.com>

On Mon, Apr 28, 2008 at 9:45 AM, Park Heesob <phasis at gmail.com> wrote:
> Hi,
>
>
> ----- Original Message -----
>  From: "Luis Lavena" <luislavena at gmail.com>
>  To: "Development and ideas for win32utils projects" <win32utils-devel at rubyforge.org>
>  Sent: Monday, April 28, 2008 7:56 PM
>  Subject: Re: [Win32utils-devel] rb_w32_select function patch
>
>  >
>  > Park,
>  >
>  > This replacement is compatible with ruby_1_8 branch or just for
>  > Sapphire? I'll love get this included in 1.8 to solve issues regarding
>  > One-Click Installer for the time being.
>  >
>  I tested code with ruby 1.8.6 p-114  compiled with VC++6.0 and VC++9.0.
>

Good thing to know. I'll check with GCC build of One-Click.

>  Regards,
>

Thank you again,

-- 
Luis Lavena
Multimedia systems
-
Human beings, who are almost unique in having the ability to learn from
the experience of others, are also remarkable for their apparent
disinclination to do so.
Douglas Adams

From djberg96 at gmail.com  Mon Apr 28 21:29:47 2008
From: djberg96 at gmail.com (Daniel Berger)
Date: Mon, 28 Apr 2008 19:29:47 -0600
Subject: [Win32utils-devel] rb_w32_select function patch
In-Reply-To: <aea95bc50804280353j66be9217ne96343322e60fbad@mail.gmail.com>
References: <aea95bc50804280353j66be9217ne96343322e60fbad@mail.gmail.com>
Message-ID: <48167A0B.2060608@gmail.com>

Heesob Park wrote:
> Hi, all
>  
> As you know, rb_w32_select don't work well with standard input.
> Here is test code:
> 
> t = Thread.new {
>    while true
>      puts "printing a line"
>      sleep 2
>    end
>  }
> 
> gets
> t.exit
> puts "exiting"
>  
> Following is the first version of patch code of rb_w32_select  function 
> in win32.c
> It is inspired by cygwin select implementation code and adopted some 
> code from it.
> The basic idea is creating thread for each file descriptor and wait events.
>  
> What's your thought about it?

I applied the patch and then tried to run the thread tests in my 
ruby_test test suite against Ruby 1.8.6 p114 on Windows XP, built with 
VC++ 8.

C:\Documents and Settings\djberge\workspace\ruby_test>ruby test_thread.rb
Loaded suite test_thread
Started
......................................................C:/rubyVC8/lib/ruby/1.8/test/unit/ui/console/testrunner.
rb:113: [BUG] Segmentation fault
ruby 1.8.6 (2008-03-03) [i386-mswin32_80]


This application has requested the Runtime to terminate it in an unusual 
way.
Please contact the application's support team for more information.

Where "test_thread.rb" is just:

# test_thread.rb
$LOAD_PATH.unshift File.join(Dir.pwd, 'lib')

Dir['test/core/thread/**/*.rb'].each{ |f|
    begin
       require f
    rescue LoadError
       next
    end
}

Or you can just run 'rake test_thread' if you have Rake installed.

You can grab "ruby_test" out of the "shards" CVS repo.

  cvs -d :pserver:anonymous at rubyforge.org:/var/cvs/shards login
cvs -d :pserver:anonymous at rubyforge.org:/var/cvs/shards checkout ruby_test

Is it possible I applied the patch wrong? Do you happen to have a 
unified diff, i.e. "diff -u" output?

Regards,

Dan


From djberg96 at gmail.com  Mon Apr 28 21:43:56 2008
From: djberg96 at gmail.com (Daniel Berger)
Date: Mon, 28 Apr 2008 19:43:56 -0600
Subject: [Win32utils-devel] win32-changejournal for the one-click
Message-ID: <48167D5C.3030907@gmail.com>

Hi,

Anyone remember how we got win32-changejournal built so that the 
one-click installer could use it? When I build the latest version with 
VC++ 8 and try to use the shared object with the one-click installer, it 
dies with with a DLL.

Yet the 0.3.1 shared object works ok. I don't remember how we pulled 
that off now, because I can't build the source for win32-changejournal 
with VC++ 6.

Thanks,

Dan

From phasis at gmail.com  Tue Apr 29 01:53:07 2008
From: phasis at gmail.com (Heesob Park)
Date: Tue, 29 Apr 2008 14:53:07 +0900
Subject: [Win32utils-devel] rb_w32_select function patch
In-Reply-To: <48167A0B.2060608@gmail.com>
References: <aea95bc50804280353j66be9217ne96343322e60fbad@mail.gmail.com>
	<48167A0B.2060608@gmail.com>
Message-ID: <aea95bc50804282253x14254bfdv8b041438aebc5153@mail.gmail.com>

Hi,

2008/4/29 Daniel Berger <djberg96 at gmail.com>:
> Heesob Park wrote:
> > Hi, all
> >  As you know, rb_w32_select don't work well with standard input.
> > Here is test code:
> >
> > t = Thread.new {
> >   while true
> >     puts "printing a line"
> >     sleep 2
> >   end
> >  }
> >
> > gets
> > t.exit
> > puts "exiting"
> >  Following is the first version of patch code of rb_w32_select  function
> in win32.c
> > It is inspired by cygwin select implementation code and adopted some code
> from it.
> > The basic idea is creating thread for each file descriptor and wait
> events.
> >  What's your thought about it?
> >
>
> I applied the patch and then tried to run the thread tests in my ruby_test
> test suite against Ruby 1.8.6 p114 on Windows XP, built with VC++ 8.
>
> C:\Documents and Settings\djberge\workspace\ruby_test>ruby test_thread.rb
> Loaded suite test_thread
> Started
> ......................................................C:/rubyVC8/lib/ruby/1.8/test/unit/ui/console/testrunner.
> rb:113: [BUG] Segmentation fault
> ruby 1.8.6 (2008-03-03) [i386-mswin32_80]
>
>
> This application has requested the Runtime to terminate it in an unusual
> way.
> Please contact the application's support team for more information.
>

I missed timeval NULL check for sleep without argument.

