[twsocket] BNF2TCP creates socket automatons from a protocol's BNF
Hello, everybody! Back in 2004 we have held here, on the twsockets mailing list, a discussion regarding a possibility of implementing a general automaton capable of handling almost any communication protocol using its BNF description. I started to work on this idea in a short time after the discussion, but unfortunately had to stop it after a while, and it has been suspended for several years. Now I've managed to resume the process, and may announce a pre-alpha product BNF2TCP. If someone is interested in the project and has a will to participate, I can share with them the source codes under the terms of non-disclosure agreement. Public distribution is not planned until beta stage. The program is written in C++, using (a little bit outdated) ICS twsockets for communications, which can be easily replaced with latest components. The program plays a role of a BNF editor/verifier/debugger and contains core classes bringing the magic into life: a LL-parser and controllers (automaton engines). The program utilizes a kinda extreme programming style and has a lot of legacy code (don't forget the years of the cease), so it needs a refactoring for adherence to good design principles and extension according to requirements and use cases which were most likely overlooked by myself. For testing purposes simplified versions of a POP3-client and server have been made. Here is how an excerpt of the BNF looks like: ... SERVER ::= Start WaitAckListen WaitAckListen ::= AckListen WaitCmdConnect WaitCmdConnect ::= CmdConnect WaitAckAccept WaitAckAccept ::= AckAccept SessionOpt | AckError Stopping SessionOpt ::= Login | SomehowDisconnect Login ::= CmdUser LoginAck LoginAck ::= AckOK PasswordOpt | AckError SessionOpt PasswordOpt ::= CmdPass PassAck | SomehowDisconnect PassAck ::= AckOK Transaction | AckError SessionOpt Transaction ::= CmdList ListAck | CmdStat StatAck | CmdRetr RetrAck | SomehowDisconnect ListAck ::= AckListResult Transaction | AckError Transaction StatAck ::= AckStatResult Transaction | AckError Transaction RetrAck ::= AckRetrResult Transaction | AckError Transaction QuitAck ::= AckOK Stopping SomehowDisconnect ::= CmdQuit QuitAck | Stop Stopping ::= AckDisconnect Stop Stop ::= CmdDisconnected Idle Idle ::= AckShutdown Down | AckStandBy WaitCmdConnect Down ::= AckEnd // = CmdList ::= 'LIST' OptParam %x0D %x0A OptParam ::= %x20 %msgnum% %msgnum% ::= number number ::= DIGIT digits digits ::= number | _ DIGIT ::= %x30-%x39 +Delimiters=None . ... This is a portion of source code: // create automaton pSocketCtrl = new SocketController; // set some event handlers pSocketCtrl-OnGetSubSource = OnGetSubSourceEvent; pSocketCtrl-OnGetParam = OnGetCtrlParam; pSocketCtrl-OnSetParam = OnSetCtrlParam; pSocketCtrl-OnDisplay = OnCtrlDisplay; pSocketCtrl-OnError = OnCtrlError; // load BNF pSocketCtrl-LoadGrammar(asBNFGrammarFile); // run pSocketCtrl-Start(); I don't have a documentation yet, but can answer your questions. You can contact me directly via e-mail, or on the list. What I plan to do next is a BNF of IMAP client as a proof of concept. Best wishes, Stan -- To unsubscribe or change your settings for TWSocket mailing list please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be
Re: [twsocket] HTTP Location
Hi, Angus, I'm not sure if it's valid, but it is used by server-side scripts for assigning some cookies and reloading the same page under new conditions: Location: / It's up to the site to prevent looping. Stan - Original Message - From: Angus Robertson - Magenta Systems Ltd an...@magsys.co.uk To: twsocket@elists.org Sent: Thursday, May 20, 2010 1:12 PM Subject: [twsocket] HTTP Location Is it valid for an HHTP header to relocate to an empty URL? Currently, HttpProt seems to parse the relocation to the same URL, and goes into a loop. HTTP/1.1 302 Found Location: http:// Temporary Redirection: http://www.telecomstrader.com/ to: http://www.telecomstrader.com/ Angus -- To unsubscribe or change your settings for TWSocket mailing list please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be -- To unsubscribe or change your settings for TWSocket mailing list please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be
Re: [twsocket] HTTP Location
Not exactly what I said ;-) I said that irrespective to what RFC could imply, some sites intentionally use redirection to the same location, and the site (server) should normally prevent endless looping. If it's not, than indeed a client should protect itself - whether it is implemented in HttpProt or outside. Other fields of http-header except Location should be also considered for proper loop detection. For example, the URL checker should accept and preserve cookies between requests and alert looping only if cookies do not change. It is not clear to me from the spec, but if it makes the express assertion that it SHOULD give the new URI in the response to non-HEAD requests, by omission it seems to me that it then MUST do so for HEAD. Either way, I wouldn't rely on a URI being given, ever. Like Stan said, it is up to the client to avoid loops. dZ. -- DZ-Jay [TeamICS] http://www.overbyte.be/eng/overbyte/teamics.html -- To unsubscribe or change your settings for TWSocket mailing list please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be -- To unsubscribe or change your settings for TWSocket mailing list please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be
