Search Results

Search found 7239 results on 290 pages for 'john michael zorko'.

Page 124/290 | < Previous Page | 120 121 122 123 124 125 126 127 128 129 130 131  | Next Page >

  • Java / Groovy Socket - Detecting the socket being closed in a non-blocking way

    - by John Arrowwood
    I'm trying to create a small HTTP proxy that can re-write the request/headers as needed to suit my requirements. If one already exists, please, point me to it. Otherwise... I've written something that ALMOST works. It can do the proxy function, but not the re-write (yet). Problem is, I can't detect when the remote socket has been closed down without doing a blocking read. It is CRITICAL for the functionality of this thing that it be able to detect the socket being closed without blocking. I have SCOURED the Java API documentation, and I can't find ANY indication that it is even possible. Here's what I have: while ( this.inbound.isConnected() && this.outbound.isConnected() ) { try { while ( ( available = readFromClient.available() ) != 0 ) { if ( available > 1024 ) available = 1024 bytesRead = readFromClient.read( buffer, 0, available ) writeToServer.write( buffer, 0, bytesRead ) } while ( ( available = readFromServer.available() ) != 0 ) { if ( available > 1024 ) available = 1024 bytesRead = readFromServer.read( buffer, 0, available ) writeToClient.write( buffer, 0, bytesRead ) } } catch (e) { print e } println "Connected: " + this.inbound.isConnected() println "Bound: " + this.inbound.isBound() println "InputShutdown: " + this.inbound.isInputShutdown() println "OutputShutdown: " + this.inbound.isOutputShutdown() print "\n"; Thread.sleep( 10 ) } The tests for the socket being closed never indicate that the socket was closed. And, as I mentioned, I can't find ANY examples of how to detect the 'END OF FILE' condition on the stream without doing a blocking read. There HAS to be a way. Does anyone here know what it is?

    Read the article

  • How can I detect when the .NET framework is "turned off"?

    - by John Myczek
    My application requires the .NET Framework version 3.5. I recently ran into a customer that had the .NET Framework installed but turned off. In this case, my installer (InstallShield 2009) does not prompt the user to install the Framework (because it is already installed) and when my application runs it crashes immediately. I tried another .NET application and it also crashes immediately. Is there any way to detect this situation and handle it more gracefully? Just detecting this during install is not ideal since the .NET Framework can be turned off at any time. Ideally, the application would be able to check and display a friendly message to the user telling them they need to turn on the .NET Framework.

    Read the article

  • LEFT OUTER JOIN SUM doubles problem

    - by Michael
    Hi I've got two tables: Table: Shopping shop_id shop_name shop_time 1 Brian 40 2 Brian 31 3 Tom 20 4 Brian 30 Table:bananas banana_id banana_amount banana_person 1 1 Brian 2 1 Brian I now want it to print: Name: Tom | Time: 20 | Bananas: 0 Name: Brian | Time: 101 | Bananas: 2 I used this code: $result = dbquery("SELECT tz.*, tt.*, SUM(shop_time) as shoptime, count(banana_amount) as bananas FROM shopping tt LEFT OUTER JOIN bananas tz ON tt.shop_name=tz.banana_person GROUP by banana_person LIMIT 40 "); while ($data5 = dbarray($result)) { echo 'Name: '.$data5["shop_name"].' | Time: '.$data5["shoptime"].' | Bananas: '.$data5["bananas"].'<br>'; } The problem is that I get this instead: Name: Tom | Time: 20 | Bananas: 0 Name: Brian | Time: 202 | Bananas: 6 I just don't know how to get around this.

    Read the article

  • Beanshell in Ant yielding, "Unable to create javax script engine for beanshell"

    - by John B.
    Greeting, I'm trying to put some Beanshell script in my Ant build.xml file. I've followed the Ant manual as well as I can but I keep getting "Unable to create javax script engine for beanshell" when I run Ant. Here is the test target I wrote mostly from examples in the Ant manual: <target name="test-target"> <script language="beanshell" setbeans="true"> <classpath> <fileset dir="c:\TEMP" includes="*.jar" /> </classpath> System.out.println("Hello world"); </script> </target> My beanshell "bsh-2.0b4.jar" file is on the script task's classpath the way the manual recommended. Hope I have the right file. I'm working in c:\TEMP right now. I've been googling and trying for a while now. Any ideas would be greatly appreciated. Thanks.

