| --- Log opened Thu Sep 20 00:00:08 2007 |
| 01:11 | | ReivZzz is now known as ReivClass |
| 02:43 | | gnolam [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has quit [Quit: Z?] |
| 03:00 | <@ToxicFrog> | Sweet bouncing jibbly fuck |
| 03:01 | <@McMartin> | ? |
| 03:01 | <@ToxicFrog> | Was Java's threading model created by inimical intelligences that watch us from beyond the veil of reality? |
| 03:01 | <@ToxicFrog> | Was it found scrawled in blood on the walls of Rl'yeh? |
| 03:01 | <@McMartin> | No. It predates anyone working out how to do it right, though. |
| 03:02 | <@McMartin> | In particular, it predates pthreads by about five revisions. |
| 03:02 | <@ToxicFrog> | Was it, perhaps, written in dire languages readable only by the mad upon ancient papyrus scrolls recovered from a cursed thread? |
| 03:02 | <@McMartin> | Get a version of the API that doesn't include all the deprecated methods, imo. |
| 03:02 | <@ToxicFrog> | No-one knows, because of its original discoverers, there are no survivors. |
| 03:02 | <@McMartin> | Oh, give me a break. |
| 03:02 | <@McMartin> | I write multithreaded code in Java rgularly with no ill effects. *twitch* |
| 03:02 | <@ToxicFrog> | Heh. |
| 03:03 | <@ToxicFrog> | Using 1.5 here, BTW. |
| 03:03 | <@ToxicFrog> | As far as I can tell, all it has some kind of bastard hybrid of a condition variable and a mutex. |
| 03:03 | <@ToxicFrog> | Presumably, you are expected to implement other structures on top of that if you need them. |
| 03:03 | <@McMartin> | You, uh, are aware of the "synchronized" keyword? |
| 03:03 | <@ToxicFrog> | Yes. |
| 03:03 | <@McMartin> | Yeah, so, that's a mutex. |
| 03:03 | <@McMartin> | A re-entrant one, even. |
| 03:03 | <@ToxicFrog> | Yeah, but inside it, you can use wait()/notify() |
| 03:04 | <@McMartin> | Yeah, but don't. |
| 03:04 | <@ToxicFrog> | Which is where the condition-variable-esque behaviour comes into play. |
| 03:04 | <@McMartin> | wait() and notify() are part of Object. |
| 03:04 | <@ToxicFrog> | I, err, kind of have to, as I need thread A to start thread B, then block until thread B signals readiness. |
| 03:04 | <@McMartin> | If you need Traditional Locks, 1.5 has them in a subpackage. |
| 03:04 | <@ToxicFrog> | Normally this is a message pass, but Java doesn't appear to have that. |
| 03:04 | <@McMartin> | Also note that condition variables only work when they are also mutex hybrids. |
| 03:05 | <@ToxicFrog> | So that makes it a barrier, whoops, doesn't have that either. |
| 03:05 | <@McMartin> | Barrier: |
| 03:05 | <@ToxicFrog> | So I have main do a synchronized(server) { server.wait(); }, and server do a synchronized(this) { notify(); } when it's ready. |
| 03:06 | <@McMartin> | That's ordinary condition variables. |
| 03:06 | <@McMartin> | wait() without a mutex barrier like that is a guaranteed race condition, and condition variables can only exist if they're woven into mutexes. |
| 03:06 | <@McMartin> | (See also SDL's implementation of them) |
| 03:07 | <@ToxicFrog> | Granted. |
| 03:07 | <@McMartin> | So, uh, 1.5 also has a subpackage of synchonization classes that I refuse to touch because I don't trust them. |
| 03:08 | <@ToxicFrog> | I'm sorely tempted to just write message passing, if I can figure out where to wedge it in. |
| 03:08 | <@McMartin> | java.util.concurrent and its subpackages. |
| 03:08 | <@ToxicFrog> | Possibly as an interface, then I can just have all of my Threads implement MessagePass. |
| 03:08 | <@McMartin> | Which includes BlockingQueue. |
| 03:09 | | * Vornicus does wonder why the java.util containers aren't threadsafe. |
| 03:09 | <@McMartin> | Vorn: Because they aren't atmoic. |
| 03:09 | <@McMartin> | atomic |
| 03:09 | <@McMartin> | Or rather, operations thereon aren't. |
| 03:09 | <@ToxicFrog> | ...except, fuck |
| 03:09 | <@ToxicFrog> | Interfaces Aren't Mixins. |
| 03:09 | <@McMartin> | Mixins are of the devil, anyway. |
| 03:10 | <@McMartin> | Give your objects message queues as fields or something so that you don't have name collisions. |
| 03:10 | <@ToxicFrog> | Hmm. I guess I can subclass Thread and...yes. That works. |
