http://samples.gotdotnet.com/quickstart/winforms/Samples/IeSourcing/VB/IESourcing.htm
> Date: Wed, 19 Apr 2006 12:23:57 -0400> From: [EMAIL PROTECTED]> Subject: Re:
> [ADVANCED-DOTNET] Click Once and Internet Zone> To:
> ADVANCED-DOTNET@DISCUSS.DEVELOP.COM> > I did not know you can host a WinFor
With Framework 2.0 I can host a Winforms control in a web page with code
running in the Internet Zone permission set. Everything works fine. However,
if I create a Click Once application which also has it's runtime permissions
limited to the internet zone it doesn't just deploy. The "Publishe
d exit, having set
up to handle the EndRead.
Sounds interesting
At 11:54 AM 3/16/2006, John Davis wrote
>Does this support chunking? If so, how does it work? Samples anyone?
J. Merrill / Analytical Software Corp
===
This list is hosted by Deve
Does this support chunking? If so, how does it work? Samples anyone?
===
This list is hosted by DevelopMentor® http://www.develop.com
View archives and manage your subscription(s) at http://discuss.develop.com
If you pass a callback method and call EndXXX in it then it is a completion
port thread.
From: Tony Hal <[EMAIL PROTECTED]>
Reply-To: "Discussion of advanced .NET topics."
To: ADVANCED-DOTNET@DISCUSS.DEVELOP.COM
Subject: [ADVANCED-DOTNET] Who calls the async socket callback
Date: Thu, 16 Feb 2
The CopyPixelOperation values don't work if the destination is a graphics
object backed by an in memory bitmap. However, it does work if the Graphics
object is from a Control (ie backed with video memory). Is there a way in
.Net to allocate a Bitmap which is allocated from video memory?
===
Anybody know what this actually does? The documentation is pretty slim, and
I can't see any samples on the web that use it. Perhaps it just tosses the
http header?
===
This list is hosted by DevelopMentor® http://www.develop.com
View archives and manage your su
g OffsetLow and OffsetHigh to see if it
can be made thread-safe.
On Mon, 19 Dec 2005 13:54:47 -0600, John Davis <[EMAIL PROTECTED]> wrote:
>I'd have to disagree. If the Seek logic was removed, BeginWriteCore
would
>be fine for multithreaded access.
>
>
>>From: Pete
O operation; which the existing FileStream
implementation is not designed to do.
http://www.peterRitchie.com/
On Mon, 19 Dec 2005 12:11:20 -0600, John Davis <[EMAIL PROTECTED]> wrote:
>Yes ... is it doing that?
>
>
>>From: Peter Ritchie <[EMAIL PROTECTED]>
>>Reply-To
. But you can't use the same overlapped structure for concurrent
writes.
On Mon, 19 Dec 2005 11:59:07 -0600, John Davis <[EMAIL PROTECTED]> wrote:
>See BeginWriteCore ... it already allocates the OVERLAPPED.
>
>>From: Peter Ritchie <[EMAIL PROTECTED]>
>>Repl
oesn't make sense.
http://www.peterRitchie.com/
On Mon, 19 Dec 2005 09:09:52 -0600, John Davis <[EMAIL PROTECTED]> wrote:
>Fair enough, we need an AsyncFileStream class which inherits from
FileStream
>and has BeginWrite/Read methods on it which contain file offset
arguments.
happen to hit the case where
you want to do something that you know would be pretty easy for the
underlying implementation you happen to be using. But that doesn't
translate into an argument that Stream is designed wrong.
--
Ian Griffiths
http://www.interact-sw.co.uk/iangblog/
> -Or
ded access with FileStream you're either going to have to synchronize
>access to a single FileStream object, or use one FileStream object per
>thread. This is exactly how FileStream is documented; there's no bug.
>
>If you want a thread-safe object to perform multi-thread
DISCUSS.DEVELOP.COM
Subject: Re: [ADVANCED-DOTNET] Possible defect in BeginWrite on FileStream
Date: Thu, 15 Dec 2005 22:42:36 -
John Davis wrote:
> Ian I'm afraid you're really mistaken on this one.
Which bit do you disagree with?