    Read the article

  • Scheme Depth-first search of a graph function

    - by John Retallack
    This is a homework question,I'm trying to do a Depth-first search function in Scheme,Here's the code I've written so far: (define explore (?(node visited) (let* ([neighbors (force (cdr node))] [next (nextNode visited neighbors)] [is-visited (member? node visited)]) (cond ;if I have no unvisited neighbours print current node and go up one level [(equal? next #f) (begin (display (car node)) (display " "))] ;if current node is not visited and I have unvisited neighbors ;print current node,mark as visited and visit it's neighbors [(and (equal? is-visited #f) (not (equal? next #f))) (begin (display (car node)) (display " ") (explore next (cons node visited)))]) ;go and visit the next neighbor (if (not (equal? (nextNode (cons next visited) neighbors) #f )) (explore (nextNode (cons next visited) neighbors) (cons node visited)))))) 'node' is the current node 'visited' is a list in witch I keep track of the nodes I visited 'nextNode' is a function that returns the first unvisited neighbor if any or #f otherwise 'member?' test's if a node is in the visited list The Graph representation is using adjacent made using references to nodes with letrec so that's why I use force in 'neighbors': Eg: (letrec ([node1 (list "NY" (delay (list node2 node3)))],where node2 and node3 are defined as node1 The problem witch I'm dealing with is that my visited lists looses track of some of the nodes I visited when I come out of recursion,How can I fix this ?

    Read the article

  • Parallel port with C#

    - by Michael S.
    Hello, I am trying to send data to LPT1 port with a C# program, unfortunately with no success.. I am using windows 7 x64. I tried both x86 and x64 (inpoutx64.dll) dll's.. With the x64 dll when I send: Output(888, 255); It just continues the program as everything went ok, but i can't see anything on my multimeter (only the static 0.02V).. I also tried the following with C++: int main () { int val = 0; printf("Enter a value\n"); scanf("%d", &val); _outp(0x378, val); getchar(); _outp(0x378, 0); return 0; } But it throws an exception: Unhandled exception at 0x01281428 in ppac.exe: 0xC0000096: Privileged instruction. I remember once I made something like this work on xp, I hope it's possible on win7 too.. Please help me with this. Thanks.

    Read the article

  • Telerik RadAlert back button cache problem

    - by Michael VS
    Hi, I'm sure you have had this one before so if you could point me to something similar.... I have a server side creation of a RadAlert window using the usual Sys.Application.remove_load and add_load procedure however the alert keeps popping up as it seems to be caching when the user hits the back button after it has been activated. I have tried to put a onclick event on a button to clear the function using remove_load before it moves to the next page however it still doesn't seem to clear it. Its used in validation so if a user inputs failed validation it pops up. If they then go and enter correct validation it then moves onto the next page. If they then use back button this is where it pops up again. Any ideas? Server side: private void Page_Load(object sender, System.EventArgs e) { if (!IsPostBack) { btnSearch.Attributes.Add("onclick", "Sys.Application.remove_load(f);"); } } private void btnSearch_Click(object sender, System.EventArgs e) { string radalertscript = "(function(){var f = function(){radalert('Welcome to RadWindow Prometheus!', 330, 210); Sys.Application.remove_load(f);};Sys.Application.add_load(f);})()"; RadAjaxManager1.ResponseScripts.Add(radalertscript); } Ive also tried using RadAjaxManager1.ResponseScripts.Clear(); before it moves on to the next page on the postback event

