Mailman 3 June 2002 - Twisted

[Twisted-Python] Licence of examples
by Benjamin Bruheim 21 Jun '02

21 Jun '02

Hello, After some speculation I stumbled across a little nut, what happens if I use LGPLed examples. Are these a part of the library and thus define the code itself as a library, or will they give Glyph a pica-percentage copyright on huge works based on the code? So, there's this file LICENSE that everyone must read through for fun. There's also an EULA coming up each time I install something on windows that I've never read through, except for fun. However, there are people believing in such inane matters, so I found this portion: 5. A program that contains no derivative of any portion of the Library, but is designed to work with the Library by being compiled or linked with it, is called a "work that uses the Library". Such a work, in isolation, is not a derivative work of the Library, and therefore falls outside the scope of this License. That states: The examples are not a part of the library, thus they are just copyrighted by glyph. So, everyone who used the pb-echo-examples, Glyph owns that code (partly). I don't know what rights he inherently has, but he has the copyright. What kind of licence should the example have? I would like to place it in Public Domain, because if stuff can be taken from there, then we should be able to place stuff there as well. Benjamin Bruheim benjamin(a)inout.no

4 4

[Twisted-Python] new doc generator, new doc
by Glyph Lefkowitz 21 Jun '02

21 Jun '02

I have just spent some time improving our documentation preprocessor as well as writing a new document and fixing some old ones. Everything's in CVS, though using generate-domdocs may be a little bit confusing, so I've put the relevant documents up on the web, too. http://zelda.twistedmatrix.com/doc/howto/plugin http://zelda.twistedmatrix.com/doc/howto/defer http://zelda.twistedmatrix.com/doc/howto/servers Of course, if you're bitten by the documentation bug too, examples for the source to all of these documents is in CVS in doc/howto/. Feel free to contribute one of your own. -- | <`'> | Glyph Lefkowitz: Traveling Sorcerer | | < _/ > | Lead Developer, the Twisted project | | < ___/ > | http://www.twistedmatrix.com |

2 1

[Twisted-Python] Generic hierarchy interface
by Jp Calderone 20 Jun '02

20 Jun '02

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Its time has come. The need is pretty clear, partially for blog/news and a lot for Reality, and I'm sure for a lot of other things I'm not aware of. Basically, what seems to be needed is an interface that allows for a protocol-defined hierarchy to protocol-specific data, organised based on protocol-specific meta-data. For example, consider email. The protocol-specific hierarchy describes message threads (which you are all hopefully familiar with). The meta-data is a portion of the xheaders ("References" and "In-Reply-To" in particular), and the data is the message body. The email implementation for this hierarchy interface would then need to parse headers and xheaders to determine the relationship, and make the appropriate messages available this available at the appropriate places. The appropriate places. There seem to be three basic relationships that are of interest: parent, child, and sibling. Accessor methods for these should be straightforward enough, and implementions would need to provide the details for each protocol of extracting them from the meta-data available. The other type of access would be absolute and relative "path" access. For example, "news://a.server.name/a.group.name/a.message.id". A Reality example: "reality://server/Sol/Earth/Kansas/Smallville/Clark_Kent" In this case, each level of depth would only need to implement a standard accessor interface to allow whatever it may contain to be accessable. I'm a little out of my element, so I hope this has been at least somewhat coherent ;) Anyone have comments, or notice and horrible flaws, or just see a plain better way to go about it? Jp - -- "One World, one Web, one Program." - Microsoft(R) promotional ad "Ein Volk, ein Reich, ein Fuhrer." - Adolf Hitler - -- 4:20pm up 27 days, 17:02, 5 users, load average: 0.00, 0.00, 0.00 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.7 (GNU/Linux) iD8DBQE9Dk7kedcO2BJA+4YRAnL/AJ45iIwfmCQSU4DXCPXtXbRhBLFTfACfYLO6 9fhOCu7iEG8mbOeTrsOdJr4= =zSD+ -----END PGP SIGNATURE-----