Do you disagree that Stream is the wrong abs
ead/write operation is done "on the background"; but, it
doesn't imply multi-threaded access to a file.
http://www.peterRitchie.com/
On Thu, 15 Dec 2005 10:01:39 -0500, John Davis <[EMAIL PROTECTED]> wrote:
>Ian I'm afraid you're really mistaken on this one
dly it's the only one we're offered
>AFAIK.)
>
>But is this a defect? It's a bug to use a stream from multiple threads
>without synchronization. It would still be a bug even with the atomic
>method you propose.
>
>
>--
>Ian Griffiths
>http://w
Just wanted to put this link here for posterity. There is what I would
call a bug in the async methods for v1.1 of the framework in that they
really aren't compatible with the current GC. You need to preallocate to
avoid a managed heap which will grow without end.
http://blogs.msdn.com/yunjin/ar
a snapshot of
the seek location when the operation starts?
--
Ian Griffiths
http://www.interact-sw.co.uk/iangblog/
> -Original Message-
> From: John Davis
>
> The way I understand it, the current BeginWrite on the FileStream
object
> will write out it's buffer at the
The way I understand it, the current BeginWrite on the FileStream object will
write out it's buffer at the location of the last Seek. The problem comes when
there are multiple threads using the same FileStream instance. Each one is
looking to write to a different offset in the file. This mean
Uncle! :)
Ian Griffiths <[EMAIL PROTECTED]> wrote:> >2) To call blocking APIs
>
> That's what I was looking for. If we eliminate blocking IO API's
> then there isn't a need to spawn off threads.
Sure, but often once you've eliminated the blocking IO APIs you're left
with nothing, because there t
x27;t shipped yet. I think
calling the current product 'legacy' code is a bit of a stretch.
(Whether you'd use this technique with databases is a separate question
of course - arranging to have hundreds of concurrent requests to the DB
might not help your performance much...)
--
Ian G
this case, ASP.NET's built-in
threading works pretty well.
The case where the technique that Fritz illustrates works best is in
heavily IO-bound work where the IO you are doing happens to force you to
use a synchronous API.
But I do concur with your assessment of your comment's value
I'm sure Fritz is a great guy and all, but that "how to create an extra thread
pool" article is one of the worst pieces I've seen yet. The whole point of
async IO is to have the async work pump through IO completion ports, this
eliminates the need for thread pools. The only value I see in Frit
e, generates optimized code for other processors as
well."
From
http://msdn.microsoft.com/msdnmag/issues/04/03/ScientificC/
also
http://www.devx.com/Intel/Article/6960
Ben
> -Original Message-
> From: Unmoderated discussion of advanced .NET topics.
[mailto:ADVANCED-
> [EM
It's been a couple years now and we still can't access parallel instructions
from within managed code. If there are no plans for vectorizing through the
jitter, at least give us IL instructions similar to the C++ intrinsics for SSE
and SSE2.
__
No, there isn't any remoting connection. But there is a socket connection between the
two.
-
Do you Yahoo!?
Yahoo! Mail Address AutoComplete - You start. We finish.
===
This list is hosted by DevelopMentor® http://www.develop.c
I just had our main eventing service crash. It is a .Net Service written in C#, all
of it is managed code. Upon attaching a debugger to the service process and looking
at the threads I noticed one of them had System.Winforms calls on it. There are no
references whatsoever in the service to Wi
If I call BeginSend, and then call it a couple more times before the first one has
completed can I rest assured that the corresponding byte buffers will will go out on
the wire in the same order? ie Are all calls to BeginSend on the same socket queued?
-
Do you
way to do this
(other than Managed C++ and P/Invoke); I am curious to know also.
Srihari
On Wed, 8 Sep 2004, John Davis wrote:
> Got asked for more details.
>
> If you look at the latest docs on FileStream.BeginRead it says ...
>
> By default, FileStream opens the operating system h
calls to the file handle. I would prefer to have my async operations fire
through an IO completion port, like the socket implementation. Usually more efficient
if done this way.
John Davis <[EMAIL PROTECTED]> wrote:
Is it possible to get the IntPtr file handle for this call without doing i
Is it possible to get the IntPtr file handle for this call without doing interop?