| 03:11 | <@ToxicFrog> | Since all message passes involve thread references anyways. |
| 03:11 | <@McMartin> | ... wait, how have you not been subclassing thread? |
| 03:11 | <@McMartin> | Just using Runnable everywhere? |
| 03:11 | <@ToxicFrog> | Well, I mean subclass Thread, and then subclass it further for my actual threads. |
| 03:12 | <@McMartin> | Right |
| 03:12 | <@ToxicFrog> | ...right. Lack of multiple returns for the lose. |
| 03:12 | <@McMartin> | I suspect java.util.concurrent.BlockingQueue will provide much entertainment as well, though I haven't used it myself. |
| 03:13 | <@McMartin> | Multiple returns are evil and non-object-oriented, so Java no can has. |
| 03:13 | <@McMartin> | Anonymously defined inner class with two fields, though, can has. |
| 03:13 | <@ToxicFrog> | Yeah. Or just Do It Right and create a Message class. |
| 03:14 | | * McMartin has a certain tendency to comically overuse anonymous classes. |
| 03:14 | <@ToxicFrog> | ISTR anonymous class definitions being hideously ugly. |
| 03:14 | <@ToxicFrog> | Otherwise I would be whipping out anonymous-functions-as-Runnables a lot more frequently. |
| 03:14 | <@McMartin> | Depends on context. |
| 03:14 | <@ToxicFrog> | (the lack of closures also makes them of lesser value to me) |
| 03:14 | <@McMartin> | When you're defining, say, iterator() on your custom collcetion it fits in nicely. |
| 03:15 | <@McMartin> | What, you haven't defined TFUtil.Applicable yet? |
| 03:15 | | * Vornicus dislikes Java specifically because of the lack of operator overloads. Which I still don't understand what the hell they were on about, but I guess that's because I've never seen them /really/ misused. |
| 03:15 | <@ToxicFrog> | I don't have a util package for Java yet, this is the first time I've used it in ages. |
| 03:15 | <@ToxicFrog> | MessagePassingThread will be the first entry in it. |
| 03:15 | <@McMartin> | x = new Runnable() { run() { System.out.println("Lolz0rz"); } }; x.run(); |
| 03:15 | <@McMartin> | If memory serves. |
| 03:16 | <@McMartin> | Vorn: iostream tops the list. |
| 03:16 | <@ToxicFrog> | Vornicus: C++'s iostream, and Java's String, both abuse them. |
| 03:17 | <@McMartin> | String offends me slightly less, but my main objection to String overloading it is that StringBuffer and StringBuilder don't, and they should and String shouldn't. |
| 03:17 | < Vornicus> | iostream is pretty bad. String, idunno. Most languages let you do that; I find that languages that use specific operators for string concatenation are rare. |
| 03:17 | <@McMartin> | String is bad because Swtring lets you do that and it's O(n^2) to do so. |
| 03:17 | < Vornicus> | ah, well |
| 03:17 | < Vornicus> | that would be an important point. |
| 03:17 | <@McMartin> | The O(n) version requires you use a different class that doesn't understand + or +=. |
| 03:18 | <@McMartin> | And which one it is by default changed from 1.4 to 1.5. |
| 03:19 | < Vornicus> | heh |
| 03:19 | <@McMartin> | Ooh, ConcurrentMap looks nice too. |
| 03:19 | <@McMartin> | Too bad my experiments require me to maintain 1.4 compliance. =( |
| 03:19 | <@McMartin> | One of the subsystems I'm using won't eat 1.5 code. |
| 03:19 | <@ToxicFrog> | Hmm. How do I get a reference to the host Thread? |
| 03:20 | <@McMartin> | It's in either System or Runtime. |
| 03:20 | <@McMartin> | If you mean "The thread running right now". |
| 03:20 | <@McMartin> | But there's a static method that will hand it to you. |
| 03:21 | <@ToxicFrog> | Neither System nor Runtime appears to have a method returning a Thread. |
| 03:21 | <@McMartin> | OK, hang on, checking the APIs. |
| 03:22 | <@McMartin> | I know it's there because PQL uses it. |
| 03:22 | <@ToxicFrog> | And yeah, that's what I'm after - Message::Message(Object) needs to know the current thread so that it can tell the destination thread where it came frmo. |
| 03:23 | <@McMartin> | Yeah. PQL can bind on "currently running thread" to match patterns of the form "This happens in one thread" |
| 03:23 | <@McMartin> | Thread.currentThread(). |
| 03:23 | <@McMartin> | It's static. |
