Mailman 3 July 2002 - Twisted

[Twisted-Python] commits
by Christopher Armstrong 14 Jul '02

14 Jul '02

Hey guys, I just wanted to say "holy crap" -- great job on all the (code AND documentation!) commits that I woke up to today. :-) The only problem is, Twisted releases are going to get behind if you guys keep this productivity up!! Guess I'll have to release again in a couple of days! ^.^' -- Chris Armstrong << radix(a)twistedmatrix.com >> http://twistedmatrix.com/users/carmstro.twistd/

2 1

[Twisted-Python] DeferredList and errback
by Andrew Bennetts 14 Jul '02

14 Jul '02

DeferredList is handy, but it isn't quite what I need for FTPClient. Here's my use-case: A file is deemed to be downloaded by FTPClient when both the 226 transfer complete message is received on the control connection, *and* the data connection is closed. DeferredList takes care of this nicely, so long as it works. In the case where the control connection immediately returns an error (e.g. file not found), one errback will be called and the other Deferred will never be called either way, effectively hanging the client. (I've got a testcase for this, but I haven't checked it in yet because having FTP tests passing is a novelty that I don't want to take away from people ;) So ideally I want DeferredList (or something like DeferredList) to wait for all callbacks, as it currently does, but fire its errback as soon as any of its Deferreds fires an errback. Later errbacks and callbacks would be ignored. I realise I'm probably too late to get this into 0.19.0, but I'm wondering what the right way to provide this is: - new class in defer.py (what would it be called?), - change the behaviour of DeferredList to do what I want, or - add a flag to the current DeferredList so it can do either behaviour? -Andrew.

2 3

[Twisted-Python] need an identity
by Sean Riley 13 Jul '02

13 Jul '02

I seem to have a problem... I need access to the identity for a client as their perspective is being created from within getPerspectiveRequest(). Although Service::getPerspectiveRequest is called by the identity instance for the client, when using the DBAuthorizer, there doesnt appear to be a way to access this identity object from within getPerspectiveRequest. The only way to lookup an identity is Authorizer::getIdentityRequest - which in this case will hit the database again to load the identity. I want the existing identity instance, not a new instance for the same identity. I know the Identity instance is out there, but I can't get at it! The DbAuthorizer does not keep an "identities" list around, so I cant look up the identity the same way you can in the DefaultAuthorizer. Is this a security issue? Is access to the identity information for a perspective not allowed? One solution would be for me would be for the identity to be passed into getPerspectiveRequest as another argument: def getPerspectiveRequest(self, name, identity): instead of the existing: def getPerspectiveRequest(self, name): This is may be too drastic an API change for just this one case... but it doesnt require changing much existing code... thoughts? ---- Sean Riley

2 1

[Twisted-Python] cBanana diffs
by Bruce Mitchener 13 Jul '02

13 Jul '02

I've made a couple of local changes to cBanana to look at improving the performance of it some. The diff is available from: http://day.cubik.org/~bruce/spread.diff The change from the malloc/memcpy/free sequence to a realloc should be pretty clear. The remainder of the changes are intended to let us allocate the python list objects in the correct size (since we know that) and to then use PyList_SET_ITEM() rather than PyList_Append(). In theory, that should help, in practice, it didn't much on simple tests. One thing that was taking a while in simple runs was the handling of LONGINT and LONGNEG values (which call back into Python rather than using C code). I've not yet optimized that at all as I don't yet have a good enough understanding. I'll also run this new cBanana.c under valgrind or Purify this week to make sure that I didn't introduce any memory leaks. Comments and suggestions are welcome. - Bruce

2 5

[Twisted-Python] IM bug bounty
by Glyph Lefkowitz 13 Jul '02

13 Jul '02

There's a but in t-im that is annoying the heck out of me and I can't figure out. If you're a Twisted hacker you can gain major brownie points by submitting a patch that fixes it. When you tab-complete a nickname on the second (wrapped) line of a t-im input box, the whole program freezes and won't respond to signals. -- | <`'> | Glyph Lefkowitz: Traveling Sorcerer | | < _/ > | Lead Developer, the Twisted project | | < ___/ > | http://www.twistedmatrix.com |

