From djberg96 at gmail.com  Sat Mar  1 11:05:05 2008
From: djberg96 at gmail.com (Daniel Berger)
Date: Sat, 01 Mar 2008 09:05:05 -0700
Subject: [Win32utils-devel] Sapphire
Message-ID: <47C97EB1.6020006@gmail.com>

It has begun:

http://www.sapphire-lang.org

Regards,

Dan

From waynev at gmail.com  Sat Mar  1 13:04:19 2008
From: waynev at gmail.com (Wayne Vucenic)
Date: Sat, 1 Mar 2008 10:04:19 -0800
Subject: [Win32utils-devel] Sapphire
In-Reply-To: <47C97EB1.6020006@gmail.com>
References: <47C97EB1.6020006@gmail.com>
Message-ID: <88c9ce410803011004n267b8da1jeb83af1d0f3ebd5d@mail.gmail.com>

Hi Dan,

Congratulations!  I read through your docs, and this looks like a great
start!

Would you be up for removing "fork" from Kernel?  IMHO, the standard library
(and especially Kernel and Object) should only contain methods that have a
good implementation (and I don't mean Cygwin) across _all_ supported
platforms.

I think it's fine for "fork" to be in some platform-specific library (like
one named Posix), but the problem with it being in Kernel is that some
programmers use it because it's there, and then are surprised that they've
written non-portable code without realizing it.

Best regards,

Wayne

On Sat, Mar 1, 2008 at 8:05 AM, Daniel Berger <djberg96 at gmail.com> wrote:

> It has begun:
>
> http://www.sapphire-lang.org
>
> Regards,
>
> Dan
> _______________________________________________
> win32utils-devel mailing list
> win32utils-devel at rubyforge.org
> http://rubyforge.org/mailman/listinfo/win32utils-devel
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/win32utils-devel/attachments/20080301/447c817e/attachment.html 

From djberg96 at gmail.com  Sat Mar  1 16:39:50 2008
From: djberg96 at gmail.com (Daniel Berger)
Date: Sat, 01 Mar 2008 14:39:50 -0700
Subject: [Win32utils-devel] Sapphire
In-Reply-To: <88c9ce410803011004n267b8da1jeb83af1d0f3ebd5d@mail.gmail.com>
References: <47C97EB1.6020006@gmail.com>
	<88c9ce410803011004n267b8da1jeb83af1d0f3ebd5d@mail.gmail.com>
Message-ID: <47C9CD26.9050801@gmail.com>

Wayne Vucenic wrote:
> Hi Dan,
> 
> Congratulations!  I read through your docs, and this looks like a great 
> start!

Thanks!

> Would you be up for removing "fork" from Kernel?  IMHO, the standard 
> library (and especially Kernel and Object) should only contain methods 
> that have a good implementation (and I don't mean Cygwin) across _all_ 
> supported platforms.

I'm absolutely going to remove fork from Kernel. There will be 'unix' 
and 'windows' libraries with their own namespace. So, if you want 
Kernel#fork, you'll have to require 'unix/kernel' explicitly. There will 
  also be unix/file, windows/file, etc.

> I think it's fine for "fork" to be in some platform-specific library 
> (like one named Posix), but the problem with it being in Kernel is that 
> some programmers use it because it's there, and then are surprised that 
> they've written non-portable code without realizing it.

I agree. :)

BTW, I've setup a sapphire-core mailing list. Visit here for options:

http://sapphire-lang.org/mailman/listinfo/sapphire-core_sapphire-lang.org

I've also setup an IRC channel, #sapphire, on Freenode.

Regards,

Dan


From djberg96 at gmail.com  Thu Mar  6 12:42:58 2008
From: djberg96 at gmail.com (Daniel Berger)
Date: Thu, 06 Mar 2008 10:42:58 -0700
Subject: [Win32utils-devel] Sapphire registration
Message-ID: <47D02D22.6040008@gmail.com>

Hey all,

If you tried to register with the Sapphire project but were 
unsuccessful, please try again. I didn't have the email settings 
properly set, so I wasn't getting the approval notifications.

In any case, I've changed the project settings so that you should be 
able to login immediately after registration, without having to wait for 
admin approval.