| 03:24 | <@ToxicFrog> | Aha. |
| 03:24 | <@ToxicFrog> | Thankye. |
| 03:24 | | * McMartin has actually never used that one in actual Java code. |
| 03:24 | <@McMartin> | Only in raw JVM assembler. |
| 03:24 | <@McMartin> | (stop looking at me like that) |
| 03:25 | <@ToxicFrog> | It should, I think, lock on the target thread... |
| 03:29 | <@ToxicFrog> | ...hmm. |
| 03:29 | <@ToxicFrog> | Ok, sender needs to block until reciever signals readiness. |
| 03:29 | <@ToxicFrog> | Reciever then blocks until sender finishes send. |
| 03:29 | <@McMartin> | The Queue won't handle that? |
| 03:29 | <@ToxicFrog> | Then sender blocks until until reciever replies, at which point the whole thing happens in reverse. |
| 03:29 | <@ToxicFrog> | Oh, right. Forgot to look at that. |
| 03:33 | <@ToxicFrog> | ....hrm. |
| 03:33 | <@ToxicFrog> | I can ask it what its remaining capacity is, but I can't set it. |
| 03:34 | <@ToxicFrog> | Aha. But it's implemented by SynchronousQueue, which looks like what I'm looking for. |
| 03:37 | | Vornicus is now known as Darius |
| 03:40 | <@ToxicFrog> | ...hmm. Can I downcast in Java? |
| 03:41 | <@ToxicFrog> | Or, more generally: Thread.currenThread() returns a Thread. |
| 03:41 | <@ToxicFrog> | I need to call MessagePassingThread methods. |
| 03:47 | <@McMartin> | Right. |
| 03:47 | <@McMartin> | MessagePassingThread t = (MessagePassingThread)Thread.currentThread(); |
| 03:47 | <@McMartin> | If this is called from, say, the main thread, you should do an instanceof check first. |
| 03:47 | <@McMartin> | If the cast fails a ClassCastException is thrown. |
| 03:51 | <@ToxicFrog> | Hmm. I wonder if I can get main to automatically be a MessagePassingThread. |
| 03:52 | <@ToxicFrog> | Eeeehehehehehehe |
| 03:53 | <@ToxicFrog> | class A1 extends MessagePassingThread { |
| 03:54 | <@ToxicFrog> | public static void main(String[] argv) { MessagePassingThread main_thread = new A1(); main_thread.start(); main_thread.join(); return; } } |
| 03:56 | <@McMartin> | Works for me |
| 03:56 | <@McMartin> | That, of course, isn't really main, but close enough. |
| 03:57 | <@ToxicFrog> | Yes. |
| 03:57 | <@McMartin> | (Wouldn't work for UQM ;_;) |
| 03:57 | <@ToxicFrog> | (;.;) |
| 03:57 | <@McMartin> | (SDL is picky about non-main threads touching the framebuffer) |
| 04:02 | <@ToxicFrog> | ;.; |
| 04:02 | <@ToxicFrog> | Error communicating with child thread: java.lang.NullPointerException |
| 04:04 | <@ToxicFrog> | God forbid the stack trace give me line numbers. |
| 04:04 | <@ToxicFrog> | Oh, right, need -g |
| 04:06 | <@ToxicFrog> | ...right. Need to allocate sendq and replyq in the constructor. |
| 04:07 | <@ToxicFrog> | ... |
| 04:07 | <@ToxicFrog> | aaargh |
| 04:09 | <@ToxicFrog> | ... |
| 04:09 | <@ToxicFrog> | So, apparently |
| 04:09 | <@ToxicFrog> | If I have a method send(Message) |
| 04:10 | <@ToxicFrog> | And another method, send(Object) |
| 04:10 | <@ToxicFrog> | Calling thread.send(null) will resolve to send(Message) and die screaming. |
| 04:12 | <@McMartin> | Overloading like that is seriously asking for trouble. |
| 04:12 | <@McMartin> | Overloads are *statically* resolve and hinge on how various things happen to have been cast. |
| 04:12 | <@ToxicFrog> | Aah. |
| 04:13 | <@McMartin> | Java and C++ are not I7. |
| 04:13 | <@ToxicFrog> | Presumably, this means that foo(Object bar) { send(bar) } will resolve to send(Object) even if bar is a Message at runtime? |
| 04:13 | <@McMartin> | (Or ML.) |
| 04:13 | <@McMartin> | That is correct. |
| 04:13 | <@McMartin> | It's basically only ever safe to overload like that if send(Object) does a type-check and then either complains or forwards to send(Message). |
| 04:14 | <@ToxicFrog> | That's what send(Object) does. |
| 04:14 | <@McMartin> | Ah. |
| 04:14 | <@McMartin> | In that case, your problem is that the expression "null" is castable to anything. |
| 04:14 | <@ToxicFrog> | And it's still not safe, because send(null) calls send(Message), which tries to call SynchronousQueue.put(null) |
| 04:14 | <@ToxicFrog> | Which is a runtime error. |
| 04:14 | <@McMartin> | Right. |