    Read the article

  • Tomcat cookies not working via my ProxyPass VirtualHost

    - by John
    Hi there. I'm having some issues with getting cookies to work when using a ProxyPass to redirect traffic on port 80 to a web-application hosted via Tomcat. My motivation for enabling cookies is to get rid of the "jsessionid=" parameter that is appended to the URLs. I've enabled cookies in my context.xml in META-INF/ for my web application. When I access the webapplication via http://url:8080/webapp it works as expected, the jsessionid parameter is not visible in the URL, instead it's stored in a cookie. When accessing my website via an apache2 virtualhost the cookies doesn't seem to work because now "jsessionid" is being appended to the URLs. How can I solve this issue? Here's my VHost configuration: <VirtualHost *:80 ServerName somedomain.no ServerAlias www.somedomain.no <Proxy * Order deny,allow Allow from all </Proxy ProxyPreserveHost Off ProxyPass / http://localhost:8080/webapp/ ProxyPassReverse / http://localhost:8080/webapp/ ErrorLog /var/log/apache2/somedomain.no.error.log CustomLog /var/log/apache2/somedomain.no.access.log combined </VirtualHost

    Read the article

  • Resharper doesn't play well with XAML?

    - by John Weldon
    I've been noticing a pattern with ReSharper (both 4.5 and 5). Very often (almost always) when I have solution-wide analysis turned on, and WPF code in my solution, ReSharper will mark a number of the .xaml.cs files as being broken. When I navigate to the file, sometimes it magically updates and displays no errors, and other times I have to open other files that are not being correctly read and close them again to force resharper to correctly analyze them. I assume it has something to do with the temporary .cs code that is generated with XAML, but does anyone know why this is actually happening, and if there is a work around? Should I just file a bug report with JetBrains? Does anyone else experience this?

    Read the article

  • Native Windows Application Development Options

    - by Michael Stum
    Long winded title, short question: If one wants to develop for Windows but not have to rely on any external dependency (no runtime, thus ruling out .net), what supported, alive and fully functioning* alternatives are there? Visual Basic 6 is dead, Visual C++ is obvious and Delphi seems to be the prime choice for that, but I wonder if there are any other alternatives? *as in: Being able to use all the Windows Features like putting an icon in the Notification Area, making the Taskbar Icon flash etc.

    Read the article

  • Updating Xml attributes with new values in a SQL Server 2008 table

    - by SMD
    I have a table in SQL Server 2008 that it has some columns. One of these columns is in Xml format and I want to update some attributes. For example my Xml column's name is XmlText and it's value in 5 first rows is such as: <Identification Name="John" Family="Brown" Age="30" /> <Identification Name="Smith" Family="Johnson" Age="35" /> <Identification Name="Jessy" Family="Albert" Age="60" /> <Identification Name="Mike" Family="Brown" Age="23" /> <Identification Name="Sarah" Family="Johnson" Age="30" /> and I want to change all Age attributes that are 30 to 40 such as below: <Identification Name="John" Family="Brown" Age="40" /> <Identification Name="Smith" Family="Johnson" Age="35" /> <Identification Name="Jessy" Family="Albert" Age="60" /> <Identification Name="Mike" Family="Brown" Age="23" /> <Identification Name="Sarah" Family="Johnson" Age="40" />

    Read the article

  • Can't get NSDate to work correctly

    - by John
    Hello, having a strange issue, must be something I'm just not seeing. I set up a variable in the .h NSDate *checkIn; @property (nonatomic, retain) NSDate *checkIn; I'm setting a variable to todays date in the initWithNibName: checkIn = [NSDate date]; I also did synthesized it as well. Now later on in my program I use it to build a tablecell with the following line cell.textLabel.text = [dateFormatter stringFromDate:checkIn]; This line kills the simulator, BAD_EXEC. If I put in a checkIn = [NSDate date]; above it, it works fine. So I'm thinking the variable isn't being stored from when I set it in the initWithNibName: Not sure why though, as my strings I do the same way are all working fine from method to method. What am I missing?

    Read the article

  • SugarCRM - how to make users see only child users?