1 0

RE: [Twisted-Python] Adding Mime types using coil
by Lewin, Karl E 12 Jul '02

12 Jul '02

I'll do my best to come up with some! Are there any "gotcha's" re: win32 I should be aware of? -----Original Message----- From: Andrew Bennetts [mailto:andrew-twisted@puzzling.org] Sent: Friday, July 12, 2002 11:40 AM To: twisted-python(a)twistedmatrix.com Subject: Re: [Twisted-Python] Adding Mime types using coil On Fri, Jul 12, 2002 at 11:19:31AM -0500, Lewin, Karl E wrote: > Ok- > > Thanks for the quick reply, I need to muck around looking at the source > more so I can ask better informed stupid questions. That was a bug in Twisted, not a stupid question. Please feel free to ask us about more of those :) -Andrew. _______________________________________________ Twisted-Python mailing list Twisted-Python(a)twistedmatrix.com http://twistedmatrix.com/cgi-bin/mailman/listinfo/twisted-python

2 1

RE: [Twisted-Python] Adding Mime types using coil
by Lewin, Karl E 12 Jul '02

12 Jul '02

Ok- Thanks for the quick reply, I need to muck around looking at the source more so I can ask better informed stupid questions. Karl -----Original Message----- From: Itamar Shtull-Trauring [mailto:twisted@itamarst.org] Sent: Friday, July 12, 2002 10:53 AM To: twisted-python(a)twistedmatrix.com Subject: Re: [Twisted-Python] Adding Mime types using coil Lewin, Karl E wrote: > I am trying to add a mime type (image/svg+xml) to a default twisted web > server (mktap web --path c:/proj/html) and although it looks like it is > changed when I add it using coil, when I close and restart coil it > doesn't show up. Any suggestions on things to look for? This is a bug in Twisted, since the dictionary is actually a class level attribute instead of an instance, so it doesn't get saved when you save the tap. I'll fix this after we release 0.19.0. _______________________________________________ Twisted-Python mailing list Twisted-Python(a)twistedmatrix.com http://twistedmatrix.com/cgi-bin/mailman/listinfo/twisted-python

2 1

[Twisted-Python] Adding Mime types using coil
by Lewin, Karl E 12 Jul '02

12 Jul '02

I am trying to add a mime type (image/svg+xml) to a default twisted web server (mktap web --path c:/proj/html) and although it looks like it is changed when I add it using coil, when I close and restart coil it doesn't show up. Any suggestions on things to look for? Platform: WinXP Pro, Python 2.2 (activestate), Twisted Twisted-0.19.0rc3 Thanks- Karl

2 1

[Twisted-Python] distributed method calls in twisted.sister
by Sean Riley 12 Jul '02

12 Jul '02

background: twisted.sister is a piece of clustering infrastructure to allow a single application or simulation to be distributed across multiple servers on multiple hosts. twisted.sister has a concept of a "parent" server which acts as a central locking facility for objects distributed on any number of "sister" servers in the network. security in distributed sister servers: example use case: player1 logged into on one sister server wishes to send a text message to player2 logged into another - unknown - sister server. I implemented a "callDistributed" on the sister server which looks like: def callDistributed(self, resourceType, resourceName, methodName, *args, **kw): this goes off to the parent which looks up the sister that has the specified resource locked, then passes the method call off to that sister to actually be invoked. all results of course being deferred so the result is passed all the way back to the originating caller. so player1 would call: self.service.callDistributed( 'avatar', 'player2', 'chatMsg', 'hi there') and player2 would get the chat message. It currently prepends "remote_" to the method name on the callee's end. This appears to be inconsistent with the common twisted notion of knowing "who" is calling a method for security reasons. Maybe there should be another type of remote method called "distributed_" which takes the identifying information of the calling distributed object as the first arguments: def distributed_chatMsg(self, resourceType, resourceName, message) where the resourecType and resourceName identify the distributed resource that invoked the distributed method call. this distributed_ method would only ever be invoked by remote objects managed in a twisted.sister network. I thought about adding the location of the distributed caller object also, but this information probably isnt required as it is known by the parent who can route any further communication. the location could also change if an object migrated between sisters which would invalidate any location information stored on a sister, so not sending it is a better option. thoughts?

3 3

[Twisted-Python] Re: [Twisted-commits] Now the rules actually work
by Bruce Mitchener 11 Jul '02

11 Jul '02

moshez CVS wrote: > Modified files: > Twisted/debian/rules 1.30 1.31 > + cp -ar doc/howto/listings/ \ > + debian/twisted-doc/usr/share/doc/twisted-doc/howto/ > + install -m 644 doc/specifications/*.html \ > + debian/twisted-doc/usr/share/doc/twisted-doc/specifications/ Should we be installing the generated .xhtml (or whatever extension) files rather than the source .html ones? - Bruce

1 0