| 04:15 | <@McMartin> | "Most specific statically resolvable type" is the overload rule, IIRC. |
| 04:15 | <@McMartin> | But you'd better nullcheck in send(Message), not send(Object). |
| 04:15 | <@ToxicFrog> | Or rewrite the API so that you always pass it Objects, and never precreated Messages. |
| 04:16 | <@McMartin> | Also possible, I suppose. |
| 04:16 | | * ToxicFrog deems this both more appropriate and easier to implement, does so |
| 04:19 | <@ToxicFrog> | Ok! |
| 04:19 | <@ToxicFrog> | We now have message-passing, in 50 lines or so. |
| 04:19 | <@McMartin> | Hooray! |
| 04:19 | <@McMartin> | Is this for work, or are you contending with Java for an unrelated reason? |
| 04:20 | <@ToxicFrog> | It's september, so this is for classes. |
| 04:20 | <@McMartin> | Ah, yes |
| 04:20 | | * ToxicFrog works in the summer - and almost never in Java, thank the numberless lords of chaos - and attends classes in the fall and winter. |
| 04:21 | <@McMartin> | I see that the rumored life-giving season of "Spring" does not exist in the northern wastelands. |
| 04:21 | <@ToxicFrog> | Well, spring is the first month or two of the summer semester. |
| 04:21 | | * McMartin is currently dusting off P65 again. |
| 04:21 | <@ToxicFrog> | But even if it's physically spring, it's logically summer. |
| 04:21 | <@ToxicFrog> | Hmm. I'm actually surprised by how easy this was. |
| 04:22 | <@McMartin> | Yes, but when I was at Berkeley, the two main-year semesters were Fall and Spring. |
| 04:22 | <@McMartin> | And that's the nature of most serious Java programs. |
| 04:22 | <@ToxicFrog> | Having a pre-existing SynchronousQueue and threading support and not needing to move Objects between VMs may be a contributing factor~ |
| 04:22 | <@McMartin> | Everything is *almost* there already; you just have to find it and tweak it to be just the way you actually need. |
| 04:22 | <@McMartin> | So the ability to rapidly search and assimilate Javadoc will speed up coding by orders of magnitude. |
| 04:23 | <@ToxicFrog> | Aah. Here we have three four-month semesters - fall, winter and summer. |
| 04:24 | <@ToxicFrog> | Ok. Now that my tools are done, I can start on the actual assignment. |
| 04:25 | | Chalcedon [~Chalcedon@Nightstar-2472.ue.woosh.co.nz] has quit [Quit: Gone] |
| 04:25 | <@ToxicFrog> | ...hmm. |
| 04:25 | <@ToxicFrog> | It would be really convenient if I could generalize send/recieve/reply to work across sockets. |
| 04:26 | <@ToxicFrog> | I feel a MessagePassingSocket class coming on. |
| 04:27 | | Chalcedon [~Chalcedon@Nightstar-2472.ue.woosh.co.nz] has joined #code |
| 04:27 | | mode/#code [+o Chalcedon] by ChanServ |
| 04:29 | <@ToxicFrog> | I think I just need to say that class Message implements Serializeable, yes? |
| 04:30 | <@McMartin> | Probably, assuming it doesn't have anything seriously freakish in it like a HashMap. |
| 04:30 | <@McMartin> | If a Message is basically just made of Strings, though, then yes, you're golden |
| 04:30 | <@ToxicFrog> | ...hrm. Upon reflection, it has a MessagePassingThread and an Object. |
| 04:31 | <@McMartin> | That could get unpleasant pretty quick. |
| 04:31 | <@ToxicFrog> | Yeaaaaah. |
| 04:31 | <@McMartin> | You'd need to implement Serializable and then override the methods that handle serialized forms. |
| 04:31 | <@McMartin> | But, uh, you're talking about RPC at that point. |
| 04:31 | <@ToxicFrog> | Which is what the course is about anyway~ |
| 04:31 | <@McMartin> | Heh heh heh |
| 04:32 | <@ToxicFrog> | Anyways...upon further consideration, a MessagePassingSocket message needs different fields anyways. |
| 04:32 | <@ToxicFrog> | In particular, sender is a socket, not a thread. |
| 04:32 | <@ToxicFrog> | And content needs to be serializeable. |
| 04:32 | <@McMartin> | You still might be able to get a top-level interface out of it, but it won't be a fake mixin, that's for sure. |
| 04:34 | <@ToxicFrog> | Yeah. I think what I need here is to Do It Right and write a Message with an interface for getting the payload out, and then an InterThreadMessage with Thread sender and Object payload, and a NetworkMessage with Socket sender, Serializeable payload and custom serialize/deserialize to handle setting up sender right when it comes off the wire. |