    - by John
    Hello, I want to change the behaviour of Sugar CRM community version. Here's what sugar currently does: 1) Admin logs into Sugar. 2) Admin clicks on Admin tab. 3) Admin creates a new user named George with admin access 4) Under user information section, Admin makes George report to Admin (in the database, it will show users.report_to_id is the admin's user_id) 5) Admin saves and logs out 6) George logs in with his password 7) George goes to admin tab. 8) George goes to list users page and sees all users, including Admin, the person he is supposed to report to. I want to change step 8 such that George is not allowed to see the user he reports to. Additionally, he is not allowed to see his grand parent user, great grand parent user, great great grand parent user etc... George should only be able to see users that he creates. How can I achieve this? Is this even possible?

    Read the article

  • perl one liner alternative to this bash "chain"?

    - by Michael Mao
    Hello everyone: I am trying to comprehend Perl following the way describe in the book "Minimal Perl". I've uploaded all source txt files onto my own server : results folder I got the output from using several bash commands in a "chain" like this: cat run*.txt | grep '^Bank[[:space:]]Balance'| cut -d ':' -f2 | grep -E '\$[0-9]+' I know this is far from the most concise and efficient, but at least it works... As our uni subject now moves onto the Perl part, I'd like to know if there is a way to get the same results in one line? I am trying something like the following code but stuck in the middle: Chenxi Mao@chenxi-a6b123bb /cygdrive/c/eMarket/output $ perl -wlne 'print; if $n=~/^Bank Balance/' syntax error at -e line 1, near "if $n" Execution of -e aborted due to compilation errors.

    Read the article

  • FPDF Page Break Question

    - by Michael
    I am using PHP and FPDF to generate a PDF with a list of items. My problem is if the item list goes on to a second or third page, I want to keep the Item Name, Quantity and Description together. Right now, it will go to a second page, but it may split up all of the details for a particular item. PLEASE HELP! <?php require_once('auth.php'); require_once('config.php'); require_once('connect.php'); $sqlitems="SELECT * FROM $tbl_items WHERE username = '" . $_SESSION['SESS_LOGIN'] . "'"; $resultitems=mysql_query($sqlitems); require_once('pdf/fpdf.php'); require_once('pdf/fpdi.php'); $pdf =& new FPDI(); $pdf->AddPage('P', 'Letter'); $pdf->setSourceFile('pdf/files/healthform/meds.pdf'); $tplIdx = $pdf->importPage(1); $pdf->useTemplate($tplIdx); $pdf->SetAutoPageBreak(on, 30); $pdf->SetTextColor(0,0,0); $pdf->Ln(10); while($rowsitems=mysql_fetch_array($resultitems)){ $pdf->SetFont('Arial','B',10); $pdf->Cell(50,4,'Item Name:',0,0,'L'); $pdf->SetFont(''); $pdf->Cell(100,4,$rowsitems['itemname'],0,0,'L'); $pdf->SetFont('Arial','B',10); $pdf->Cell(50,4,'Quantity:',0,0,'L'); $pdf->SetFont(''); $pdf->Cell(140,4,$rowsitems['itemqty'],0,1,'L'); $pdf->SetFont('Arial','B'); $pdf->Cell(50,4,'Description:',0,0,'L'); $pdf->SetFont(''); $pdf->Cell(140,4,$rowsitems['itemdesc'],0,1,'L'); } $pdf->Output('Items.pdf', 'I'); ?>