[C#]public FileStream( IntPtr handle, FileAccess access, bool ownsHandle, int
bufferSize, bool isAsync);
-
Do you Yahoo!?
New and Improved Yahoo! Mail - 100MB free stora
ternalWaitForCompletion on LazyAsyncResult) on the internal Event
for you. What makes you think EndSend will not work in this form?
Garry
On Wed, 4 Aug 2004 19:29:46 -0700, John Davis wrote:
>Do a WaitOne on the WaitHandle in the IAsyncResult, returned from
BeginSend, before calling EndSe
Do a WaitOne on the WaitHandle in the IAsyncResult, returned from BeginSend, before
calling EndSend. Or pass a callback to the BeginSend, and then call EndSend within
the callback. A lot of the current socket documentation is just plain wrong. I
reported it a couple weeks ago, and got a respo
I would really like to see a test app posted that reproduces this behavior. My best
guess is that the problem is due to lack of synchronization somewhere.
"J. Merrill" <[EMAIL PROTECTED]> wrote:I'm surprised that the framework code doesn't
allocate a large block of (possibly unmanaged) memory t
Yes, synchonization out the wazoo.
In terms of debugging deadlocks, etc. I always put my server code in a dll and then
run it inside a console app when I'm debugging. And have it run in a service when
things are in production. This makes your debugging, esecially things like deadlocks,
a lot
Our server only has one appdomain, and it works great. Another thing you really need
to be careful about is the whole callback thing. Assume that your callbacks will be
coming back on multiple IO threads. This will definitely be the case when running on
a multi-processor box (or hyperthreaded
Just realized I misunderstood the original post. Yep, sounds like a virus or
something. I've never experienced your problem.
John Davis <[EMAIL PROTECTED]> wrote:Recently finished writing an eventing server
using async sockets. All I can say is that the performance is nothin
it'll probably take around 3x longer.
Have a good one,
John Davis
-
Do you Yahoo!?
Vote for the stars of Yahoo!'s next ad campaign!
===
This list is hosted by DevelopMentor® http://www.develop.com
Some .NET courses
it'll probably take around 3x longer.
Have a good one,
John Davis
Garry Barclay <[EMAIL PROTECTED]> wrote:
It appears that we are NOT receiving the (completion port) callback from
an asynchronous socket (BeginSend/EndSend) call.
The sockets are clients and the failure seems to be r
the TCP protocol will pipe the
1024 bytes, and wait patiently until I am ready to receive the data and
deliver it to me. So, it shouldn't matter if I receive it in one chunk or
4 smaller chunks.
At least, this is my understanding of the whole sockets thing :-)
Srihari
On Mon, 28 Jun 2004, John Davis
anyone wants me too. $500/hr :)
Regards,
John Davis
http://www.dallas-csharp-sig.com
-
Do you Yahoo!?
New and Improved Yahoo! Mail - Send 10MB messages!
===
This list is hosted by DevelopMentor® http://www.develop.com
Some .NET courses
If an async web method call times out before the corresponding EndXXX gets
called, does the corresponding HttpContext object automatically get freed?
===
This list is hosted by DevelopMentor® http://www.develop.com
Some .NET courses you may be interested in:
NEW!
Yes, Control.Invoke uses the message pump. The message pump is the queue
you speak of. I'm guessing the mechanism uses a PostMessage with a WM_USER
+ x.
- Original Message -
From: "John Elliot" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Monday, March 01, 2004 5:34 AM
Subject: Re:
ular receive request because it'll be handled by the pool of
threads dedicated to async IO requests, if the OS supports IO completion
ports.
--
Ian Griffiths
DevelopMentor
> -Original Message-
> From: John Davis [mailto:[EMAIL PROTECTED]
>
> Someone pointed this out to me the o
Someone pointed this out to me the other day in the docs ...
When your application calls BeginReceive, the system will use a separate thread to
execute the specified callback method, and will block on EndReceive until the Socket
reads data or throws an exception.
I would hope that the callback d
It works for me. I've got some sample code on my website.
www.pcprogramming.com
- Original Message -
From: "Borisov, William" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Thursday, April 03, 2003 4:51 PM
Subject: [ADVANCED-DOTNET] Remoting Events Framework version 1.1
> I'm trying
The most efficient way to program sockets is via io completion ports. If
you use the (Begin/End)*** methods on the Socket class, and the host OS
supports io completion ports, these calls will get mapped to io completion
ports via the thread pool.