| 04:35 | <@McMartin> | First, do your homework~ |
| 04:35 | <@McMartin> | Then, do the stuff that will come in handy later. |
| 04:37 | <@ToxicFrog> | This will come in handy for the homework immediately, though. |
| 04:37 | <@McMartin> | Ah |
| 04:38 | <@ToxicFrog> | Since it involves at least two nodes, each one consisting of two threads and communicating with other nodes over TCP. |
| 04:38 | <@McMartin> | Aha, I see. |
| 04:38 | <@ToxicFrog> | (also: re: Mixins being evil and wrong: Serializeable is totally a mixin even if the docs call it an interface, thank god.) |
| 04:39 | <@McMartin> | (Well, OK. But serialization should be done in a non-OO way.) |
| 04:39 | <@McMartin> | (<3 Pickle & friends) |
| 04:39 | <@McMartin> | (And it's not really a mixin, either, it's a pair of methods that are always defined, and the interface is basically just a notice about whether or not those methods will work.) |
| 04:42 | <@ToxicFrog> | (which, from the programmer's point of view, is functionally indistinguishable unless you're doing deep VM wizardry) |
| 04:42 | <@ToxicFrog> | ...hmm. |
| 04:42 | <@McMartin> | (oh wate) |
| 04:42 | <@ToxicFrog> | There doesn't seem to be a way for the deserializer to figure out where it came from. |
| 04:42 | <@ToxicFrog> | (yeah, yeah) |
| 04:43 | <@McMartin> | (yeah. This is a "take advice from me with a grain of salt" marker, really, because I'm closer to the VM than the language.) |
| 04:43 | <@McMartin> | (Java Generics in particular are totally beyond me because they vanish at the VM level at present.) |
| 04:44 | <@ToxicFrog> | Hmm. Ok, I can't make the semantics the same without some wacky select()oid hackery. |
| 04:44 | <@ToxicFrog> | So I'll settle for MessagePassingSocket.recv(), MPS.send() and MPS.reply() |
| 04:44 | <@McMartin> | Ah well |
| 04:45 | <@McMartin> | You can solve it later with an extra layer of indirection. |
| 04:45 | <@ToxicFrog> | Indeed. |
| 04:46 | <@ToxicFrog> | (the issue here is that for true identicality, I need MessagePassingThread.recv() to block until it recieves a message from any MessagePassing(Thread|Socket), and for replying to that message to send thread-to-thread or over the network as appropriate) |
| 04:46 | <@McMartin> | (Hmm, yes.) |
| 04:47 | <@ToxicFrog> | (at which point, communication wise, there's no difference between a thread on this machine and a thread on another machine except that the latter requires you to explicitly ask for a connection before you can send to or recieve from it) |
| 04:48 | <@ToxicFrog> | (Maybe I should just implement an Erlang VM in Java...) |
| 04:48 | <@McMartin> | (Not Easier) |
| 04:53 | <@ToxicFrog> | (yeah. ;.;) |
| 04:53 | <@ToxicFrog> | Hmm. One would think that there's some kind of generic select()oid in Java that lets me block on both a queue and a set of sockets, and returns when any of them are hot. |
| 04:55 | <@ToxicFrog> | Hmm. There's Selector, but that requires SelectableChannels. |
| 04:55 | <@ToxicFrog> | Not only is SerializeableQueue not one, it's not an interface, so I can't even subclass SQ and add it. |
| 04:56 | <@ToxicFrog> | Maybe I should just implement the entire ITC mechanism with sockets under the hood. |
| 04:56 | <@ToxicFrog> | But this means serialization/deserialization overhead on every single goddamn message pass. |
| 05:01 | <@McMartin> | I don't recommend it |
| 05:02 | <@McMartin> | Though it is very enterprisey. |
| 05:02 | <@McMartin> | =P |
| 05:02 | <@ToxicFrog> | Feh. |
| 05:04 | <@ToxicFrog> | I could busywait, but ew. |
| 05:06 | <@ToxicFrog> | ... |
| 05:06 | <@ToxicFrog> | I could spawn a sister thread responsible for listening to sockets and putting stuff recieved from them in the queue. |
| 05:08 | <@McMartin> | home |
| 05:08 | <@ToxicFrog> | ta. |
| 05:21 | <@McMartin> | And back |
| 05:29 | <@ToxicFrog> | Keh. I can't even blame Java for this; I can't think of any language in which this is easy except the ones that have it intrinsically. |