    Read the article

  • C# Active Directory - Check username / password

    - by Michael G
    I'm using the following code on Windows Vista Ultimate SP1 to query our active directory server to check the user name and password of a user on a domain. public Object IsAuthenticated() { String domainAndUsername = strDomain + @"\" + strUser; DirectoryEntry entry = new DirectoryEntry(_path, domainAndUsername, strPass); SearchResult result; try { //Bind to the native AdsObject to force authentication. DirectorySearcher search = new DirectorySearcher(entry) { Filter = ("(SAMAccountName=" + strUser + ")") }; search.PropertiesToLoad.Add("givenName"); // First Name search.PropertiesToLoad.Add("sn"); // Last Name search.PropertiesToLoad.Add("cn"); // Last Name result = search.FindOne(); if (null == result) { return null; } //Update the new path to the user in the directory. _path = result.Path; _filterAttribute = (String)result.Properties["cn"][0]; } catch (Exception ex) { return new Exception("Error authenticating user. " + ex.Message); } return user; } the target is using .NET 3.5, and compiled with VS 2008 standard I'm logged in under a domain account that is a domain admin where the application is running. The code works perfectly on windows XP; but i get the following exception when running it on Vista: System.DirectoryServices.DirectoryServicesCOMException (0x8007052E): Logon failure: unknown user name or bad password. at System.DirectoryServices.DirectoryEntry.Bind(Boolean throwIfFail) at System.DirectoryServices.DirectoryEntry.Bind() at System.DirectoryServices.DirectoryEntry.get_AdsObject() at System.DirectoryServices.DirectorySearcher.FindAll(Boolean findMoreThanOne) at System.DirectoryServices.DirectorySearcher.FindOne() at Chain_Of_Custody.Classes.Authentication.LdapAuthentication.IsAuthenticated() at System.DirectoryServices.DirectoryEntry.Bind(Boolean throwIfFail) at System.DirectoryServices.DirectoryEntry.Bind() at System.DirectoryServices.DirectoryEntry.get_AdsObject() at System.DirectoryServices.DirectorySearcher.FindAll(Boolean findMoreThanOne) at System.DirectoryServices.DirectorySearcher.FindOne() at Chain_Of_Custody.Classes.Authentication.LdapAuthentication.IsAuthenticated() I've tried changing the authentication types, I'm not sure what's going on. See also: http://stackoverflow.com/questions/290548/c-validate-a-username-and-password-against-active-directory

    Read the article

  • Download HTML and Images with WGet without first few lines

    - by St. John Johnson
    I'm attempting to use wget with the -p option to download specific documents and the images linked in the HTML. The problem is, the site that is hosting the HTML has some non-html information preceding the HTML. This is causing wget to not interpret the document as HTML and doesn't search for images. Is there a way to have wget strip the first X lines and/or force searching for images? Example URL: http://www.sec.gov/Archives/edgar/data/13239/000119312510070346/ds4.htm First Lines of Content: <DOCUMENT> <TYPE>S-4 <SEQUENCE>1 <FILENAME>ds4.htm <DESCRIPTION>FORM S-4 <TEXT> <HTML><HEAD> <TITLE>Form S-4</TITLE> Last Lines of Content: </BODY></HTML> </TEXT> </DOCUMENT>

    Read the article

  • MKMapView within UITableView

    - by john-london
    I have a MKMapView within a UITableviewCell and I want to be able to disable scrolling of the table view when the user starts to drag within the map view. My plan had been to override the MKMapView touchesBegan/Moved etc methods and if they were called to then disable scrolling in the parent view. Scrolling in the parent view would be enabled again only when the touches in the MKMapView ended or were cancelled. My problem is that the touchesBegan/Moved methods inside MKMapView are not called - the problem is also discussed here: stackoverflow discussion. My table view is within a tab control, so I don't think I can extend UIWindow the way they describe, as there are several views that could be within the window, depending on which tab is active. How can I access the MKMapView touch events please?

    Read the article

  • Why can't I source my vimrc after installing vundle?