[twsocket] Filtering HTTP content - an approach to choose
Hi, everybody! I need to implement a kind of HTTP content filter for an end-user PC. The easiest way is probably using a proxy like GpHTTPProxy. But it's a little bit difficult to integrate the proxy into all existing browsers and ensure that it is not switched off by unathorized user. Another option is to build some stuff over WinPcap, but it's not good because the application will require 3-d party product to be installed. I'd prefer to build a self-contained application, capable of analyzing traffic as firewalls and anti-viral software do. Could someone suggest a source to start from? Any other approaches to mention? Thanks in advance. Stan -- To unsubscribe or change your settings for TWSocket mailing list please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be
Re: [twsocket] WSAECONNABORTED (10053) when connecting to server
What is the servers you are trying to connect to? Microsoft products often use abortion as a preferred method for disconnecting, i.e. they don't use graceful close. As for the reason why the servers decided to disconnect - this is another question. Suppose, some restriction on a number of concurrent connections may by applied. -Oorspronkelijk bericht- Van: twsocket-boun...@elists.org [mailto:twsocket-boun...@elists.org] Namens Anton Sviridov Verzonden: dinsdag 27 april 2010 12:00 Aan: twsocket@elists.org Onderwerp: [twsocket] WSAECONNABORTED (10053) when connecting to server Hello all, I faced strange issue in my app. I develop an application which opens ~15 connections to several servers, ~3-5 connections per server. After establishing the connection it receives data flow. But when I start it, some connections fail with WSAECONNABORTED (10053) - and the oddest thing is that they fail after receiving some data. Reconnect after 5 sec pause works fine. I asked our admin, but he says the PC which my app runs on has no restrictions on the router at all. Moreover, I have no problem on my own PC with XP (PC where I run app has W2003 server). So it is unlikely bug in my or ICS code. I found description of MS bug here: http://support.microsoft.com/?kbid=931319 and applied the patch provided, but it only turned the things worse so I uninstalled it. Do you have any advice? Maybe someone have already faced with such issue? -- Anton -- To unsubscribe or change your settings for TWSocket mailing list please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be -- To unsubscribe or change your settings for TWSocket mailing list please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be -- To unsubscribe or change your settings for TWSocket mailing list please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be
Re: [twsocket] GpHTTPProxy v2.4 (the patched 2.0)
Unfortunately, I can't do that. As you may recollect, it took approx 1.5 years for me to find a free time for fixing the bug. ICS has changed tremendously during this period, and I did not follow the changes, alas. One of the reasons for this is that the old version works stable and upgrade is needless. As for the bug itself, the last discussion seems dated by september 2008, under the subject GpHTTPProxy doesn't work correctly, so anybody who wants to know about the issue, can search archives and read. - Original Message - From: Francois PIETTE francois.pie...@skynet.be To: ICS support mailing twsocket@elists.org Sent: Wednesday, April 21, 2010 7:59 PM Subject: Re: [twsocket] GpHTTPProxy v2.4 (the patched 2.0) If anybody wants to continue work on the GpHTTPProxy I would be happy to revoke any rights on that unit (except as the original author). I would also change the license if required . Maybe Stanislav will take the lead on this component ? -- francois.pie...@overbyte.be The author of the freeware multi-tier middleware MidWare The author of the freeware Internet Component Suite (ICS) http://www.overbyte.be -- To unsubscribe or change your settings for TWSocket mailing list please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be
[twsocket] GpHTTPProxy v2.4 (the patched 2.0)
Hello everybody, In the past some people have posted messages here exposing their interest in a HTTP-proxy. They tried to use the GpHTTPProxy component, but it has one critical bug with persistent connections. Now it seems that I managed to fix the bug using the proprietary sources provided by Primoz Gabrijelcic as a reference for me by a non-disclosure agreement. As the agreement concerns the proprietary sources only, and the patched source is the GpHTTPProxy indeed, I suppose the agreement does not cover the patched version and it can be published. Moreover, Primoz seems not having time (long since) for patching the stuff himself, and does not respond. So, I think I could post or submit the patched version to the overbyte site. The version is not thoroughly tested, and I can't provide a support for it, because I work on it in my spare time. Also, the patch does not deal with chunked encoding and leaves it unsupported. Please, note that the original version of GpHTTPProxy used for the patch is outdated with regard to ICS itslef, though this is the latest version of GpHTTPProxy publicly available. I will NOT port the GpHTTPProxy v.2.4 to the latest ICS version in the near future. Francois, Wilfried, Arno, Angus, fellows - what do you think? Best wishes, Stan Korotky. -- To unsubscribe or change your settings for TWSocket mailing list please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be