>
> Is it possible I applied the patch wrong? Do you happen to have a unified
> diff, i.e. "diff -u" output?
>
I added some safe check code.
Here is diff -u output:

--- win32.c.org	2008-04-29 14:44:41.000000000 +0900
+++ win32.c	2008-04-29 14:32:59.000000000 +0900
@@ -2046,6 +2046,117 @@
     return fileset->fd_count;
 }

+static HANDLE main_select_event=NULL;
+static HANDLE *select_thread_list=NULL;
+
+static struct {
+  int vk;
+  int val[4];
+} keytable[]  = {
+	      /* N S C A */
+  {VK_LEFT,	{1,1,1,1}},
+  {VK_RIGHT,	{1,1,1,1}},
+  {VK_UP,	{1,1,1,1}},
+  {VK_DOWN,	{1,1,1,1}},
+  {VK_PRIOR,	{1,1,1,1}},
+  {VK_NEXT,	{1,1,1,1}},
+  {VK_HOME,	{1,1,1,1}},
+  {VK_END,	{1,1,1,1}},
+  {VK_INSERT,	{1,1,1,1}},
+  {VK_DELETE,	{1,1,1,1}},
+  {VK_F1,	{1,1,0,0}},
+  {VK_F2,	{1,1,0,0}},
+  {VK_F3,	{1,1,0,0}},
+  {VK_F4,	{1,1,0,0}},
+  {VK_F5,	{1,1,0,0}},
+  {VK_F6,	{1,1,1,0}},
+  {VK_F7,	{1,1,0,0}},
+  {VK_F8,	{1,1,0,0}},
+  {VK_F9,	{1,1,0,0}},
+  {VK_F10,	{1,1,0,0}},
+  {VK_F11,	{1,0,0,0}},
+  {VK_F12,	{1,0,0,0}},
+  {VK_NUMPAD5,	{1,0,0,0}},
+  {VK_CLEAR,	{1,0,0,0}},
+  {'6',		{0,0,1,0}},
+  {0,		{0,0,0,0}}
+};
+
+static int
+is_non_ascii_key(INPUT_RECORD* irec)
+{
+#define NORMAL  0
+#define SHIFT	1
+#define CONTROL	2
+#define ALT	3
+  int modifier_index = NORMAL;
+  int i;
+
+  if (irec->Event.KeyEvent.dwControlKeyState & SHIFT_PRESSED)
+    modifier_index = SHIFT;
+  else if (irec->Event.KeyEvent.dwControlKeyState &
+		(LEFT_CTRL_PRESSED | RIGHT_CTRL_PRESSED))
+    modifier_index = CONTROL;
+  else if (irec->Event.KeyEvent.dwControlKeyState &
+		(LEFT_ALT_PRESSED | RIGHT_ALT_PRESSED))
+    modifier_index = ALT;
+
+  for (i = 0; keytable[i].vk; i++)
+    if (irec->Event.KeyEvent.wVirtualKeyCode == keytable[i].vk)
+      return keytable[i].val[modifier_index];
+
+  if (irec->Event.KeyEvent.uChar.AsciiChar) {
+  	return 1;
+  }
+  return 0; 	
+}
+
+static DWORD WINAPI
+select_read_thread(PVOID argp)
+{
+    HANDLE fd_handle = (HANDLE)argp;
+    DWORD ret = 1;
+    DWORD mode = 0;
+
+    mode = 0;
+    GetConsoleMode(fd_handle,&mode);
+    if(mode!=0) {
+    	// RealConsole Input
+    	INPUT_RECORD irec;
+    	DWORD events_read=0;
+
+        for(;;) {
+        	ret = WaitForSingleObject(main_select_event,0);
+        	if(ret!=WAIT_TIMEOUT) break;
+        	if(!PeekConsoleInput(fd_handle,&irec,1,&events_read))
+        		break;
+        	if (irec.EventType == KEY_EVENT && irec.Event.KeyEvent.bKeyDown &&
+			(irec.Event.KeyEvent.uChar.AsciiChar || is_non_ascii_key(&irec))) {
+	       			return ret;
+        	}
+        	if(events_read>0)
+       		   ReadConsoleInput(fd_handle, &irec, 1, &events_read);
+        }
+        return ret;	
+    }
+    return ret;
+}
+
+static DWORD WINAPI
+select_write_thread(PVOID argp)
+{
+    HANDLE fd_handle = (HANDLE)argp;
+    DWORD ret = 1;
+    DWORD mode = 0;
+
+    GetConsoleMode(fd_handle,&mode);
+    if(mode!=0) {
+        // RealConsole Output
+        return ret;	
+    }
+    return ret;
+}
+
 long
 rb_w32_select (int nfds, fd_set *rd, fd_set *wr, fd_set *ex,
 	       struct timeval *timeout)
@@ -2057,6 +2168,10 @@
     fd_set trap;
 #endif /* USE_INTERRUPT_WINSOCK */
     int file_nfds;
+    DWORD val;
+    int fd;
+    int i;
+    DWORD exit_code;

     if (!NtSocketsInitialized) {
 	StartSockets();
@@ -2074,6 +2189,55 @@
     file_nfds += extract_file_fd(wr, &file_wr);
     if (file_nfds)
     {
+        main_select_event = CreateEvent(NULL,TRUE,FALSE,"main_select_event");
+        if(main_select_event == NULL)
+        {
+	  fprintf(stderr,"CreateEvent failed (%d)\n", GetLastError());
+	  return -1;
+        }
+        select_thread_list = malloc(sizeof(HANDLE)*(file_nfds+1));
+
+        for(i=0; i<file_nfds; i++)
+        {
+          if(i<file_rd.fd_count) {
+            select_thread_list[i] = CreateThread(NULL,0,select_read_thread,
+            	(PVOID)file_rd.fd_array[i],0,&val);
+          }
+          else {
+            select_thread_list[i] = CreateThread(NULL,0,select_write_thread,
+            	(PVOID)file_wr.fd_array[i-file_rd.fd_count],0,&val);
+          }          	
+          if (select_thread_list[i] == NULL)
+          {
+            printf("CreateThread failed (%d)\n", GetLastError());
+            return -1;
+          }
+        }
+
+        select_thread_list[file_nfds] = interrupted_event;
+        if(timeout)
+            r = WaitForMultipleObjects(file_nfds+1,select_thread_list,FALSE,
+        	timeout->tv_sec * 1000 + timeout->tv_usec / 1000);   	        	
+        else
+            r =
WaitForMultipleObjects(file_nfds+1,select_thread_list,FALSE,INFINITE);
+        if (!SetEvent(main_select_event) )
+        {
+          printf("SetEvent failed (%d)\n", GetLastError());
+          return -1;
+        }
+       for(i=0;i<file_nfds;i++) {
+           exit_code = 0;
+           GetExitCodeThread(select_thread_list[i],&exit_code);
+           if(exit_code==STILL_ACTIVE) {
+          	TerminateThread(select_thread_list[i],1);
+           }
+           CloseHandle(select_thread_list[i]);
+       }
+       CloseHandle(main_select_event);
+       free(select_thread_list);
+       if(r==WAIT_TIMEOUT)
+        	return 0;
+
 	// assume normal files are always readable/writable
 	// fake read/write fd_set and return value
 	if (rd) *rd = file_rd;