    - by John N
    I'm trying to get my Vim to update on the fly after editing vimrc. So I followed the instructions at Vimcast which basically source vimrc every time you hit save. But that doesn't work for some reason (when I save my vimrc it doesn't give any errors), so I decided to run source $HOME/.vimrc manually and here's what I got: -bash: Configuration file for vim set nocompatible : command not found -bash: Plugin Management { filetype off : command not found -bash: .vimrc: line 7: syntax error near unexpected token `(' -bash: .vimrc: line 7: ` call vundle#rc()' And here's my Vim info (I'm running Mac OS X 10.7.3): VIM - Vi IMproved 7.3 (2010 Aug 15, compiled Jul 31 2011 19:27:29) Compiled by [email protected] Normal version without GUI. Features included (+) or not (-): -arabic +autocmd -balloon_eval -browse +builtin_terms +byte_offset +cindent -clientserver -clipboard +cmdline_compl +cmdline_hist +cmdline_info +comments -conceal +cryptv +cscope +cursorbind +cursorshape +dialog_con +diff +digraphs -dnd -ebcdic -emacs_tags +eval +ex_extra +extra_search -farsi +file_in_path +find_in_path +float +folding -footer +fork() -gettext -hangul_input +iconv +insert_expand +jumplist -keymap -langmap +libcall +linebreak +lispindent +listcmds +localmap -lua +menu +mksession +modify_fname +mouse -mouseshape -mouse_dec -mouse_gpm -mouse_jsbterm -mouse_netterm -mouse_sysmouse +mouse_xterm +multi_byte +multi_lang -mzscheme +netbeans_intg -osfiletype +path_extra -perl +persistent_undo +postscript +printer -profile -python -python3 +quickfix +reltime -rightleft -ruby +scrollbind +signs +smartindent -sniff +startuptime +statusline -sun_workshop +syntax +tag_binary +tag_old_static -tag_any_white -tcl +terminfo +termresponse +textobjects +title -toolbar +user_commands +vertsplit +virtualedit +visual +visualextra +viminfo +vreplace +wildignore +wildmenu +windows +writebackup -X11 -xfontset -xim -xsmp -xterm_clipboard -xterm_save system vimrc file: "$VIM/vimrc" user vimrc file: "$HOME/.vimrc" user exrc file: "$HOME/.exrc" fall-back for $VIM: "/usr/share/vim" Compilation: gcc -c -I. -D_FORTIFY_SOURCE=0 -Iproto -DHAVE_CONFIG_H -arch i386 -arch x86_64 -g -Os -pipe Linking: gcc -arch i386 -arch x86_64 -o vim -lncurses Could anybody tell me what's wrong here? Thanks so much!

    Read the article

  • Binding an ASP.NET GridView Control to a string array

    - by Michael Kniskern
    I am trying to bind an ASP.NET GridView control to an string array and I get the following item: A field or property with the name 'Item' was not found on the selected data source. What is correct value I should use for DataField property of the asp:BoundField column in my GridView control. Here is my source code: ASPX page <asp:GridView ID="MyGridView" runat="server" AutoGenerateColumns="false"> <Columns> <asp:BoundField DataField="Item" /> <asp:CommandField ButtonType="Link" ShowSelectButton="true" SelectText="Click Me!" /> </Columns> </asp:GridView> Code Behind: string[] MyArray = new string[1]; MyArray[0] = "My Value"; MyGridView.DataSource = MyArray; MyGridView.DataBind(); UPDATE I need to have the AutoGenerateColumns attribute set to false because I need to generate additional asp:CommandField columns. I have updated my code sample to reflect this scenarion

