Mailman 3 April 2004 - Twisted

Re: [Twisted-Python] no factory.disconnect used the docs!?! (btw thx Itamar )
by exarkun＠divmod.com April 23, 2004

April 23, 2004

On Fri, 23 Apr 2004 12:50:58 -0500, "john nielsen" <jn(a)who.net> wrote: >Once Itamar mentioned closing connections, then I realized I needed to add: > > d.addCallback(lambda _: factory.disconnect()) > > The descriptors are now cleaned up. > > An interesting point is that I cannot find any documentation that > actually uses: factory.disconnect() > > Did I miss it? Where in the docs do you see this? There is no such method on factories. You want to … [View More]

2 1

[Twisted-Python] IFoo(x, persist=1)
by Itamar Shtull-Trauring April 23, 2004

April 23, 2004

I know this is used a lot (the weakref persistence stuff). Should it be option of adaptation or perhaps of adapter registration? Trying to see how to fit this in to zope3's, which doesn't do this, and it'd be nice to have it as part of their core code rather than ours. -- Itamar Shtull-Trauring http://itamarst.org Looking for a job -- http://itamarst.org/resume.html

4 4

[Twisted-Python] no factory.disconnect used the docs!?! (btw thx Itamar )
by john nielsen April 23, 2004

April 23, 2004

Once Itamar mentioned closing connections, then I realized I needed to add: d.addCallback(lambda _: factory.disconnect()) The descriptors are now cleaned up. An interesting point is that I cannot find any documentation that actually uses: factory.disconnect() Did I miss it? john ----- Original Message ----- From: twisted-python-request(a)twistedmatrix.com Date: Thu, 22 Apr 2004 12:00:05 -0600 To: twisted-python(a)twistedmatrix.com Subject: Twisted-Python Digest, Vol 1, Issue 1218 > … [View More]Send Twisted-Python mailing list submissions to > twisted-python(a)twistedmatrix.com > > To subscribe or unsubscribe via the World Wide Web, visit > http://twistedmatrix.com/cgi-bin/mailman/listinfo/twisted-python > or, via email, send a message with subject or body 'help' to > twisted-python-request(a)twistedmatrix.com > > You can reach the person managing the list at > twisted-python-owner(a)twistedmatrix.com > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of Twisted-Python digest..." > > > Today's Topics: > > 1. Re: Transporting pb.Copyable classes? (exarkun(a)divmod.com) > 2. Re: Transporting pb.Copyable classes? (Jasper Phillips) > 3. Unwinding the Twisted Web (a progress report) (David Reid) > 4. Re: Unwinding the Twisted Web (a progress report) > (Itamar Shtull-Trauring) > 5. how to have a server talk to a server (john nielsen) > 6. Re: how to have a server talk to a server (Eric Mangold) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Wed, 21 Apr 2004 18:53:37 GMT > From: <exarkun(a)divmod.com> > Subject: Re: [Twisted-Python] Transporting pb.Copyable classes? > To: "Twisted discussion stuff" <twisted-python(a)twistedmatrix.com> > Message-ID: <20040421185337.9235.1895087205.divmod.quotient.13@ohm> > Content-Type: text/plain > > On Wed, 21 Apr 2004 09:57:47 -0700 (PDT), Jasper Phillips <jasper(a)peak.org> wrote: > > > > [snip] > > > > > > All this seems to me like frantic contortion to avoid a bug in jelly. > > Shouldn't it be able to translate passed class objects into the remote > > version, provided they have been properly registered? Is there some reason > > why not that I am missing? > > Indeed. While I would do this differently, jelly should certainly be able to handle your approach. Let's get into specifics, shall we? Here's some code: > > >>> from twisted.spread import jelly > >>> class Foo: pass > ... > >>> f = Foo() > >>> f.bar = Foo > >>> jelly.jelly(f) > ['__main__.Foo', ['dictionary', ['bar', ['class', '__main__.Foo']]]] > > So far, so good. > > >>> from twisted.spread import pb > >>> class Foo(pb.Copyable): > ... pass > ... > >>> f = Foo() > >>> f.bar = Foo > >>> jelly.jelly(f) > ['__main__.Foo', ['dictionary', ['bar', ['class', '__main__.Foo']]]] > > Hmm, that works too. I expected it not to, actually. I suppose something you're doing with globalSecurity or setCopierForClassTree() may be mucking things up. > > > > > I appreciate the code suggestions, but basicaly I just want to know whether > > this is expected behavior, or whether I should post a demonstrative test > > case and suggest a fix. > > Posting a demonstrative test case would be excellent :) I suspect there is no bug in jelly, just a disagreement in expectations