Regards,

Park Heesob

From luislavena at gmail.com  Tue Apr 29 02:55:50 2008
From: luislavena at gmail.com (Luis Lavena)
Date: Tue, 29 Apr 2008 03:55:50 -0300
Subject: [Win32utils-devel] rb_w32_select function patch
In-Reply-To: <aea95bc50804282253x14254bfdv8b041438aebc5153@mail.gmail.com>
References: <aea95bc50804280353j66be9217ne96343322e60fbad@mail.gmail.com>
	<48167A0B.2060608@gmail.com>
	<aea95bc50804282253x14254bfdv8b041438aebc5153@mail.gmail.com>
Message-ID: <71166b3b0804282355n1f105392h8defbafd83c9152f@mail.gmail.com>

On Tue, Apr 29, 2008 at 2:53 AM, Heesob Park <phasis at gmail.com> wrote:
> Hi,
>
> [snip]

>  I added some safe check code.
>  Here is diff -u output:
>

Hello Park,

I just tested your patch for rb_w32_select and found, even gets() is
no longer blocking other threads from being executed, everything stops
when you start typing something...

Compared to other other platforms, the behavior differs, so maybe is
not the best solution...

I thought your previous patch (2005, 2006 maybe?) workaround that
issue. But don't remember quite right.

In any case, this looks good, one step closer :-)

(btw, platform i386-mingw32)

-- 
Luis Lavena
Multimedia systems
-
Human beings, who are almost unique in having the ability to learn from
the experience of others, are also remarkable for their apparent
disinclination to do so.
Douglas Adams

From phasis at gmail.com  Tue Apr 29 03:07:39 2008
From: phasis at gmail.com (Heesob Park)
Date: Tue, 29 Apr 2008 16:07:39 +0900
Subject: [Win32utils-devel] win32-changejournal for the one-click
In-Reply-To: <48167D5C.3030907@gmail.com>
References: <48167D5C.3030907@gmail.com>
Message-ID: <aea95bc50804290007nc03b8d4wed25a5c621e144b0@mail.gmail.com>

Hi,

2008/4/29 Daniel Berger <djberg96 at gmail.com>:
> Hi,
>
> Anyone remember how we got win32-changejournal built so that the one-click
> installer could use it? When I build the latest version with VC++ 8 and try
> to use the shared object with the one-click installer, it dies with with a
> DLL.
>
> Yet the 0.3.1 shared object works ok. I don't remember how we pulled that
> off now, because I can't build the source for win32-changejournal with VC++
> 6.
>
> Thanks,
>
> Dan
I guess you installed Platform SDK.

Regards,

Park Heesob

From phasis at gmail.com  Tue Apr 29 05:32:48 2008
From: phasis at gmail.com (Heesob Park)
Date: Tue, 29 Apr 2008 18:32:48 +0900
Subject: [Win32utils-devel] rb_w32_select function patch
In-Reply-To: <71166b3b0804282355n1f105392h8defbafd83c9152f@mail.gmail.com>
References: <aea95bc50804280353j66be9217ne96343322e60fbad@mail.gmail.com>
	<48167A0B.2060608@gmail.com>
	<aea95bc50804282253x14254bfdv8b041438aebc5153@mail.gmail.com>
	<71166b3b0804282355n1f105392h8defbafd83c9152f@mail.gmail.com>
Message-ID: <aea95bc50804290232p2e51461dh9120091a1f249d8d@mail.gmail.com>

Hi,

2008/4/29 Luis Lavena <luislavena at gmail.com>:
> On Tue, Apr 29, 2008 at 2:53 AM, Heesob Park <phasis at gmail.com> wrote:
> > Hi,
> >
> > [snip]
>
> >  I added some safe check code.
> >  Here is diff -u output:
> >
>
> Hello Park,
>
> I just tested your patch for rb_w32_select and found, even gets() is
> no longer blocking other threads from being executed, everything stops
> when you start typing something...
>
> Compared to other other platforms, the behavior differs, so maybe is
> not the best solution...
>
> I thought your previous patch (2005, 2006 maybe?) workaround that
> issue. But don't remember quite right.
>
That patch is not working well as you expected.

> In any case, this looks good, one step closer :-)
>
> (btw, platform i386-mingw32)
>
Here is a modified version for two step closer :)

--- win32.c.org	2008-04-29 14:44:41.000000000 +0900
+++ win32.c	2008-04-29 18:25:26.000000000 +0900
@@ -2046,6 +2046,65 @@
     return fileset->fd_count;
 }

+static HANDLE main_select_event=NULL;
+static HANDLE *select_thread_list=NULL;
+
+static DWORD WINAPI
+select_read_thread(PVOID argp)
+{
+    HANDLE fd_handle = (HANDLE)argp;
+    DWORD ret = 1;
+    DWORD mode = 0;
+
+    GetConsoleMode(fd_handle,&mode);
+    if(mode!=0) {
+    	// RealConsole Input
+    	DWORD num_events,num_events_read,pre_num=0;
+    	INPUT_RECORD *input_record;
+    	int i;
+
+        while(1) {
+            if(WaitForSingleObject(main_select_event,0)!=WAIT_TIMEOUT) break;
+	    GetNumberOfConsoleInputEvents(fd_handle, &num_events);
+	    if(pre_num != num_events) {
+	        input_record = (INPUT_RECORD
*)malloc(sizeof(INPUT_RECORD)*num_events);
+	        PeekConsoleInput(fd_handle, input_record, num_events,
&num_events_read);
+		for(i=0;i<num_events;i++) {
+	            if ((input_record[i].EventType == KEY_EVENT) &&
input_record[i].Event.KeyEvent.bKeyDown) {
+			if(input_record[i].Event.KeyEvent.uChar.AsciiChar==13) {   /*
carriage return */
+				free(input_record);
+	       			return ret;
+			}
+			else if(input_record[i].Event.KeyEvent.uChar.AsciiChar==26) { /*
^Z - end of file */
+				free(input_record);
+	       			return ret;
+			}
+	     	    }
+		}
+		free(input_record);
+		pre_num = num_events;
+           } 	
+       }
+       return ret;
+    }
+    return ret;
+}
+
+static DWORD WINAPI
+select_write_thread(PVOID argp)
+{
+    HANDLE fd_handle = (HANDLE)argp;
+    DWORD ret = 1;
+    DWORD mode = 0;
+
+    GetConsoleMode(fd_handle,&mode);
+    if(mode!=0) {
+        // RealConsole Output
+        return ret;
+    }
+    return ret;
+}
+
 long
 rb_w32_select (int nfds, fd_set *rd, fd_set *wr, fd_set *ex,
 	       struct timeval *timeout)
@@ -2057,6 +2116,10 @@
     fd_set trap;
 #endif /* USE_INTERRUPT_WINSOCK */
     int file_nfds;
+    DWORD val;
+    int fd;
+    int i;
+    DWORD exit_code;