Re: [twsocket] Full URL decode (unescape)
I've looked at the module before writing the initial post. What do you mean exactly? There is no a function for unescaping. - Original Message - From: Francois PIETTE francois.pie...@skynet.be To: ICS support mailing twsocket@elists.org Sent: Sunday, January 10, 2010 7:05 PM Subject: Re: [twsocket] Full URL decode (unescape) Have a look at OverbyteIcsUrl unit. -- francois.pie...@overbyte.be http://www.overbyte.be - Original Message - From: Stanislav stas...@orc.ru To: ICS support mailing twsocket@elists.org Sent: Sunday, January 10, 2010 12:07 AM Subject: [twsocket] Full URL decode (unescape) Hi everyone, Perhaps, this is a silly question, but I don't see a function for unescaping URLs. HttpSrv module does have TextToHtmlText and RemoveHtmlSpecialChars functions, but does not have anything for reverse translation. UrlDecode does not look at html entities either. Of course, we can code such things in our applications on our own, but why doesn't HTTP server component (or IcsUrl) provide unescaping, if it provides escaping? If I overlooked it, please, point me out the place where to look at. Best wishes, Stanislav. -- To unsubscribe or change your settings for TWSocket mailing list please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be -- To unsubscribe or change your settings for TWSocket mailing list please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be -- To unsubscribe or change your settings for TWSocket mailing list please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be
Re: [twsocket] Full URL decode (unescape)
As I wrote, there is no a method to provide the reverse functionality for TextToHtmlText and RemoveHtmlSpecialChars, i.e. unescaping html entities, such as amp; etc., into single characters. By the way, both names of the functions seems to be misleading, IMHO. TextToHtmlText should read something like HtmlEntityEncode or HtmlSpecialCharsEncode. RemoveHtmlSpecialChars looks like a stripped down version of the TextToHtmlText. Why do we have both? With the existing names used for back compatibility, I'd suggest to introduce new alias (either of abovementioned). And add HtmlEntity/SpecialCharsDecode of course. - Original Message - From: Francois PIETTE francois.pie...@skynet.be To: ICS support mailing twsocket@elists.org Sent: Monday, January 11, 2010 9:35 PM Subject: Re: [twsocket] Full URL decode (unescape) UrlDecode ? If this doesn't fit your need, please give an example of what you need. -- francois.pie...@overbyte.be The author of the freeware multi-tier middleware MidWare The author of the freeware Internet Component Suite (ICS) http://www.overbyte.be - Original Message - From: Stanislav stas...@orc.ru To: ICS support mailing twsocket@elists.org Sent: Monday, January 11, 2010 9:15 AM Subject: Re: [twsocket] Full URL decode (unescape) I've looked at the module before writing the initial post. What do you mean exactly? There is no a function for unescaping. - Original Message - From: Francois PIETTE francois.pie...@skynet.be To: ICS support mailing twsocket@elists.org Sent: Sunday, January 10, 2010 7:05 PM Subject: Re: [twsocket] Full URL decode (unescape) Have a look at OverbyteIcsUrl unit. -- francois.pie...@overbyte.be http://www.overbyte.be - Original Message - From: Stanislav stas...@orc.ru To: ICS support mailing twsocket@elists.org Sent: Sunday, January 10, 2010 12:07 AM Subject: [twsocket] Full URL decode (unescape) Hi everyone, Perhaps, this is a silly question, but I don't see a function for unescaping URLs. HttpSrv module does have TextToHtmlText and RemoveHtmlSpecialChars functions, but does not have anything for reverse translation. UrlDecode does not look at html entities either. Of course, we can code such things in our applications on our own, but why doesn't HTTP server component (or IcsUrl) provide unescaping, if it provides escaping? If I overlooked it, please, point me out the place where to look at. Best wishes, Stanislav. -- -- To unsubscribe or change your settings for TWSocket mailing list please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be
[twsocket] Full URL decode (unescape)