which can easily be resolved once there is a concrete example to talk about. > > Jp > > > > ------------------------------ > > Message: 2 > Date: Wed, 21 Apr 2004 14:03:47 -0700 (PDT) > From: Jasper Phillips <jasper(a)peak.org> > Subject: Re: [Twisted-Python] Transporting pb.Copyable classes? > To: Twisted discussion stuff <twisted-python(a)twistedmatrix.com> > Message-ID: <Pine.LNX.4.53.0404211301540.9678(a)a.shell.peak.org> > Content-Type: TEXT/PLAIN; charset=US-ASCII > > > On Wed, 21 Apr 2004 exarkun(a)divmod.com wrote: > > > On Wed, 21 Apr 2004 09:57:47 -0700 (PDT), Jasper Phillips <jasper(a)peak.org> wrote: > > > > > > All this seems to me like frantic contortion to avoid a bug in jelly. > > > Shouldn't it be able to translate passed class objects into the remote > > > version, provided they have been properly registered? Is there some reason > > > why not that I am missing? > > > > Indeed. While I would do this differently, jelly should certainly be > > able to handle your approach. Let's get into specifics, shall we? > > Here's some code: > > I may end up doing it differently, but one thing at a time! I'm in the > midst of upgrading to new cred, and already have plenty of scope with which > to hang myself. ;-) > > I'm not particularily attached to the static class approach, it just > happened to be what came to mind when prototyping. I'm definitely curious > now what you would do... I've been programming in a vacuum for quite a > while. :-( > > > > >>> from twisted.spread import jelly > > >>> class Foo: pass > > ... > > >>> f = Foo() > > >>> f.bar = Foo > > >>> jelly.jelly(f) > > ['__main__.Foo', ['dictionary', ['bar', ['class', '__main__.Foo']]]] > > > > So far, so good. > > > > >>> from twisted.spread import pb > > >>> class Foo(pb.Copyable): > > ... pass > > ... > > >>> f = Foo() > > >>> f.bar = Foo > > >>> jelly.jelly(f) > > ['__main__.Foo', ['dictionary', ['bar', ['class', '__main__.Foo']]]] > > > > Hmm, that works too. I expected it not to, actually. I suppose > > something you're doing with globalSecurity or setCopierForClassTree() > > may be mucking things up. > > This test case doesn't seem quite right... Why doesn't it fail since Foo > hasn't been registered? Hmmm, looks like jelly()'s taster arg defaults to > DummySecurityOptions()... > > What about registering Foo, and then calling with jelly.globalSecurity?: > ---- Code ---- > from twisted.spread import jelly, pb > > class Foo( pb.Copyable ): pass > jelly.setUnjellyableForClass( Foo, Foo ) > #jelly.globalSecurity.allowInstancesOf( Foo ) # uncomment to fix > > if __name__ == '__main__': > f = Foo() > f.bar = Foo > print jelly.jelly( f, taster=jelly.globalSecurity ) > > ---- End Code ---- > > This gives the Insecure Jelly exception I'm seeing, from jelly.py:482... > However, comment out `f.bar = Foo` or uncomment allowInstancesOf() and it > works. > > Oh, and I'm using twisted 1.2 and python 2.3.3. > > > [snip] > > Posting a demonstrative test case would be excellent :) I suspect there > > is no bug in jelly, just a disagreement in expectations which can easily > > be resolved once there is a concrete example to talk about. > > Is this a decent example? As for a fix, it looks like the solution is there > in SecurityOptions.allowsInstancesOf(), but that it's just not exposed. > > -Jasper > > > > > ------------------------------ > > Message: 3 > Date: Thu, 22 Apr 2004 03:56:47 -0700 > From: David Reid <dreid(a)dreid.org> > Subject: [Twisted-Python] Unwinding the Twisted Web (a progress > report) > To: twisted-python <twisted-python(a)twistedmatrix.com>, twisted-web > <twisted-web(a)twistedmatrix.com> > Message-ID: <1082631407.28634.20.camel@localhost> > Content-Type: text/plain > > I had been trying to straighten out the model for the split since my > first post to twisted-web regarding it. Now that radix, and exarkun > have split twisted.news out of the core codebase as "lowdown". Since > then I have begun work on my own source tree of twisted.web, the module > has been split as "unwound" which is my proposed name as mentioned in > the numerous naming threads (to be installed as tmlabs.unwound if > desired.) > > So far I've done the following: > > * duplicate the source tree renaming the toplevel module and changing > the imports accordingly. > > * run the tests (everything passes just as it does in twisted.web) > > * removed modules that were deprecated BY woven, and added a > Deprecation Warning message to unwound.woven.