Regards,

Dan

From phasis at gmail.com  Thu Mar  6 19:58:25 2008
From: phasis at gmail.com (Heesob Park)
Date: Fri, 7 Mar 2008 09:58:25 +0900
Subject: [Win32utils-devel] Sapphire registration
In-Reply-To: <47D02D22.6040008@gmail.com>
References: <47D02D22.6040008@gmail.com>
Message-ID: <aea95bc50803061658l47bcee3fg6966b2072988580b@mail.gmail.com>

Hi,

2008/3/7, Daniel Berger <djberg96 at gmail.com>:
>
> Hey all,
>
> If you tried to register with the Sapphire project but were
> unsuccessful, please try again. I didn't have the email settings
> properly set, so I wasn't getting the approval notifications.
>
> In any case, I've changed the project settings so that you should be
> able to login immediately after registration, without having to wait for
> admin approval.


In that case, please remove my pending account phasis68.
I cannot register again for the error message : "login and email has already
beed taken".

Regards,
>
> Dan



Regards,

Park Heesob
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/win32utils-devel/attachments/20080307/6c6860da/attachment.html 

From djberg96 at gmail.com  Fri Mar 21 13:30:13 2008
From: djberg96 at gmail.com (Daniel Berger)
Date: Fri, 21 Mar 2008 11:30:13 -0600
Subject: [Win32utils-devel] Fwd: Is select() untrustworthy on Windows when
	used with handles from popen3?
In-Reply-To: <006c01c8852a$8dd423b0$670aa8c0@wax>
References: <006c01c8852a$8dd423b0$670aa8c0@wax>
Message-ID: <6037b70c0803211030va512fa1s267dc89cf957fa15@mail.gmail.com>

Any ideas?

Dan


---------- Forwarded message ----------
From: Steve Shreeve <steve.shreeve at gmail.com>
Date: Thu, Mar 13, 2008 at 10:51 AM
Subject: Is select() untrustworthy on Windows when used with handles
from popen3?
To: djberg96 at gmail.com




Dan,

Thanks for your work on the win32util modules.

I've created an expect-like library that allows one to easily script
interactions with non-GUI applications. Everything works great when
connecting to the target system using a socket, or using PTY.spawn on
Linux. On Windows, I get quite a bit of grief. Here's an example that
has me confused:

d:\>irb
irb(main):001:0> require 'win32/open3'
=> true
irb(main):002:0> i,o,e = Open3.popen3('date')
=> [#<IO:0x2b42a4c>, #<IO:0x2b42a24>, #<IO:0x2b42a10>, 3164]
irb(main):003:0> o.sysread(70)
=> "The current date is: Thu 03/13/2008 \nEnter the new date: (mm-dd-yy) "
irb(main):004:0> select([o],nil,nil,1)
=> [[#<IO:0x2b42a24>], [], []]
irb(main):005:0> o.sysread(1)
# => here, it hangs... even though it *says* there's input to read ]==

Since we hang at this spot, I'm sorta hosed... how can I trust select()?

Steve

==

ps - I'd be happy to pay for your time to assist with this.

From Daniel.Berger at qwest.com  Mon Mar 24 11:32:04 2008
From: Daniel.Berger at qwest.com (Berger, Daniel)
Date: Mon, 24 Mar 2008 10:32:04 -0500
Subject: [Win32utils-devel] GetTempPath vs GetEnvironmentVariable vs getenv
Message-ID: <7524A45A1A5B264FA4809E2156496CFB023D30F2@ITOMAE2KM01.AD.QINTRA.COM>

Hi all,
 
I discovered some interesting quirks with regards to getting environment
variable information on Windows. It's because 'TMP' is considered a
system environment variable, which you can't change with
SetEnvironmentVariable() according to the docs.
 
I also noticed that GetTempPath() returns a full path if you set the
value to a relative path.
 