Hi everyone, Perhaps, this is a silly question, but I don't see a function for unescaping URLs. HttpSrv module does have TextToHtmlText and RemoveHtmlSpecialChars functions, but does not have anything for reverse translation. UrlDecode does not look at html entities either. Of course, we can code such things in our applications on our own, but why doesn't HTTP server component (or IcsUrl) provide unescaping, if it provides escaping? If I overlooked it, please, point me out the place where to look at. Best wishes, Stanislav. -- To unsubscribe or change your settings for TWSocket mailing list please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be
[twsocket] Smoothed and weighted round robin distribution
Hi, everybody, I need to implement a kind of weighted round robin distribution for HTTP ICS-based reverse-proxy. I've found some info about packets distribution but neither about users (sessions) distribution. The difference is that, according to the algorithms, packets can be left in a queue (so load balancing is achieved), while users should not. The given data is a list of N servers and their relative weights W[n]. Could someone suggest a link for appropriate algorithm? I tried to use the well-known round robin modification of constructing a virtual cicrle, divided into N sectors, each of which is proportional to corresponding server weight. But it may produce very uneven distribution in time, that is for a period it draws a series with some servers probabilities far from their weights, then it begins to draw a series with other servers probabilities far from their weights. I could show an example, but it's not necessary, I think. For a large period the mean values approximate to the weights. But I'd like to have a short time smooth distribution. Thanks in advance. Best wishes, Stanislav Korotky. -- To unsubscribe or change your settings for TWSocket mailing list please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be
Re: [twsocket] Clustering/failover support for ICS servercomponents--especially Web Server
Hello, Don't you think that reverse-proxy is sufficient for that purposes, with no need to modify servers themselves? Failover facilities should be external to a server, imho, because they may fail if they are embedded into a server which undergoes a failure. Best wishes, Stanislav Korotky. - Original Message - From: Fastream Technologies ga...@fastream.com To: ICS support mailing twsocket@elists.org Sent: Friday, October 23, 2009 11:29 AM Subject: [twsocket] Clustering/failover support for ICS servercomponents--especially Web Server Hello, I wonder if this is in somebody's to-do list. Don't you see the demand I see in this? If so, maybe we can collect money similar to what we did for the ICS-SSL project which succeeded. Best Regards, SubZero -- To unsubscribe or change your settings for TWSocket mailing list please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be -- To unsubscribe or change your settings for TWSocket mailing list please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be
Re: [twsocket] TPop3Cli is halted while receiving a message over 4Mb
(taking into account that they disclose themselves in such a negligible, infrequent manner)? Do you mean that it doesn't always occur with the same 4MB message ? Yes. Today the program fetches the same message (it remains on a server) successfully (several times for testing). Now I can't investigate the matter, alas. Best wishes, Stanislav Korotky, Russia, Moscow, CET +02:00 -- [EMAIL PROTECTED] http://www.overbyte.be -- To unsubscribe or change your settings for TWSocket mailing list please goto http://www.elists.org/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be
[twsocket] TPop3Cli is halted while receiving a message over 4 Mb
Hello All! I use TPop3Cli in my application. I thought it worked well so far, but I've just stumbled on a strange thing. The application tries to fetch new message with a size over 4 Mb. According to a progress indicator I have in the application the message is downloaded till 97% but the rest of it never arrives, i.e. there is no more data available to an underlying socket, and the application waits the missing data for ever (actually, timeout). An ordinary mail client fetches the same message without a problem. I do understand that this seems to be a bug in my application, but considering that the application successfully works with million (!) of messages since several last years (i.e. it is tested to a certain degree) and does not produce any exceptions under debugger I don't have an idea how to track down problems with such a message. Could someone imagine mistakes of any kind which could lead to similiar symptoms (taking into account that they disclose themselves in such a negligible, infrequent manner)? Best wishes, Stanislav Korotky, Russia, Moscow, CET +02:00 -- To unsubscribe or change your settings for TWSocket mailing list please goto http://www.elists.org/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be
[twsocket] How to determine DNS settings?
Hello All! Is it possible to determine current DNS address used by a system if DNS is obtained automatically, that is DNS is not written into the registry? Best wishes, Stanislav Korotky, Russia, Moscow, CET +02:00 -- To unsubscribe or change your settings for TWSocket mailing list please goto http://www.elists.org/mailman/listinfo/twsocket Visit our website at http://www.overbyte.be