| 05:30 | <@McMartin> | Synchronization Is Hard (tm). |
| 05:30 | <@ToxicFrog> | The hard bit is synchronizing on sockets and on other threads simultaneously. |
| 05:30 | <@ToxicFrog> | Without just going "all communication goes over sockets" |
| 05:52 | | Darius [~vorn@ServicesOp.Nightstar.Net] has quit [Ping Timeout] |
| 05:55 | | Darius [~vorn@64.252.41.ns-3279] has joined #code |
| 05:56 | | Darius is now known as NSGuest-1138 |
| 05:58 | | NSGuest-1138 is now known as Darius |
| 07:04 | | GeekSoldier|bed is now known as GeekSoldier|work |
| 07:29 | | Darius is now known as Vornicus |
| 07:59 | | Vornicus is now known as Vornicus-Latens |
| 08:07 | | You're now known as TheWatcher |
| 08:21 | | Thaqui [~Thaqui@Nightstar-26933.jetstream.xtra.co.nz] has quit [Ping Timeout] |
| 08:24 | | Thaqui [~Thaqui@Nightstar-26933.jetstream.xtra.co.nz] has joined #code |
| 08:24 | | mode/#code [+o Thaqui] by ChanServ |
| 10:34 | | GeekSoldier|work [~Rob@Nightstar-4842.pools.arcor-ip.net] has quit [Quit: I quit!] |
| 10:52 | | Chalcedon [~Chalcedon@Nightstar-2472.ue.woosh.co.nz] has quit [Quit: Gone] |
| 11:08 | | GeekSoldier|work [~Rob@Nightstar-4842.pools.arcor-ip.net] has joined #code |
| 11:12 | | gnolam [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has joined #Code |
| 11:12 | | mode/#code [+o gnolam] by ChanServ |
| 11:31 | | Attilla [~The.Attil@194.72.70.ns-11849] has quit [Quit: <Insert Humorous and/or serious exit message here>] |
| 11:39 | | Attilla [~The.Attil@194.72.70.ns-11849] has joined #code |
| 11:56 | < jerith> | Spot the bug: |
| 11:56 | < jerith> | public String getPrefix(String uri) { |
| 11:56 | < jerith> | if( uri.equals(XMLConstants.XML_NS_URI) ) |
| 11:56 | < jerith> | return XMLConstants.XML_NS_PREFIX; |
| 11:56 | < jerith> | if( uri.equals(XMLConstants.XMLNS_ATTRIBUTE_NS_URI) ) |
| 11:56 | < jerith> | return XMLConstants.XMLNS_ATTRIBUTE; |
| 11:56 | < jerith> | if( uri==null ) |
| 11:56 | < jerith> | throw new IllegalArgumentException(); |
| 11:56 | < jerith> | //... |
| 11:56 | < jerith> | } |
| 11:56 | < jerith> | I can't even fix it, because it's code generated by someone else's tool at compile-time. |
| 11:57 | < Vornicus-Latens> | cute |
| 11:58 | < jerith> | Note that I found it because Eclipse throws a warning. |
| 11:59 | < jerith> | I even fixed it, before I realised it was generated code. |
| 12:00 | < Vornicus-Latens> | Eclipse is awesome. |
| 12:01 | < jerith> | It is, but only for Java code. |
| 12:01 | < jerith> | Java /needs/ lots of static analysis and automated refactorings. |
| 13:07 | | Kyrre [~Z@87.72.35.ns-3885] has quit [Ping Timeout] |
| 13:09 | | Kyrre [~Z@87.72.35.ns-3885] has joined #Code |
| 13:10 | | KarmaBot [~fark.off@87.72.35.ns-3885] has quit [Ping Timeout] |
| 13:10 | | KarmaBot [~fark.off@87.72.35.ns-3885] has joined #Code |
| 13:10 | | mode/#code [+v KarmaBot] by ChanServ |
| 13:11 | | McMartin [~mcmartin@Nightstar-904.dsl.pltn13.sbcglobal.net] has quit [Ping Timeout] |
| 13:13 | | McMartin [~mcmartin@Nightstar-904.dsl.pltn13.sbcglobal.net] has joined #code |
| 13:13 | | mode/#code [+o McMartin] by ChanServ |
| 13:22 | < GeekSoldier|work> | i just started using eclipse for python. |
| 13:22 | | GeekSoldier|work is now known as GeekSoldier |
| 14:05 | < jerith> | GeekSoldier: How's it working for you? |
| 14:06 | < jerith> | Two guys here use it for Ruby, but I prefer a normal editor |
| 14:09 | < GeekSoldier> | not too bad. I'm still getting used to it. |
| 14:10 | < GeekSoldier> | I like being able to have all my project files open in the same editor. I had been using IDLE for it before. |
| 14:10 | < jerith> | Ah. |
| 14:11 | | * jerith hands GeekSoldier {emacs,vim} and ipython, which is the best Python dev system he has yet discovered. |
| 14:17 | < GeekSoldier> | golly... I don't think I've enough fingers for emacs. |
| 14:25 | < jerith> | Heh. It's not that bad once you get used to it. |
| 14:26 | < jerith> | And it's a price worth paying, IMHO. |
| 14:26 | < jerith> | But vim's also a good Python editor. |
| 14:35 | | AnnoDomini [AnnoDomini@Nightstar-29100.neoplus.adsl.tpnet.pl] has joined #Code |