__init__ > > If there are no complaints about this I'm asking someone (like radix) to > help me complete the rest of the steps necessary to make this the > official split so i can start working on changes and enhancements. The > things I can't do on my own, atleast not as part of twistedmatrix.com is > setup the SVN repo and the project page. > > Also it appears if radix and exarkun did not solve (or atleast did not > implmenet a solution for) the tap naming conflicts. I.E. if lowdown or > tmlabs.lowdown is going to use mktap news or mktap lowdown instead. > > Other than that, Unwound is currently a useable tree of twisted.web At > revision 10601. > > (I'm open to changing the name to either tmlabs.web or some cute/catchy > name other than 'unwound', but ultimately I think that is a policy > decision that falls on the core twisted devs.) > > -- > David Reid > Claim to fame: "I actually wrote code that used twisted.web.webwidgets!" > > > > > ------------------------------ > > Message: 4 > Date: Thu, 22 Apr 2004 11:21:46 -0400 > From: Itamar Shtull-Trauring <itamar(a)itamarst.org> > Subject: Re: [Twisted-Python] Unwinding the Twisted Web (a progress > report) > To: Twisted discussion stuff <twisted-python(a)twistedmatrix.com> > Cc: twisted-web(a)twistedmatrix.com > Message-ID: <1082647306.27964.10.camel@sheriffpony> > Content-Type: text/plain > > On Thu, 2004-04-22 at 06:56, David Reid wrote: > > > So far I've done the following: > > > > * duplicate the source tree renaming the toplevel module and changing > > the imports accordingly. > > >From what I've gathered from talking to Donovan and James this is *not* > how the new twisted.web version is going to be developed. Donovan said > he'd write out a roadmap for the mailing list after talking it over. > > -- > Itamar Shtull-Trauring http://itamarst.org > Looking for a job -- http://itamarst.org/resume.html > > > > > ------------------------------ > > Message: 5 > Date: Thu, 22 Apr 2004 10:46:30 -0500 > From: "john nielsen" <jn(a)who.net> > Subject: [Twisted-Python] how to have a server talk to a server > To: twisted-python(a)twistedmatrix.com > Message-ID: <20040422154630.2989779004F(a)ws1-14.us4.outblaze.com> > Content-Type: text/plain; charset="iso-8859-1" > > I see some examples where you have a client call pb.PBClientFactory to talk to a server which called pb.PBServerFactory. > > What if you have 2 servers (both running pb.PBServerFactory) that accept clients and also want to talk to each other? Would the first server also make a pb.PBClientFactory to talk to the second server (seems like you would have to be careful to not leak descriptors since the reactor would never close). > > There's probably a standard way to do it, that I just missed. > > Thanks for any help, > > john > > -- > ___________________________________________________________ > Sign-up for Ads Free at Mail.com > http://promo.mail.com/adsfreejump.htm > > > > > ------------------------------ > > Message: 6 > Date: Thu, 22 Apr 2004 11:38:55 -0600 > From: Eric Mangold <teratorn(a)world-net.net> > Subject: [Twisted-Python] Re: how to have a server talk to a server > To: Twisted discussion stuff <twisted-python(a)twistedmatrix.com> > Message-ID: <opr6vcm5nfi4eeqj(a)mail.oaktech.net> > Content-Type: text/plain; format=flowed; charset=iso-8859-1 > > On Thu, 22 Apr 2004 10:46:30 -0500, john nielsen <jn(a)who.net> wrote: > > > I see some examples where you have a client call pb.PBClientFactory to > > talk to a server which called pb.PBServerFactory. > > > > What if you have 2 servers (both running pb.PBServerFactory) that accept > > clients and also want to talk to each other? Would the first server > > also make a pb.PBClientFactory to talk to the second server > > Yes. > > > (seems like you would have to be careful to not leak descriptors since > > the reactor would never close). > > Not sure what you mean by that. > > -Eric > > > > > > ------------------------------ > > _______________________________________________ > Twisted-Python mailing list > Twisted-Python(a)twistedmatrix.com > http://twistedmatrix.com/cgi-bin/mailman/listinfo/twisted-python > > > End of Twisted-Python Digest, Vol 1, Issue 1218 > *********************************************** > -- ___________________________________________________________ Sign-up for Ads Free at Mail.com http://promo.mail.com/adsfreejump.htm [View Less]