     if (!NtSocketsInitialized) {
 	StartSockets();
@@ -2074,6 +2137,57 @@
     file_nfds += extract_file_fd(wr, &file_wr);
     if (file_nfds)
     {
+        main_select_event = CreateEvent(NULL,TRUE,FALSE,"main_select_event");
+        if(main_select_event == NULL)
+        {
+	  fprintf(stderr,"CreateEvent failed (%d)\n", GetLastError());
+	  return -1;
+        }
+        select_thread_list = malloc(sizeof(HANDLE)*(file_nfds+1));
+
+        for(i=0; i<file_nfds; i++)
+        {
+          if(i<file_rd.fd_count) {
+            select_thread_list[i] = CreateThread(NULL,0,select_read_thread,
+            	(PVOID)file_rd.fd_array[i],0,&val);
+          }
+          else {
+            select_thread_list[i] = CreateThread(NULL,0,select_write_thread,
+            	(PVOID)file_wr.fd_array[i-file_rd.fd_count],0,&val);
+          }
+          if (select_thread_list[i] == NULL)
+          {
+            printf("CreateThread failed (%d)\n", GetLastError());
+            return -1;
+          }
+        }
+
+        select_thread_list[file_nfds] = interrupted_event;
+        if(timeout)
+            r = WaitForMultipleObjects(file_nfds+1,select_thread_list,FALSE,
+        	timeout->tv_sec * 1000 + timeout->tv_usec / 1000);
+        else
+            r =
WaitForMultipleObjects(file_nfds+1,select_thread_list,FALSE,INFINITE);
+        if (!SetEvent(main_select_event) )
+        {
+          printf("SetEvent failed (%d)\n", GetLastError());
+          return -1;
+        }
+       	for(i=0;i<file_nfds;i++) {
+             while(1) {
+                  exit_code = 0;
+        	  GetExitCodeThread(select_thread_list[i],&exit_code);
+        	  if(exit_code!=STILL_ACTIVE) {
+        	      CloseHandle(select_thread_list[i]);
+        	      break;
+        	  }
+            }
+       	}
+        CloseHandle(main_select_event);
+        free(select_thread_list);
+        if(r==WAIT_TIMEOUT||r==file_nfds)
+        	return 0;
+
 	// assume normal files are always readable/writable
 	// fake read/write fd_set and return value
 	if (rd) *rd = file_rd;



Regards,

Park Heesob

From djberg96 at gmail.com  Tue Apr 29 08:24:37 2008
From: djberg96 at gmail.com (Daniel Berger)
Date: Tue, 29 Apr 2008 06:24:37 -0600
Subject: [Win32utils-devel] rb_w32_select function patch
In-Reply-To: <aea95bc50804290232p2e51461dh9120091a1f249d8d@mail.gmail.com>
References: <aea95bc50804280353j66be9217ne96343322e60fbad@mail.gmail.com>	<48167A0B.2060608@gmail.com>	<aea95bc50804282253x14254bfdv8b041438aebc5153@mail.gmail.com>	<71166b3b0804282355n1f105392h8defbafd83c9152f@mail.gmail.com>
	<aea95bc50804290232p2e51461dh9120091a1f249d8d@mail.gmail.com>
Message-ID: <48171385.1030304@gmail.com>

Heesob Park wrote:
> Hi,
> 
> 2008/4/29 Luis Lavena <luislavena at gmail.com>:
>> On Tue, Apr 29, 2008 at 2:53 AM, Heesob Park <phasis at gmail.com> wrote:
>>> Hi,
>>>
>>> [snip]
>>>  I added some safe check code.
>>>  Here is diff -u output:
>>>
>> Hello Park,
>>
>> I just tested your patch for rb_w32_select and found, even gets() is
>> no longer blocking other threads from being executed, everything stops
>> when you start typing something...
>>
>> Compared to other other platforms, the behavior differs, so maybe is
>> not the best solution...
>>
>> I thought your previous patch (2005, 2006 maybe?) workaround that
>> issue. But don't remember quite right.
>>
> That patch is not working well as you expected.
> 
>> In any case, this looks good, one step closer :-)
>>
>> (btw, platform i386-mingw32)
>>
> Here is a modified version for two step closer :)

<snip>

That works. Passes my (limited) thread test suite, too. :)

Awesome. Many thanks. I'll commit this in the CVS repo for Sapphire.

Dan

From djberg96 at gmail.com  Tue Apr 29 08:29:59 2008
From: djberg96 at gmail.com (Daniel Berger)
Date: Tue, 29 Apr 2008 06:29:59 -0600
Subject: [Win32utils-devel] win32-changejournal for the one-click
In-Reply-To: <aea95bc50804290007nc03b8d4wed25a5c621e144b0@mail.gmail.com>
References: <48167D5C.3030907@gmail.com>
	<aea95bc50804290007nc03b8d4wed25a5c621e144b0@mail.gmail.com>
Message-ID: <6037b70c0804290529m31b0a625yb2baff184814cf21@mail.gmail.com>

On Tue, Apr 29, 2008 at 1:07 AM, Heesob Park <phasis at gmail.com> wrote:
> Hi,
>
>  2008/4/29 Daniel Berger <djberg96 at gmail.com>:
>
>
> > Hi,
>  >
>  > Anyone remember how we got win32-changejournal built so that the one-click
>  > installer could use it? When I build the latest version with VC++ 8 and try
>  > to use the shared object with the one-click installer, it dies with with a
>  > DLL.
>  >
>  > Yet the 0.3.1 shared object works ok. I don't remember how we pulled that
>  > off now, because I can't build the source for win32-changejournal with VC++
>  > 6.
>  >
>  > Thanks,
>  >
>  > Dan
>  I guess you installed Platform SDK.