| 14:35 | | mode/#code [+o AnnoDomini] by ChanServ |
| 14:37 | < MinceR> | hay |
| 14:37 | | ReivClass is now known as ReivZzz |
| 14:39 | <@gnolam> | http://www.youtube.com/watch?v=Mal6XbN5cEg <- "QBasic Nerd" |
| 15:00 | | ReivZzz is now known as ReivSLEP |
| 15:10 | | Thaleon [~Matt_@70.91.160.ns-4029] has joined #Code |
| 15:11 | | Thaleon [~Matt_@70.91.160.ns-4029] has quit [Quit: ] |
| 15:26 | | AnnoDomini is now known as Jesus |
| 15:26 | | Jesus is now known as AnnoDomini |
| 15:43 | | Thaqui [~Thaqui@Nightstar-26933.jetstream.xtra.co.nz] has quit [Ping Timeout] |
| 16:28 | < jerith> | So, Java. |
| 16:29 | < jerith> | I have some code that uses .toGMTString(), which is deprecated. |
| 16:29 | < jerith> | Is there a drop-in replacement? |
| 16:29 | < jerith> | Google is being unhelpful in this regard. |
| 16:31 | <@AnnoDomini> | The Java documentation tells me it's been replaced with DateFormat.format(Date date). |
| 16:33 | < jerith> | But I'm not sure how to get a date in the right timezone. |
| 16:33 | < jerith> | I should probably RTFM, but it's Java and I'm tired. :-/ |
| 16:41 | < jerith> | Bloody hell, now I know where Ruby gets it from! |
| 16:41 | < jerith> | 'The format is locale independent and digits must be taken from the Basic Latin block of the Unicode standard. No daylight saving time transition schedule can be specified with a custom time zone ID. If the specified string doesn't match the syntax, "GMT" is used.' |
| 16:42 | < MinceR> | yay for proper error handling. |
| 16:43 | < jerith> | Because throwing a WrongTimezoneFormatYouBlitheringIdiotException is /clearly/ not the right thing to do there... |
| 16:43 | < MinceR> | clearly. |
| 16:43 | < jerith> | Although you can actually tell that you have an error there if you look and you have GMT. |
| 16:44 | < jerith> | Well if it's a user-specified string, you can't. |
| 16:49 | < jerith> | Although it's envvar based, so proper error handling is rather more difficult than throwing a WrongTimezoneFormatYouBlitheringIdiotException. |
| 16:50 | < jerith> | Oops, mischan. |
| 16:51 | | gnolam is now known as gnoperkele |
| 16:58 | | AnnoDomini is now known as AnnoDomini2|Dalek2 |
| 17:01 | | AnnoDomini2|Dalek2 is now known as AnnoDomini |
| 17:21 | | You're now known as TheWatcher[afk] |
| 17:57 | | Thaqui [~Thaqui@Nightstar-26933.jetstream.xtra.co.nz] has joined #code |
| 17:57 | | mode/#code [+o Thaqui] by ChanServ |
| 18:28 | | Vornotron [~vorn@64.252.44.ns-4484] has joined #code |
| 18:29 | | Vornicus-Latens [~vorn@ServicesOp.Nightstar.Net] has quit [Ping Timeout] |
| 18:33 | | You're now known as TheWatcher |
| 18:38 | | Vornotron is now known as Vornicus |
| 18:45 | | GeekSoldier [~Rob@Nightstar-4842.pools.arcor-ip.net] has quit [Ping Timeout] |
| 18:45 | | GeekSoldier [~Rob@Nightstar-5400.pools.arcor-ip.net] has joined #code |
| 19:25 | | Forj [~Forj@Nightstar-2472.ue.woosh.co.nz] has joined #code |
| 19:25 | | mode/#code [+o Forj] by ChanServ |
| 19:32 | | gnoperkele is now known as gnolam |
| 19:35 | < MinceR> | SSLServerSocketFactory sslserversocketfactory = (SSLServerSocketFactory)SSLServerSocketFactory.getDefault(); |
| 19:35 | < MinceR> | gotta love java. |
| 19:39 | < jerith> | No, you don't. |
| 19:40 | < jerith> | You can use Python instead. |
| 19:40 | < jerith> | Speaking of which... |
| 19:40 | | * jerith goes to look at his Python code. |
| 19:41 | < MinceR> | i wish i could use python instead on tomorrow's class |
| 19:41 | < GeekSoldier> | yeah. I'm trying to do something, and it's gettin me steamed... convert an integer to a list... my grey matter is not working tonight. I shall prevail, though. |
| 19:41 | < jerith> | Umm... integer to list? |
| 19:41 | < jerith> | Do you want to put it into a list of one element? |
| 19:41 | < MinceR> | but just java isn't enough for them, they've got to use java 1.4 |
| 19:41 | < jerith> | foo = [bar] |
| 19:42 | < GeekSoldier> | for example: |
| 19:42 | < jerith> | MinceR: Ugly loops. :-/ |
| 19:42 | < GeekSoldier> | 1234 -> [1,2,3,4] |
| 19:42 | < MinceR> | and ugly containers |
| 19:42 | < jerith> | Ah. |