Here's a little program that demonstrates what I'm talking about:
 
# Compare getenv/putenv, GetEnvironmentVariable/SetEnvironmentVariable
require 'windows/api'
include Windows
 
GetTempPath = API.new('GetTempPath','LP','L')
 
GetEnvironmentVariable = API.new('GetEnvironmentVariable', 'PPL', 'L')
SetEnvironmentVariable = API.new('SetEnvironmentVariable', 'PP', 'B')
 
Getenv = API.new('getenv', 'P', 'P', 'msvcrt')
Putenv = API.new('_putenv', 'P', 'I', 'msvcrt')
 
# ENV
def get_env1(var)
   ENV[var]
end
 
# GetEnvironmentVariable
def get_env2(var)
   buf = 0.chr * 260
   GetEnvironmentVariable.call('TMP', buf, buf.length)
   buf.unpack('A*').first
end
  
# GetTempPath, which looks for 'TMP' first
def get_env3
   buf = 0.chr * 260
   GetTempPath.call(buf.length, buf)
   buf.unpack('A*').first
end
 
# getenv()
def get_env4(var)
   Getenv.call(var)
end
 
puts get_env1('TMP') # C:\DOCUME~1\djberge\LOCALS~1\Temp
puts get_env2('TMP') # C:\DOCUME~1\djberge\LOCALS~1\Temp
puts get_env3        # C:\DOCUME~1\djberge\LOCALS~1\Temp\
puts get_env4('TMP') # C:\DOCUME~1\djberge\LOCALS~1\Temp
 
# Set 'TMP' using Ruby's ENV[]=
ENV['TMP'] = "C:\\foo\\bar"
puts "=" * 40
 
puts get_env1('TMP') # C:\foo\bar
puts get_env2('TMP') # C:\foo\bar
puts get_env3        # C:\foo\bar\
puts get_env4('TMP') # C:\DOCUME~1\djberge\LOCALS~1\Temp - oops
 
# Set 'TMP' using SetEnvironmentVariable
SetEnvironmentVariable.call('TMP', "C:\\zap")
puts "=" * 40
 
puts get_env1('TMP') # C:\zap
puts get_env2('TMP') # C:\zap
puts get_env3        # C:\zap\
puts get_env4('TMP') # C:\DOCUME~1\djberge\LOCALS~1\Temp - oops
 
# Set 'TMP' using putenv
Putenv.call("TMP=C:\\blah")
puts "=" * 40
 
puts get_env1('TMP') # C:\blah
puts get_env2('TMP') # C:\blah
puts get_env3        # C:\blah\
puts get_env4('TMP') # C:\blah - ah, there we go
 
# Set 'TMP' to a relative path
Putenv.call("TMP=dan")
puts "=" * 40
 
puts get_env1('TMP') # dan
puts get_env2('TMP') # dan
puts get_env3        # C:\Documents and Settings\djberge\My
Documents\dan - full
puts get_env4('TMP') # dan

The odd thing about TMP is that my shell value doesn't match the
registry setting. The 'set' output shows:
 
TMP=C:\DOCUME~1\djberge\LOCALS~1\Temp
 
But the registry entry, under
HKEY_LOCAL_MACHINE\System\CurrentControlSet\Control\Session
Manager\Environment, is set to:
 
C:\Temp
 
Anyway, just thought I'd share.
 
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.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/win32utils-devel/attachments/20080324/bed374a4/attachment-0001.html 

From phasis at gmail.com  Mon Mar 24 22:31:53 2008
From: phasis at gmail.com (Heesob Park)
Date: Tue, 25 Mar 2008 11:31:53 +0900
Subject: [Win32utils-devel] GetTempPath vs GetEnvironmentVariable vs
	getenv
In-Reply-To: <7524A45A1A5B264FA4809E2156496CFB023D30F2@ITOMAE2KM01.AD.QINTRA.COM>
References: <7524A45A1A5B264FA4809E2156496CFB023D30F2@ITOMAE2KM01.AD.QINTRA.COM>
Message-ID: <aea95bc50803241931j566e7ddeq72e8e98aa8e11dce@mail.gmail.com>

Hi,

2008/3/25, Berger, Daniel <Daniel.Berger at qwest.com>:

>  Hi all,
>
> I discovered some interesting quirks with regards to getting environment
> variable information on Windows. It's because 'TMP' is considered a system
> environment variable, which you can't change with SetEnvironmentVariable()
> according to the docs.
>

I don't think so. 'TMP' can be changed with SetEnvironmentVariable().
But it affected only the current running process and child processes.
Environment variable can be categorized into three levels: process, user and
system level.
You can change user and System environment variable persistently via
registry setting.