Got a shared object you can build for me so I can distribute this as a
gem that will work with the one-click?

Thanks,

Dan

From phasis at gmail.com  Tue Apr 29 09:25:54 2008
From: phasis at gmail.com (Park Heesob)
Date: Tue, 29 Apr 2008 22:25:54 +0900
Subject: [Win32utils-devel] win32-changejournal for the one-click
References: <48167D5C.3030907@gmail.com><aea95bc50804290007nc03b8d4wed25a5c621e144b0@mail.gmail.com>
	<6037b70c0804290529m31b0a625yb2baff184814cf21@mail.gmail.com>
Message-ID: <003e01c8a9fc$9006da00$9a7ba8c0@mycomputer>

Hi,
----- Original Message ----- 
From: "Daniel Berger" <djberg96 at gmail.com>
To: "Development and ideas for win32utils projects" <win32utils-devel at rubyforge.org>
Sent: Tuesday, April 29, 2008 9:29 PM
Subject: Re: [Win32utils-devel] win32-changejournal for the one-click


> On Tue, Apr 29, 2008 at 1:07 AM, Heesob Park <phasis at gmail.com> wrote:
>> Hi,
>>
>>  2008/4/29 Daniel Berger <djberg96 at gmail.com>:
>>
>>
>> > Hi,
>>  >
>>  > Anyone remember how we got win32-changejournal built so that the one-click
>>  > installer could use it? When I build the latest version with VC++ 8 and try
>>  > to use the shared object with the one-click installer, it dies with with a
>>  > DLL.
>>  >
>>  > Yet the 0.3.1 shared object works ok. I don't remember how we pulled that
>>  > off now, because I can't build the source for win32-changejournal with VC++
>>  > 6.
>>  >
>>  > Thanks,
>>  >
>>  > Dan
>>  I guess you installed Platform SDK.
> 
> Got a shared object you can build for me so I can distribute this as a
> gem that will work with the one-click?
> 
Here is Platform SDK compatible with VC++6.0

http://www.microsoft.com/msdownload/platformsdk/sdkupdate/psdk-full.htm

I attached a shared object for one-click. 

Regards,

Park Heesob
-------------- next part --------------
A non-text attachment was scrubbed...
Name: changejournal.so
Type: application/octet-stream
Size: 24655 bytes
Desc: not available
URL: <http://rubyforge.org/pipermail/win32utils-devel/attachments/20080429/b0512776/attachment-0001.so>

From Daniel.Berger at qwest.com  Tue Apr 29 12:14:24 2008
From: Daniel.Berger at qwest.com (Berger, Daniel)
Date: Tue, 29 Apr 2008 11:14:24 -0500
Subject: [Win32utils-devel] ZwXX functions
Message-ID: <7524A45A1A5B264FA4809E2156496CFB023D3217@ITOMAE2KM01.AD.QINTRA.COM>

Hi all,

I noticed that some of the Zw functions would be very handy to use.
While they're exported by NTDLL, it seems they're part of the Windows
Driver Kit, and those headers don't ship with Visual C++ by default. Or
do they? I don't have them with Visual C++ 8 anyway.

What I'm mainly after is the ZwQueryInformationFile for use with
File::Stat, so I can get information more easily from a handle. This
would be useful for a custom IO#stat implementation, where I'll have a
fileno but not a file name.

While it's technically part of the WDK, there's no reason it can't be
used for regular file operations as well.

Any suggestions?

Thanks,

Dan


This communication is the property of Qwest and may contain confidential or
privileged information. Unauthorized use of this communication is strictly 
prohibited and may be unlawful.  If you have received this communication 
in error, please immediately notify the sender by reply e-mail and destroy 
all copies of the communication and any attachments.

From phasis at gmail.com  Tue Apr 29 22:15:04 2008
From: phasis at gmail.com (Heesob Park)
Date: Wed, 30 Apr 2008 11:15:04 +0900
Subject: [Win32utils-devel] ZwXX functions
In-Reply-To: <7524A45A1A5B264FA4809E2156496CFB023D3217@ITOMAE2KM01.AD.QINTRA.COM>
References: <7524A45A1A5B264FA4809E2156496CFB023D3217@ITOMAE2KM01.AD.QINTRA.COM>
Message-ID: <aea95bc50804291915s3aa71665m932d423f1b80d56b@mail.gmail.com>

Hi,

2008/4/30 Berger, Daniel <Daniel.Berger at qwest.com>:
> Hi all,
>
> I noticed that some of the Zw functions would be very handy to use.
> While they're exported by NTDLL, it seems they're part of the Windows
> Driver Kit, and those headers don't ship with Visual C++ by default. Or
> do they? I don't have them with Visual C++ 8 anyway.
>
> What I'm mainly after is the ZwQueryInformationFile for use with
> File::Stat, so I can get information more easily from a handle. This
> would be useful for a custom IO#stat implementation, where I'll have a
> fileno but not a file name.
>
> While it's technically part of the WDK, there's no reason it can't be
> used for regular file operations as well.
>
> Any suggestions?
>
You cannot call ZwXX functions directly in application.
Instead, there is NtXX functions including NtQueryInformationFile.

Refer to http://www.codeguru.com/Cpp/W-P/system/processesmodules/article.php/c2827/
It has sample program and source code availabe.