| 19:42 | < GeekSoldier> | it will involve a loop and mod. |
| 19:42 | < GeekSoldier> | I'm getting it now. |
| 19:43 | < jerith> | Easier to go via a string, actually. |
| 19:43 | | gnolam is now known as gnobitter |
| 19:43 | < GeekSoldier> | list[] = str(1234)? |
| 19:44 | < jerith> | [int(c) for c in str(a)] |
| 19:44 | < GeekSoldier> | I always forget the elegance of list comprehensions. |
| 19:44 | < jerith> | If you're looping over it, you may want to replace the square brackets with round and get a generator comprehension. |
| 19:52 | | Forj [~Forj@Nightstar-2472.ue.woosh.co.nz] has quit [Quit: Gone] |
| 20:05 | | GeekSoldier is now known as GeekSoldier|bed |
| 20:06 | < jerith> | 'Night GS. |
| 20:07 | < GeekSoldier|bed> | goodnight. |
| 20:07 | < GeekSoldier|bed> | I solved my thingy. thanks for the help! |
| 20:07 | < jerith> | :-) |
| 20:07 | < jerith> | Yay for solved thingies! |
| 20:41 | <@McMartin> | MinceR: 1.4? That is unfortunate, though I don't quite trust 1.6 yet. |
| 20:41 | <@McMartin> | After awhile you get used to the boilerplate. |
| 20:42 | <@McMartin> | And you stop using long varnames since the typecasts are doing that for you. =P |
| 20:43 | < MinceR> | i don't get used to it since i avoid it most of the time. :> |
| 20:43 | | * McMartin would recommend renaming sslSocketFactory something like "f". |
| 20:53 | < jerith> | Ugh. Why can't Python make everything return something? |
| 20:53 | <@McMartin> | It does. |
| 20:53 | <@McMartin> | "None" is a value. |
| 20:54 | < jerith> | " ".join(a.split().remove('baz')) |
| 20:54 | < jerith> | Now I have to use three lines and a temporary variable. |
| 20:54 | < jerith> | Because .remove() returns None instead of the array. |
| 20:54 | < jerith> | s/array/list/ |
| 20:55 | <@McMartin> | Um |
| 20:55 | <@McMartin> | You don't want remove there |
| 20:55 | <@McMartin> | You want a filter |
| 20:55 | <@McMartin> | remove is destructive |
| 20:55 | <@McMartin> | " ".join([x for x in a.split if x != 'baz']) |
| 20:55 | < jerith> | I don't care, it's on a temporary array. |
| 20:55 | < jerith> | s/array/list/ |
| 20:55 | < jerith> | Again. |
| 20:57 | < jerith> | Much as I like list comprehensions, that's not really a clean way to do it. |
| 20:57 | | * jerith does it that way anyway. |
| 21:32 | | Chalcedon [~Chalcedon@Nightstar-2472.ue.woosh.co.nz] has joined #code |
| 21:33 | | mode/#code [+o Chalcedon] by ChanServ |
| 22:48 | | KarmaBot [~fark.off@87.72.35.ns-3885] has quit [Ping Timeout] |
| 22:49 | | Kyrre [~Z@87.72.35.ns-3885] has quit [Ping Timeout] |
| 23:02 | | AnnoDomini [AnnoDomini@Nightstar-29100.neoplus.adsl.tpnet.pl] has quit [Ping Timeout] |
| 23:03 | <@McMartin> | I agree it's kind of a wart in the API, after some pondering of this, but I think on the whole I like it |
| 23:03 | <@McMartin> | Because it means I won't accidentally mutate something I was intending to hold constant, because the imperative routines all return None and can't be chained the way functionals can. |
| 23:04 | | AnnoDomini [AnnoDomini@Nightstar-29029.neoplus.adsl.tpnet.pl] has joined #Code |
| 23:04 | | mode/#code [+o AnnoDomini] by ChanServ |
| 23:07 | | Attilla [~The.Attil@194.72.70.ns-11849] has quit [Quit: <Insert Humorous and/or serious exit message here>] |
| 23:08 | < MinceR> | afaik it's because they didn't want confusion regarding whether the value is changed or the result of the change is returned. |
| 23:17 | | You're now known as TheWatcher[T-2] |
| 23:20 | | You're now known as TheWatcher[zZzZ] |
| 23:31 | | * McMartin makes Notepad++ understand P65, wonders how to export this. |
| 23:34 | | Attilla [~The.Attil@194.72.70.ns-11849] has joined #code |
| 23:41 | | * MinceR makes Notepad++ understand P90. |
| 23:46 | | gnobitter is now known as gnolam |
| 23:55 | < Attilla> | By shooting it? |
| 23:55 | < MinceR> | that's the only way :> |
| 23:56 | <@ToxicFrog> | McMartin: find the rc files, publish the relevant lines? |
| 23:57 | | * ToxicFrog reimplements Message as an abstract class |
| 23:59 | <@ToxicFrog> | Hmm. |
| --- Log closed Fri Sep 21 00:00:01 2007 |