>  The odd thing about TMP is that my shell value doesn't match the registry
> setting. The 'set' output shows:
>
> TMP=C:\DOCUME~1\djberge\LOCALS~1\Temp
>
> But the registry entry, under
> HKEY_LOCAL_MACHINE\System\CurrentControlSet\Control\Session
> Manager\Environment, is set to:
>
> C:\Temp
>


User environment variables are stored in the Registry in the key:

HKEY_CURRENT_USER\Environment

System variables are found in the key:
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session
Manager\Environment

Refer to http://vlaurie.com/computers2/Articles/environment.htm

Regards,

Park Heesob
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/win32utils-devel/attachments/20080325/64b0e14d/attachment.html 

From djberg96 at gmail.com  Thu Mar 27 07:47:17 2008
From: djberg96 at gmail.com (Daniel Berger)
Date: Thu, 27 Mar 2008 05:47:17 -0600
Subject: [Win32utils-devel] Fwd: CreateProcess function prototype
In-Reply-To: <002701c88ff1$c8342510$660aa8c0@wax>
References: <002701c88ff1$c8342510$660aa8c0@wax>
Message-ID: <6037b70c0803270447y69fbd3e1g237676ca6f83f010@mail.gmail.com>

Hm, I think he's write about params 3 and 4. Do we have to resort to
the address for the ENV data structure? I can't remember now.

Regards,

Dan


---------- Forwarded message ----------
From: Steve Shreeve <steve.shreeve at gmail.com>
Date: Thu, Mar 27, 2008 at 4:03 AM
Subject: CreateProcess function prototype
To: Daniel Berger <djberg96 at gmail.com>




Dan,

Also, the function prototype for CreateProcess on line 65 of
windows\process.rb is incorrect. It should be:

      API.new('CreateProcess', 'PPPPLLLPPP', 'B')

And, it might even be better to be:

      API.new('CreateProcess', 'PPPPLLPPPP', 'B')

Thanks,

Steve

From phasis at gmail.com  Thu Mar 27 08:31:43 2008
From: phasis at gmail.com (Park Heesob)
Date: Thu, 27 Mar 2008 21:31:43 +0900
Subject: [Win32utils-devel] Fwd: CreateProcess function prototype
References: <002701c88ff1$c8342510$660aa8c0@wax>
	<6037b70c0803270447y69fbd3e1g237676ca6f83f010@mail.gmail.com>
Message-ID: <001201c89006$852d2130$9a7ba8c0@mycomputer>

Hi,
----- Original Message ----- 
From: "Daniel Berger" <djberg96 at gmail.com>
To: "win32utils-devel" <win32utils-devel at rubyforge.org>
Sent: Thursday, March 27, 2008 8:47 PM
Subject: [Win32utils-devel] Fwd: CreateProcess function prototype


> Hm, I think he's write about params 3 and 4. Do we have to resort to
> the address for the ENV data structure? I can't remember now.
>
I remember ENV is not just a pointer but a pointer of pointer.
So it should be an address.

Regards,
Park Heesob

From Daniel.Berger at qwest.com  Thu Mar 27 10:16:08 2008
From: Daniel.Berger at qwest.com (Berger, Daniel)
Date: Thu, 27 Mar 2008 09:16:08 -0500
Subject: [Win32utils-devel] Fwd: CreateProcess function prototype
In-Reply-To: <001201c89006$852d2130$9a7ba8c0@mycomputer>
References: <002701c88ff1$c8342510$660aa8c0@wax><6037b70c0803270447y69fbd3e1g237676ca6f83f010@mail.gmail.com>
	<001201c89006$852d2130$9a7ba8c0@mycomputer>
Message-ID: <7524A45A1A5B264FA4809E2156496CFB023D3112@ITOMAE2KM01.AD.QINTRA.COM>

 