Regards,

Park Heesob

From phasis at gmail.com  Wed Apr 30 04:08:35 2008
From: phasis at gmail.com (Heesob Park)
Date: Wed, 30 Apr 2008 17:08:35 +0900
Subject: [Win32utils-devel] rb_w32_select function patch
In-Reply-To: <48171385.1030304@gmail.com>
References: <aea95bc50804280353j66be9217ne96343322e60fbad@mail.gmail.com>
	<48167A0B.2060608@gmail.com>
	<aea95bc50804282253x14254bfdv8b041438aebc5153@mail.gmail.com>
	<71166b3b0804282355n1f105392h8defbafd83c9152f@mail.gmail.com>
	<aea95bc50804290232p2e51461dh9120091a1f249d8d@mail.gmail.com>
	<48171385.1030304@gmail.com>
Message-ID: <aea95bc50804300108x1f1bf7d4q13249ce7ae326469@mail.gmail.com>

Hi,

2008/4/29 Daniel Berger <djberg96 at gmail.com>:
>
> <snip>
>
> That works. Passes my (limited) thread test suite, too. :)
>
> Awesome. Many thanks. I'll commit this in the CVS repo for Sapphire.
>
> Dan
>
Now, I have made rb_w32_select function to work with pipe .
Here is a test code:

readPipe, writePipe = IO.pipe
t = Thread.new { while true; sleep 1; puts "got #{readPipe.readline}"; end }

i = 1
while true
  i += 1
  sleep 1
  puts "hello from main"
  if i>3
    writePipe.puts "test #{i}\n"
  end
end

t.join


Here is patch diff output:

--- win32.c.org	2008-04-29 14:44:41.000000000 +0900
+++ win32.c	2008-04-30 16:46:32.000000000 +0900
@@ -2046,6 +2050,91 @@
     return fileset->fd_count;
 }

+static HANDLE main_select_event=NULL;
+static HANDLE *select_thread_list=NULL;
+
+static DWORD WINAPI
+select_read_thread(PVOID argp)
+{
+    HANDLE fd_handle = (HANDLE)argp;
+    DWORD ret = 1;
+    DWORD mode = 0;
+
+    GetConsoleMode(fd_handle,&mode);
+    if(mode) {
+        /* Console Input */
+        DWORD num_events,num_events_read,pre_num=0;
+        INPUT_RECORD *input_record;
+        int i;
+
+        while(1) {
+            if(WaitForSingleObject(main_select_event,0)!=WAIT_TIMEOUT) break;
+            GetNumberOfConsoleInputEvents(fd_handle, &num_events);
+            if(pre_num != num_events) {
+                input_record = (INPUT_RECORD
*)malloc(sizeof(INPUT_RECORD)*num_events);
+                PeekConsoleInput(fd_handle, input_record, num_events,
&num_events_read);
+                for(i=0;i<num_events;i++) {
+                    if ((input_record[i].EventType == KEY_EVENT) &&
input_record[i].Event.KeyEvent.bKeyDown) {
+
if(input_record[i].Event.KeyEvent.uChar.AsciiChar==13) {   /* carriage
return */
+                            free(input_record);
+                            return ret;
+                        }
+                        else
if(input_record[i].Event.KeyEvent.uChar.AsciiChar==26) { /* ^Z - end
of file */
+                            free(input_record);
+                            return ret;
+                        }
+                    }
+                }
+                free(input_record);
+                pre_num = num_events;
+           }
+        }
+        return ret;
+    } else {
+        DWORD state;
+        ret = GetNamedPipeHandleState(fd_handle,&state,NULL,NULL,NULL,NULL,0);
+        if(ret) {
+            /* Pipe Input */
+            int bytes_avail=0;
+            while(1) {
+
if(WaitForSingleObject(main_select_event,0)!=WAIT_TIMEOUT) break;
+                ret = PeekNamedPipe(fd_handle,NULL,0,NULL,&bytes_avail,NULL);
+                if(bytes_avail>0) {
+                    return bytes_avail;
+              }
+            }
+        }
+        return ret;
+    }
+    return ret;
+}
+
+static DWORD WINAPI
+select_write_thread(PVOID argp)
+{
+    HANDLE fd_handle = (HANDLE)argp;
+    DWORD ret = 1;
+    DWORD mode = 0;
+
+    GetConsoleMode(fd_handle,&mode);
+    if(mode) {
+        /* Console Output */
+        return ret;
+    } else {
+        DWORD state;
+        ret = GetNamedPipeHandleState(fd_handle,&state,NULL,NULL,NULL,NULL,0);
+        if(ret) {
+            /* Pipe output */
+            int bytes_written=0;
+            ret = WriteFile(fd_handle,NULL,0,&bytes_written,NULL);
+            if(ret) {
+                return ret;
+            }
+        }
+    }
+    return ret;
+}
+
 long
 rb_w32_select (int nfds, fd_set *rd, fd_set *wr, fd_set *ex,
 	       struct timeval *timeout)
@@ -2074,6 +2163,80 @@
     file_nfds += extract_file_fd(wr, &file_wr);
     if (file_nfds)
     {
+        DWORD val,exit_code;
+        int i;
+        main_select_event = CreateEvent(NULL,TRUE,FALSE,NULL);
+        if(main_select_event == NULL)
+        {
+          printf("CreateEvent failed (%d)\n", GetLastError());
+          return -1;
+        }
+        select_thread_list = malloc(sizeof(HANDLE)*(file_nfds+1));
+
+        for(i=0; i<file_nfds; i++)
+        {
+          if(i<file_rd.fd_count) {
+            select_thread_list[i] = CreateThread(NULL,0,select_read_thread,
+                (PVOID)file_rd.fd_array[i],0,&val);
+          }
+          else {
+            select_thread_list[i] = CreateThread(NULL,0,select_write_thread,
+                (PVOID)file_wr.fd_array[i-file_rd.fd_count],0,&val);
+          }
+          if (select_thread_list[i] == NULL)
+          {
+            printf("CreateThread failed (%d)\n", GetLastError());
+            return -1;
+          }
+        }
+
+        select_thread_list[file_nfds] = interrupted_event;
+        if(timeout)
+            r = WaitForMultipleObjects(file_nfds+1,select_thread_list,FALSE,
+                timeout->tv_sec * 1000 + timeout->tv_usec / 1000);
+        else
+            r = WaitForMultipleObjects(file_nfds+1,select_thread_list,FALSE,
+                INFINITE);
+
+        if (!SetEvent(main_select_event) )
+        {
+          printf("SetEvent failed (%d)\n", GetLastError());
+          return -1;
+        }
+        /* thread cleanup */
+        for(i=0;i<file_nfds;i++) {
+            while(1) {
+                  exit_code = 0;
+                  GetExitCodeThread(select_thread_list[i],&exit_code);
+                  if(exit_code!=STILL_ACTIVE) {
+                      CloseHandle(select_thread_list[i]);
+                      break;
+                  }
+            }
+        }
+        free(select_thread_list);
+        CloseHandle(main_select_event);
+
+        if(r==WAIT_TIMEOUT||r==file_nfds) { /* timeout or interrupt */
+                FD_ZERO(&file_rd);
+                FD_ZERO(&file_wr);
+                *rd = file_rd;
+                *wr = file_wr;
+                return 0;
+        }
+        if(r<file_rd.fd_count) { /* read ready */
+                FD_ZERO(&file_wr);
+                *rd = file_rd;
+                *wr = file_wr;
+                return 1;
+        }
+        else { /* write ready */
+                FD_ZERO(&file_rd);
+                *rd = file_rd;
+                *wr = file_wr;
+                return 1;
+        }
+
 	// assume normal files are always readable/writable
 	// fake read/write fd_set and return value
 	if (rd) *rd = file_rd;