    Read the article

  • [MFC] I can't re-parent a window

    - by John
    Following on from this question, now I have a clearer picture what's going on... I have a MFC application with no main window, which exposes an API to create dialogs. When I call some of these methods repeatedly, the dialogs created are parented to each other instead of all being parented to the desktop... I have no idea why. But anyway even after creation, I am unable to change the parent back to NULL or CWnd::GetDesktopWindow()... if I call SetParent followed by GetParent, nothing has changed. So apart from the really weird question of why Windows is magically parenting each dialog to the last one created, is there anything I'm missing to be able to set these windows as children of the desktop? UPDATED: I have found the reason for all this, but not the solution. From my dialog constructor, we end up in: BOOL CDialog::CreateIndirect(LPCDLGTEMPLATE lpDialogTemplate, CWnd* pParentWnd, void* lpDialogInit, HINSTANCE hInst) { ASSERT(lpDialogTemplate != NULL); if (pParentWnd == NULL) pParentWnd = AfxGetMainWnd(); m_lpDialogInit = lpDialogInit; return CreateDlgIndirect(lpDialogTemplate, pParentWnd, hInst); } Note: if (pParentWnd == NULL)pParentWnd = AfxGetMainWnd(); The call-stack from my dialog constructor looks like this: mfc80d.dll!CDialog::CreateIndirect(const DLGTEMPLATE * lpDialogTemplate=0x005931a8, CWnd * pParentWnd=0x00000000, void * lpDialogInit=0x00000000, HINSTANCE__ * hInst=0x00400000) mfc80d.dll!CDialog::CreateIndirect(void * hDialogTemplate=0x005931a8, CWnd * pParentWnd=0x00000000, HINSTANCE__ * hInst=0x00400000) mfc80d.dll!CDialog::Create(const char * lpszTemplateName=0x0000009d, CWnd * pParentWnd=0x00000000) mfc80d.dll!CDialog::Create(unsigned int nIDTemplate=157, CWnd * pParentWnd=0x00000000) MyApp.exe!CMyDlg::CMyDlg(CWnd * pParent=0x00000000)

    Read the article

  • Odd tcp deadlock under windows

    - by John Robertson
    We are moving large amounts of data on a LAN and it has to happen very rapidly and reliably. Currently we use windows TCP as implemented in C++. Using large (synchronous) sends moves the data much faster than a bunch of smaller (synchronous) sends but will frequently deadlock for large gaps of time (.15 seconds) causing the overall transfer rate to plummet. This deadlock happens in very particular circumstances which makes me believe it should be preventable altogether. More importantly if we don't really know the cause we don't really know it won't happen some time with smaller sends anyway. Can anyone explain this deadlock? Deadlock description (OK, zombie-locked, it isn't dead, but for .15 or so seconds it stops, then starts again) The receiving side sends an ACK. The sending side sends a packet containing the end of a message (push flag is set) The call to socket.recv takes about .15 seconds(!) to return About the time the call returns an ACK is sent by the receiving side The the next packet from the sender is finally sent (why is it waiting? the tcp window is plenty big) The odd thing about (3) is that typically that call doesn't take much time at all and receives exactly the same amount of data. On a 2Ghz machine that's 300 million instructions worth of time. I am assuming the call doesn't (heaven forbid) wait for the received data to be acked before it returns, so the ack must be waiting for the call to return, or both must be delayed by something else. The problem NEVER happens when there is a second packet of data (part of the same message) arriving between 1 and 2. That part very clearly makes it sound like it has to do with the fact that windows TCP will not send back a no-data ACK until either a second packet arrives or a 200ms timer expires. However the delay is less than 200 ms (its more like 150 ms). The third unseemly character (and to my mind the real culprit) is (5). Send is definitely being called well before that .15 seconds is up, but the data NEVER hits the wire before that ack returns. That is the most bizarre part of this deadlock to me. Its not a tcp blockage because the TCP window is plenty big since we set SO_RCVBUF to something like 500*1460 (which is still under a meg). The data is coming in very fast (basically there is a loop spinning out data via send) so the buffer should fill almost immediately. According to msdn the buffer being full and at least one pending send should cause the data to be sent (though in another place it mentions that there various "heuristics" used in deciding when a send hits the wire). Anway, why the sender doesn't actually send more data during that .15 second pause is the most bizarre part to me. The information above was captured on the receiving side via wireshark (except of course the socket.recv return times which were logged in a text file). We tried changing the send buffer to zero and turning off Nagle on the sender (yes, I know Nagle is about not sending small packets - but we tried turning Nagle off in case that was part of the unstated "heuristics" affecting whether the message would be posted to the wire. Technically microsoft's Nagle is that a small packet isn't sent if the buffer is full and there is an outstanding ACK, so it seemed like a possibility).

    Read the article

< Previous Page | 120 121 122 123 124 125 126 127 128 129 130 131  | Next Page >