1 0

[Twisted-Python] server to server leaking descriptors
by john nielsen April 23, 2004

April 23, 2004

I'm trying to figure out server to server communication. Here is example code that works but it leaks descriptors until you run out. In this example, the client talks to server1 on port 7000, server1 talks to server2 on 7001, then the client gets the result from server2 on 7001. Server1 uses pb.PBClientFactory() to send data to server2, which is the part that leaks descriptors. What is the appropriate way for a server to talk to a server? Thanks for any info, john ------------ from twisted.… [View More]

2 1

[Twisted-Python] how to do server to server defined at runtime?
by john nielsen April 23, 2004

April 23, 2004

I am trying to figure out a way to have a server talk to another server, you do not know about except at runtime. In other words, a client tells the server to talk to another server. I figured about the problem I had earlier, and moving the connectTCP outside of PBServerFactory, fixes that problem, which makes sense since it is now not part of the reacotr. But, then I do not know how to specify who to talk to at runtime. For example in this code, how to I move connectTCP into the adder class? … [View More]

1 0

[Twisted-Python] PB and hashed passwords
by Uwe C. Schroeder April 23, 2004

April 23, 2004

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, just found something in pb (twisted version 1.2.0) that's a bit strange. I've written a credentials checker and set self.credentialInterfaces = (credentials.IUsernamePassword,) So far, so good. However the following error is thrown: No checker for twisted.cred.credentials.IUsernameHashedPassword, twisted.cred.credentials.ICredentials, twisted.spread.pb.IUsernameMD5Password I traced it back to pb._PortalAuthChallenger It only implements … [View More]

4 6

[Twisted-Python] how to have a server talk to a server
by john nielsen April 22, 2004