Regards,

Park Heesob

From Daniel.Berger at qwest.com  Wed Apr 30 10:08:54 2008
From: Daniel.Berger at qwest.com (Berger, Daniel)
Date: Wed, 30 Apr 2008 09:08:54 -0500
Subject: [Win32utils-devel] rb_w32_select function patch
In-Reply-To: <aea95bc50804300108x1f1bf7d4q13249ce7ae326469@mail.gmail.com>
References: <aea95bc50804280353j66be9217ne96343322e60fbad@mail.gmail.com><48167A0B.2060608@gmail.com><aea95bc50804282253x14254bfdv8b041438aebc5153@mail.gmail.com><71166b3b0804282355n1f105392h8defbafd83c9152f@mail.gmail.com><aea95bc50804290232p2e51461dh9120091a1f249d8d@mail.gmail.com><48171385.1030304@gmail.com>
	<aea95bc50804300108x1f1bf7d4q13249ce7ae326469@mail.gmail.com>
Message-ID: <7524A45A1A5B264FA4809E2156496CFB023D321B@ITOMAE2KM01.AD.QINTRA.COM>

 

> -----Original Message-----
> From: win32utils-devel-bounces at rubyforge.org 
> [mailto:win32utils-devel-bounces at rubyforge.org] On Behalf Of 
> Heesob Park
> Sent: Wednesday, April 30, 2008 2:09 AM
> To: Development and ideas for win32utils projects
> Subject: Re: [Win32utils-devel] rb_w32_select function patch
> 
> Hi,
> 
> 2008/4/29 Daniel Berger <djberg96 at gmail.com>:
> >
> > <snip>
> >
> > That works. Passes my (limited) thread test suite, too. :)
> >
> > Awesome. Many thanks. I'll commit this in the CVS repo for Sapphire.
> >
> > Dan
> >
> Now, I have made rb_w32_select function to work with pipe .
> Here is a test code:
> 
> readPipe, writePipe = IO.pipe
> t = Thread.new { while true; sleep 1; puts "got 
> #{readPipe.readline}"; end }
> 
> i = 1
> while true
>   i += 1
>   sleep 1
>   puts "hello from main"
>   if i>3
>     writePipe.puts "test #{i}\n"
>   end
> end
> 
> t.join

<snip>

Excellent, thanks!

With this in place I'm wondering how open3 could be implemented to work
on Windows without fork and without resorting to a C extension. Or if
IO.popen needs to be reworked. Any thoughts?

Thanks,

Dan


This communication is the property of Qwest and may contain confidential or
privileged information. Unauthorized use of this communication is strictly 
prohibited and may be unlawful.  If you have received this communication 
in error, please immediately notify the sender by reply e-mail and destroy 
all copies of the communication and any attachments.

From Daniel.Berger at qwest.com  Wed Apr 30 10:45:04 2008
From: Daniel.Berger at qwest.com (Berger, Daniel)
Date: Wed, 30 Apr 2008 09:45:04 -0500
Subject: [Win32utils-devel] rb_w32_select function patch
In-Reply-To: <aea95bc50804290232p2e51461dh9120091a1f249d8d@mail.gmail.com>
References: <aea95bc50804280353j66be9217ne96343322e60fbad@mail.gmail.com><48167A0B.2060608@gmail.com><aea95bc50804282253x14254bfdv8b041438aebc5153@mail.gmail.com><71166b3b0804282355n1f105392h8defbafd83c9152f@mail.gmail.com>
	<aea95bc50804290232p2e51461dh9120091a1f249d8d@mail.gmail.com>
Message-ID: <7524A45A1A5B264FA4809E2156496CFB023D321C@ITOMAE2KM01.AD.QINTRA.COM>

 

> -----Original Message-----
> From: win32utils-devel-bounces at rubyforge.org 
> [mailto:win32utils-devel-bounces at rubyforge.org] On Behalf Of 
> Heesob Park
> Sent: Tuesday, April 29, 2008 3:33 AM
> To: Development and ideas for win32utils projects
> Subject: Re: [Win32utils-devel] rb_w32_select function patch
> 
> Hi,
> 
> 2008/4/29 Luis Lavena <luislavena at gmail.com>:
> > On Tue, Apr 29, 2008 at 2:53 AM, Heesob Park 
> <phasis at gmail.com> wrote:
> > > Hi,
> > >
> > > [snip]
> >
> > >  I added some safe check code.
> > >  Here is diff -u output:
> > >
> >
> > Hello Park,
> >
> > I just tested your patch for rb_w32_select and found, even 
> gets() is 
> > no longer blocking other threads from being executed, 
> everything stops 
> > when you start typing something...
> >
> > Compared to other other platforms, the behavior differs, so 
> maybe is 
> > not the best solution...
> >
> > I thought your previous patch (2005, 2006 maybe?) workaround that 
> > issue. But don't remember quite right.
> >
> That patch is not working well as you expected.
> 
> > In any case, this looks good, one step closer :-)
> >
> > (btw, platform i386-mingw32)
> >
> Here is a modified version for two step closer :)

<snip>

The only difference I see between this and a Unix version is that
anything I type on the command line doesn't appear until I hit the Enter
key. On Unix (or with Jruby on Windows), I see whatever I type on the
console right away. I'm guessing it's a cmd.exe or native thread issue.
It's nothing big, but I thought I'd mention it.

Regards,

Dan


This communication is the property of Qwest and may contain confidential or
privileged information. Unauthorized use of this communication is strictly 
prohibited and may be unlawful.  If you have received this communication 
in error, please immediately notify the sender by reply e-mail and destroy 
all copies of the communication and any attachments.