4 3

[Twisted-Python] A Farewell to Arms
by Glyph Lefkowitz 19 Jun '02

19 Jun '02

As detailed in my previous message, the .arm() method and patterns which required its use with respect to Deferreds were problematic: http://twistedmatrix.com/pipermail/twisted-python/2002-May/001117.html This is no longer the case. Rejoice! For the average application author, this means that a confusing and obscure corner of the API has been swept out of their view, where it never should have been in the first place. For the Twisted systems programmer, this means that you have more predictable and consistent control over execution of a Deferred's processing chain regardless of whether you "own" it or not. It is all around a good thing. As most good changes do, this resulted in a reduction of the total amount of code in Twisted by a nontrivial number of lines. It also fixed some previously intractible bugs relating to AOT and Marmalade being unable to deal with copy_reg registered types correctly. Even better, the document in CVS (doc/howto/defer.html) has been updated to describe this behavior, down to the subtle details of implementation. -- | <`'> | Glyph Lefkowitz: Traveling Sorcerer | | < _/ > | Lead Developer, the Twisted project | | < ___/ > | http://www.twistedmatrix.com |

1 0

[Twisted-Python] DICE Job - SENIOR GAME SYSTEMS DESIGNER
by Itamar Shtull-Trauring 17 Jun '02

17 Jun '02

<http://www.dice.com/DandL/s/stuart.01T0SR_DEs.html>

2 2

[Twisted-Python] twisted.enterprise.row
by Andrew Bennetts 16 Jun '02

16 Jun '02

Quick question: Why doesn't twisted.enterprise.row use DB-API's Cursor.description feature, instead of PostgreSQL specific gunk? A cursory glance at what row needs and what Cursor.description provides suggests that you could and should use that. -Andrew.

1 0

[Twisted-Python] pb.setFactoryForClass()
by Sean Riley 16 Jun '02

16 Jun '02

proposal for new functionality in pb: In addition to the existing pb.setCopierForClass() I'd like a pb.setFactoryForClass() which would take a factory method instead of a class as its argument. This would allow serialized objects to be re-constructed on the other end of network connections by a factory instead of just be creating a new class. I have a use case where objects are created locally on both the client and server and these objects need to be passed as arguments to remote method calls. The best way to do it seems to be to send a unique ID over the wire and have a factory method transform the ID into the correct object on the other side. A trivial example: class Thing(pb.Copyable) allIDs = {} def __init__(self, id): self.id = id Thing.allIDs[id] = self def getStateToJellyFor(self, perspective): return {"id":self.id} def thingFactory(state): return Thing.allIDs.get(state["id"],None) pb.setFactoryToCopyFor("Thing", thingFactory) Things could then be send over the network as very small objects (just IDs) as long as both side have already loaded all the "Things" from local storage. ---- "If it's not running programs or fusing atoms, it's just bending space.", Ken Macleod "That's it, I'm outta here.", Homer Simpson's Brain Sean Riley sean(a)ninjaneering.com

2 1

[Twisted-Python] BUG: setting timeout on clientTCP() will invoke connectionFailed twice
by Ivo van der Wijk 15 Jun '02

15 Jun '02

Hi, The following code: -- from twisted.protocols.http import HTTPClient from twisted.internet import reactor class foo(HTTPClient): def connectionFailed(self): print "Failed" reactor.clientTCP("localhost", 8888, foo(), 2) reactor.run() -- Will print "Failed" twice. Without the timeout, only one "Failed" will be printed (of course make sure you connect to something that fails immediately when testing this :) It's kind of silly to keep te timeout after the connection has already failed - it should be cancelled, right? Cheers Ivo -- Drs. I.R. van der Wijk -=- Brouwersgracht 132 Amaze Internet Services V.O.F. 1013 HA Amsterdam, NL -=- T +31-20-4688336 F +31-20-4688337 Linux/Web/Zope/SQL/MMBase W http://www.amaze.nl E info(a)amaze.nl Network Solutions W http://vanderwijk.info E ivo(a)amaze.nl Consultancy PGP http://vanderwijk.info/pgp -=-