- Original Message -
From: "Nigel" <[EMAIL
Using the async begin/end method is 10x easier than doing this directly from
the Win32 API, and just as efficient. One can pass arbitrary(generic) state
around using the "object state" param.
- Original Message -
From: "Fiach Reid" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Friday,
Bingo, look at the IHttpHandler interface for the server side. You'll need
to create your own signaling stream if you want to reach clients behind a
firewall. You'll consume this stream on the client side via the
HttpWebRequest class. Also be aware that some firewalls close the socket
after so m
gt; -- keeping everything from being marked for collection.
>
> -Mike
> http://staff.develop.com/woodring
> http://www.develop.com/devresources
>
> [1]
> http://discuss.develop.com/archives/wa.exe?A2=ind0012C&L=DOTNET&P=R22579
> &I=-3
>
> > -Original Me
s,
> Patrik
>
> > -Original Message-
> > From: John Davis [mailto:[EMAIL PROTECTED]]
> > Sent: 2003-01-13 20:46
> > To: [EMAIL PROTECTED]
> > Subject: Re: [ADVANCED-DOTNET] IP Server Questions...
> >
> >
> > Along this line ... is the Socket
Along this line ... is the Socket class ever going to get layered on top of
IO completion ports?
- Original Message -
From: "Scott Burrington" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Monday, January 13, 2003 12:32 PM
Subject: [ADVANCED-DOTNET] IP Server Questions...
> Hi all,
>
and any Java app. Am I right, or have I
overseen
> an important issue?
>
>
> -Ursprüngliche Nachricht-
> Von: John Davis [mailto:jdavis@;PCPROGRAMMING.COM]
> Gesendet: Mittwoch, 23. Oktober 2002 14:09
> An: [EMAIL PROTECTED]
> Betreff: Re: [ADVANCED-DOTNET] host and network byte order?
Java is big endian (the MSB (most significant byte) comes first). .Net
appears to be little endian (LSB (least significant byte) comes first). But
I don't think the standard really specifies anything. So, I guess it just
defaults to whatever the native ordering is for the CPU. For Intel CPU's
t
int32)
> IL_00e5: throw
> IL_00e6: ldarg.0
> IL_00e7: call instance class
[mscorlib]System.Collections.ArrayList
> System.Net.Sockets.Socket::get_AcceptQueue()
> IL_00ec: ldloc.0
> IL_00ed: callvirt instance int32
> [mscorlib]System.Collections.ArrayList::Add(object)
>
On NT kernel OS's BeginAccept calls AcceptEx, the completion of this comes
back through GetQueuedCompletionStatus (in the thread pool) which is then
used to call your AsyncCallback(OnConnected). Most our there who have not
used IO completion ports don't realize how sweet this is. It's now possib
Look at the BeginRead method on Socket, the event you're looking for is
actually a callback.
- Original Message -
From: "Jeff Kryzer" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Friday, September 06, 2002 3:52 PM
Subject: [ADVANCED-DOTNET] No Socket Receive event
> I've searched a
>that you're hitting the default HTTP 1.1 2-concurrent-connections to 1
>server per process maximum when .Net is trying to fetch the application
>config file? There are ways to open more than 2 concurrent connections, but
>I don't know them from the top of my head. Search MSDN
Couldn't get an answer in the last group so I'm askin the big guns now.
The socket ctor acts strangely when executed within ieexec.exe The first
call can take over a minute. If I attach the debugger and break, I get the
call stack below when using a release build.
-
Had the same problem. I gave up on "Everything" and am just using
"FullTrust". And believe it or not I am still hitting problems within
ieexec.exe when the socket library initializes. Documentation is sorely
lacking on all this security stuff. Windows forms controls act one way,
href launched
Why is it that I can programmatically add code groups to RootCodeGroup of
the "Machine" policy level. But if I try to add a code group to any of the
descendants ie Trusted_Zone it doesn't show up in the ".Net Framework
Configuration tool"
What's really got me puzzled is that I can manually add c
61 matches
Mail list logo