From Daniel.Berger at qwest.com  Wed Apr 30 15:13:10 2008
From: Daniel.Berger at qwest.com (Berger, Daniel)
Date: Wed, 30 Apr 2008 14:13:10 -0500
Subject: [Win32utils-devel] Playing with NtQueryInformationFile
Message-ID: <7524A45A1A5B264FA4809E2156496CFB023D3222@ITOMAE2KM01.AD.QINTRA.COM>

Hi all,

I'm trying to get the allocation size of a file via a file handle
(rather than its name). The example below works for FileNameInformation
but I can't get it to work as expected for FileStandardInformation.

Here's some sample code:

# query_test.rb
require 'windows/handle'
require 'windows/error'
include Windows::Handle
include Windows::Error

NtQueryInformationFile = API.new('NtQueryInformationFile', 'LPPLL', 'L',
'ntdll')

# http://msdn.microsoft.com/en-us/library/cc232064.aspx
FileNameInformation     = 9
FileStandardInformation = 5
STATUS_SUCCESS          = 0

fh = File.open('test.txt', 'w')
fh.puts "hello"

handle = get_osfhandle(fh.fileno)

if handle == INVALID_HANDLE_VALUE
   puts "ERROR, get_osfhandle() : " + get_last_error
   fh.close rescue nil
   File.delete('test.txt')
   exit
end

# Excessive but harmless (?)
io_status_block  = 0.chr * 512
file_information = 0.chr * 512

status = NtQueryInformationFile.call(
   handle,
   io_status_block,
   file_information,
   file_information.size,
   # FileStandardInformation # Doesn't work as expected
   FileNameInformation       # But this does
)

if status != STATUS_SUCCESS
   puts "ERROR, NtQueryInformationFile() : #{status}"
   fh.close
   File.delete('test.txt')
   exit
end

# p file_information[0, 8].unpack('L')[0] # Zero ??? # Use with
FileStandardInformation

p file_information[0,4].unpack('L')[0] / 2 # Unicode
p file_information[4,file_information.length].tr("\000",'').strip

fh.close

# end query_test.rb

The above will print the file name (without the drive letter) as well as
the length of the file name, which varies depending on whatever path you
created 'test.txt' on.

However, if I try to replace "FileNameInformation" with
"FileStandardInformation" and inspect the first 8 bytes of
file_information (the AllocationSize), it's always zero. The only thing
I get out of the buffer at all is the NumberOfLinks, which is always 1.

What am I doing wrong?

Thanks,

Dan


This communication is the property of Qwest and may contain confidential or
privileged information. Unauthorized use of this communication is strictly 
prohibited and may be unlawful.  If you have received this communication 
in error, please immediately notify the sender by reply e-mail and destroy 
all copies of the communication and any attachments.

From phasis at gmail.com  Wed Apr 30 20:41:42 2008
From: phasis at gmail.com (Heesob Park)
Date: Thu, 1 May 2008 09:41:42 +0900
Subject: [Win32utils-devel] Playing with NtQueryInformationFile
In-Reply-To: <7524A45A1A5B264FA4809E2156496CFB023D3222@ITOMAE2KM01.AD.QINTRA.COM>
References: <7524A45A1A5B264FA4809E2156496CFB023D3222@ITOMAE2KM01.AD.QINTRA.COM>
Message-ID: <aea95bc50804301741r24627adan9cbca754f033c819@mail.gmail.com>

Hi,

2008/5/1 Berger, Daniel <Daniel.Berger at qwest.com>:
> Hi all,
>
> I'm trying to get the allocation size of a file via a file handle
> (rather than its name). The example below works for FileNameInformation
> but I can't get it to work as expected for FileStandardInformation.
>
> Here's some sample code:
>
> # query_test.rb
> require 'windows/handle'
> require 'windows/error'
> include Windows::Handle
> include Windows::Error
>
> NtQueryInformationFile = API.new('NtQueryInformationFile', 'LPPLL', 'L',
> 'ntdll')
>
> # http://msdn.microsoft.com/en-us/library/cc232064.aspx
> FileNameInformation     = 9
> FileStandardInformation = 5
> STATUS_SUCCESS          = 0
>
> fh = File.open('test.txt', 'w')
> fh.puts "hello"
>
> handle = get_osfhandle(fh.fileno)
>
> if handle == INVALID_HANDLE_VALUE
>   puts "ERROR, get_osfhandle() : " + get_last_error
>   fh.close rescue nil
>   File.delete('test.txt')
>   exit
> end
>
> # Excessive but harmless (?)
> io_status_block  = 0.chr * 512
> file_information = 0.chr * 512
>
> status = NtQueryInformationFile.call(
>   handle,
>   io_status_block,
>   file_information,
>   file_information.size,
>   # FileStandardInformation # Doesn't work as expected
>   FileNameInformation       # But this does
> )
>
> if status != STATUS_SUCCESS
>   puts "ERROR, NtQueryInformationFile() : #{status}"
>   fh.close
>   File.delete('test.txt')
>   exit
> end
>
> # p file_information[0, 8].unpack('L')[0] # Zero ??? # Use with
> FileStandardInformation
>
> p file_information[0,4].unpack('L')[0] / 2 # Unicode
> p file_information[4,file_information.length].tr("\000",'').strip
>
> fh.close
>
> # end query_test.rb
>
> The above will print the file name (without the drive letter) as well as
> the length of the file name, which varies depending on whatever path you
> created 'test.txt' on.
>
> However, if I try to replace "FileNameInformation" with
> "FileStandardInformation" and inspect the first 8 bytes of
> file_information (the AllocationSize), it's always zero. The only thing
> I get out of the buffer at all is the NumberOfLinks, which is always 1.
>
> What am I doing wrong?
>

Insert fh.flush atfer fh.puts "hello" must be required.
Following code will work for you:

fh = File.open('test.txt', 'w')
fh.puts "hello"
fh.flush
handle = get_osfhandle(fh.fileno)
io_status_block  = 0.chr * 8
file_information = 0.chr * 0x18
status = NtQueryInformationFile.call(
   handle,
   io_status_block,
   file_information,
   file_information.size,
   FileStandardInformation
)

For the actual buffer size, Reter to

http://undocumented.ntinternals.net/
http://undocumented.ntinternals.net/UserMode/Undocumented%20Functions/NT%20Objects/File/FILE_INFORMATION_CLASS.html

Regards,

Park Heesob