> -----Original Message-----
> From: win32utils-devel-bounces at rubyforge.org 
> [mailto:win32utils-devel-bounces at rubyforge.org] On Behalf Of 
> Park Heesob
> Sent: Thursday, March 27, 2008 6:32 AM
> To: Development and ideas for win32utils projects
> Subject: Re: [Win32utils-devel] Fwd: CreateProcess function prototype
> 
> Hi,
> ----- Original Message -----
> From: "Daniel Berger" <djberg96 at gmail.com>
> To: "win32utils-devel" <win32utils-devel at rubyforge.org>
> Sent: Thursday, March 27, 2008 8:47 PM
> Subject: [Win32utils-devel] Fwd: CreateProcess function prototype
> 
> 
> > Hm, I think he's write about params 3 and 4. Do we have to resort to
> > the address for the ENV data structure? I can't remember now.
> >
> I remember ENV is not just a pointer but a pointer of pointer.
> So it should be an address.

Yes, that's right.

It looks like I never caught the mistake in params 3 and 4 because I
mistyped the hash keys in the Process.create method with regards to
process_inherit and thread_inherit, so they were never actually being
used. There were a couple of mistakes there, actually.

Anyway, I'll fix win32-process and windows-pr tonight.

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 djberg96 at gmail.com  Thu Mar 27 19:03:07 2008
From: djberg96 at gmail.com (Daniel Berger)
Date: Thu, 27 Mar 2008 17:03:07 -0600
Subject: [Win32utils-devel] IO.socketpair addition -- very helpful!
In-Reply-To: <006a01c89022$810f23c0$660aa8c0@wax>
References: <006a01c89022$810f23c0$660aa8c0@wax>
Message-ID: <6037b70c0803271603h12a69215h6aaa5be8e0d5e784@mail.gmail.com>

Steve,

Thanks for this.

May I recommend joining the win32utils-devel mailing list? You would
be most welcome with ideas like this!

Regards,

Dan

On Thu, Mar 27, 2008 at 9:52 AM, Steve Shreeve <steve.shreeve at gmail.com> wrote:
>
>
> Dan,
>
> I had nearly pulled my hair out over the past week trying to figure out why
> I couldn't spawn a process on Windows and communicate with it. A few things
> I found out later:
>
> 1) The 'fork' implementation is totally gorked and should be renamed to
> something other that 'fork'... it's more like 'clone_me_and_do_over()'.
> 2) On Windows, pipes are essentially worthless because select() always says
> there's data, which will take you into a black hole if you try to read it.
>
> I *did* figure out, however, that sockets actually behave quite nicely on
> Windows. Also, I found your latest win32/process has some nice additions
> (modulo the few tweakers we've been discussing recently).
>
> In order to create a fairly clean way to interact with subprocesses, I've
> now got the code below.
>
> Let me know if this is helpful.
>
> Thanks,
>
> Steve
>
> ==
>
> require 'socket'
>
> STDIN.sync = STDOUT.sync = STDERR.sync = true
>
> class IO
>   def self.socketpair(sync=true)
>     if RUBY_PLATFORM.include?('win32')
>       tcp = TCPServer.new('127.0.0.1', 0)
>       one = TCPSocket.new('127.0.0.1', tcp.addr[1])
>       two = tcp.accept and tcp.close
>     else
>       one, two = UNIXSocket.socketpair
>     end
>     one.sync = two.sync = true if sync
>     [one, two]
>   end
>
>   def self.spawn(cmd=nil)
>     io, child_io = IO.socketpair
>
>     require 'win32/process'
>     child = Process.create(
>       'app_name' => "cmd /k #{cmd}", # the "/k" keeps the process around
> when it's done
>       'process_inherit' => true, # not yet sure if this one is actually
> needed
>       'thread_inherit' => true, # not yet sure if this one is actually
> needed
>       'startup_info' => {
>         'stdin'  => child_io,
>         'stdout' => child_io,
>         'stderr' => File.open('nul', 'wb') # ignore STDERR, for now (what
> about sync? close?)
>       }
>     )
>     at_exit do
>       Process.TerminateProcess(child.process_handle, child.process_id)
>       Process.CloseHandle(child.process_handle)
>     end
>     child_io.close
>
>     io
>   end
> end
>
> io = IO.spawn('date')
>
> # communicate using "io" and control threads...
>