1 0

[Twisted-Python] small README changes
by Bruce Mitchener 15 Jun '02

15 Jun '02

Hey all, I started to check out Twisted today by running some of the stuff out of the distribution rather than only reading docs. I'd noted a few problems or unclear points in that process. The attached patch addresses most of them. Is the setup.py needed for anything if you aren't going to install Twisted into site-libs? Also, does one of the FTP tests fail for everyone? One of the marmalade tests failed on Python 2.2 (error about the NodeList type). Cheers, - Bruce Index: README =================================================================== RCS file: /cvs/Twisted/README,v retrieving revision 1.87 diff -u -r1.87 README --- README 2 Jun 2002 04:38:28 -0000 1.87 +++ README 15 Jun 2002 06:15:38 -0000 @@ -44,8 +44,8 @@ What can I do with it? - To install Twisted, just make sure the Twisted-$VERSION/twisted directory - is in the "PYTHONPATH" environment variable. For example, if you extracted + To install Twisted, just make sure the Twisted-$VERSION directory is in + the "PYTHONPATH" environment variable. For example, if you extracted Twisted-0.18.0.tar.gz to /home/bob/, then you would have something like: export PYTHONPATH=$PYTHONPATH:/home/bob/Twisted-0.18.0/ @@ -79,6 +79,9 @@ % admin/runtests + Some of the tests may fail if you are running Python 2.2 or don't have + the Python XML support installed. + If you're feeling more brave, you can try the new "acceptance tests". These require some setup and are mainly for the developers to decide if it's OK to release, but: @@ -95,6 +98,8 @@ Note that the following examples only create the .tap files with the servers inside of them: to actually run the servers, see "Running Servers" below. + Also note that the commands used (mktap, coil, twistd, etc) can be found + in the bin directory of your Twisted installation. The normal type of server to create is a webserver. You can run this command::

1 0

Re: [Twisted-Python] no 'headers' in server.Request instance problem
by Jon Dyte 13 Jun '02

13 Jun '02

glyph(a)twistedmatrix.com wrote :-> From: Itamar Shtull-Trauring <twisted(a)itamarst.org> > Subject: Re: [Twisted-Python] no 'headers' in server.Request instance problem > Date: Thu, 13 Jun 2002 01:37:41 +0200 > > > BTW - you don't need to use twisted.web.server.Request - since the > > refactor to HTTP you only need twisted.protocols.http.Request, no need > > to use twisted.web unless you are using it as a web framework. > > Actually you don't need to use Request at all. If you want this to work with > other twisted.web servers, it would be best to structure your work as additions > to Resource, not Request. > > What functionality are you adding at the Request level? > > > -- > | <`'> | Glyph Lefkowitz: Traveling Sorcerer | > | < _/ > | Lead Developer, the Twisted project | > | < ___/ > | http://www.twistedmatrix.com | I'm not really adding any functionality. I was merely making a previous Twisted/Quixote script work with Twisted-0.18.0, which is how I found the reported problem. This script is so short I include it here. Jon """qserv Demo of an HTTP server built on top of Twisted Python that serves a Quixote application, and also supports an SMTP interface. """ # created 2002/03/19, AMK __revision__ = "$Id$" import string, cStringIO from quixote.http_request import HTTPRequest from quixote.publish import Publisher from twisted.cred.service import Service from twisted.internet.app import Application from twisted.mail import mail from twisted.protocols import protocol, http from twisted.web import server # Ports this server will listen on HTTP_PORT = 8098 SMTP_PORT = 8099 import quixote quixote.enable_ptl() publisher = Publisher('quixote.demo') publisher.read_config('quixote.conf') publisher.setup_logs() class QuixoteTWRequest (http.Request): def process(self): environ = self.create_environment() ## this seek is important self.content.seek(0,0) stdin = self.content request = HTTPRequest(stdin, environ) stdout = cStringIO.StringIO() stderr = cStringIO.StringIO() publisher.publish(stdin, stdout, stderr, environ) output = stdout.getvalue() #self.write(output) #self.setResponseCode(request.response.status_code) #for hdr, value in request.response.headers.items(): # self.setHeader(hdr, value) #self.finish() ## as per message on quixote list self.startedWriting = 1 self.write(output.replace("Status:", "HTTP/1.1")) self.finish() def create_environment (self): host = self.getHeader('host') or self.getHost() if ':' in host: serverName = string.split(host, ':')[0] else: serverName = host content_type = self.getHeader('content-type') typ, remote_addr, remote_port = self.transport.getPeer() env = { "PATH_INFO": self.path, "SERVER_SOFTWARE": server.version, "SERVER_NAME": serverName, "GATEWAY_INTERFACE": "CGI/1.1", "REMOTE_ADDR": remote_addr, "REMOTE_PORT": str(remote_port), "SERVER_PROTOCOL": self.clientproto, "SERVER_PORT": str(HTTP_PORT), "REQUEST_METHOD": self.method, "SCRIPT_NAME": "", "SCRIPT_FILENAME": "", "REQUEST_URI": self.uri, "CONTENT_TYPE": content_type, } return env class QuixoteFactory (http.HTTPFactory): def buildProtocol (self, addr): h = http.HTTPChannel() h.requestFactory = QuixoteTWRequest h.factory = self return h def run (): app = Application('Quixote') app.listenTCP(HTTP_PORT, QuixoteFactory()) app.run(save=0) if __name__ == '__main__': run()

1 0