April 22, 2004

I see some examples where you have a client call pb.PBClientFactory to talk to a server which called pb.PBServerFactory. What if you have 2 servers (both running pb.PBServerFactory) that accept clients and also want to talk to each other? Would the first server also make a pb.PBClientFactory to talk to the second server (seems like you would have to be careful to not leak descriptors since the reactor would never close). There's probably a standard way to do it, that I just missed. Thanks … [View More]

2 1

[Twisted-Python] Unwinding the Twisted Web (a progress report)
by David Reid April 22, 2004

April 22, 2004

I had been trying to straighten out the model for the split since my first post to twisted-web regarding it. Now that radix, and exarkun have split twisted.news out of the core codebase as "lowdown". Since then I have begun work on my own source tree of twisted.web, the module has been split as "unwound" which is my proposed name as mentioned in the numerous naming threads (to be installed as tmlabs.unwound if desired.) So far I've done the following: * duplicate the source tree renaming … [View More]

2 1

Re: [Twisted-Python] Transporting pb.Copyable classes?
by exarkun＠divmod.com April 21, 2004

April 21, 2004

On Wed, 21 Apr 2004 09:57:47 -0700 (PDT), Jasper Phillips <jasper(a)peak.org> wrote: > > [snip] > > > All this seems to me like frantic contortion to avoid a bug in jelly. > Shouldn't it be able to translate passed class objects into the remote > version, provided they have been properly registered? Is there some reason > why not that I am missing? Indeed. While I would do this differently, jelly should certainly be able to handle your approach. Let's get … [View More]

2 1

Re: [Twisted-Python] Transporting pb.Copyable classes?
by exarkun＠divmod.com April 21, 2004

April 21, 2004

On Tue, 20 Apr 2004 17:48:00 -0700 (PDT), Jasper Phillips <jasper(a)peak.org> wrote: > > On Wed, 21 Apr 2004, Andrew Bennetts wrote: > > > On Tue, Apr 20, 2004 at 04:20:32PM -0700, Jasper Phillips wrote: > > > > > > I've been passing around instances of pb.Copyable classes for some time, > > > registering them via pb.setCopierForClassTree()... However, this doesn't > > > allow me to pass the classes themselves. > > > > … [View More]> > I dug around in jelly and found jelly.globalSecurity.allowInstancesOf(), > > > which seems to do what I want, but makes me nervious as it's pretty > > > buried... > > > > > > Passing classes seems like a reasonable thing to do, but now I wonder if > > > there is some reason that jelly doesn't allow it (except via a buried > > > method call)? Or is this just an accidental misfeature? > > > > Why not just pass the name of the class, and unserialise that as > > appropriate? > > > > I'm not a PB expert (or even close), but I always lean towards passing data > > rather than code (or code-like things like classes) over remote method calls > > when I can :) > > That's certainly feasible, but then I have an extra unserialize step > cluttering my code as the classes passed have attributes I want to access. > In general I'm passing game state which exists naturally as code, and while > I could always serialize and unserialize it, why bother if I don't have to? > > Plus, isn't unserializing classes based upon their name exactly what jelly > does? > > On the other hand it's debateable whether the things I'm passing _should_ be > classes at all (they represent a type of terrain, with statically accessible > fields describing their game effects). However, the quality of my code is > really a seperate issue. > Just to be sure I'm reading this correctly, you have something like... class Foo: bar = [] Foo.bar.append(baz) someRemoteRef.callRemote('doStuff', Foo) where you want 'remote_doStuff' to receive a class object whose "bar" attribute is [baz] (or possibly [RemoteReference(baz)], or *something* containing baz)? If not, disregard the rest of my post :) Otherwise ... You're going to have difficulty using PB with this approach. PB can serialize classes, but it does so in exactly the way Andrew described: it looks up their name and sends it, and the other side turns the name back into an object. Any mutable class state you change on one side will *not* be reflected on the other side. I'd recommend moving your mutable state off these class objects and onto instances of some sort. Otherwise, you will have to implement special jelliers/unjelliers for the class objects you want treated specially. Jp [View Less]

3 3