how do you know it's not a guarantee? (the atomic i++, I mean)

gotta go

hoompf... too late, it seems

oh, cybereal already left

try catching your error and calling getCause().printStackTrace() on it
or do you get all that anyway from the outermost Throwable anyway?

it's probably in the JLS
~JLS

http://java.sun.com/docs/books/jls/

listenerList.add(SelectionChangedEventListener.class, listener); Bound mismatch: The generic method add(ClassT, T) of type EventListenerList is not applicable for the arguments (ClassJSvgCanvasEnh.SelectionChangedEventListener, JSvgCanvasEnh.SelectionChangedEventListener)...

I couldn't find it in the JLS, but I didn't look very hard.

As far as I can see, ClassJSvgCanvasEnh.SelectionChangedEventListener, JSvgCanvasEnh.SelectionChangedEventListener) correspondes very well to the prototype: (ClassT, T)

B is instantized from a method within A. Is it possible to get a reference to the A object within method x ?

(so.. like reference to "parent class")

hmmm (I didn't look at all :-) maybe it's in some kind of VM specification
A.this
and not "parent class", more "outer class"

how would i be able to use a thread in a servlet?

bigmojo, like usually?

wont i get a exception when the servlet host is context switchet

uhm
Thread object is like any other java host object and i don't see how calling start() would cause any problems

http://java.sun.com/docs/books/jls/third_edition/html/memory.html#17.7 "Non-atomic Treatment of double and long "

well... I couldn't find it in JLS and I did look hard (perhaps not hard enough)

Yes, I found that bit. I meant ++ not being atomic on int.

yeah, I was just about to say that
in the postfix/prefix ++/-- operator description, there is nothing about it being atomic or not...

i have a no-GUI application and want to run some code on exit, how can this be done?

http://www.google.it/search?q=java+shutdown+hook&ie=utf-8&oe=utf-8&aq=t&rls=org.mozilla:en-US:official&client=firefox-a

~tell paulweb515 about javadoc Runtime.addShutdownHook(*)

paulweb515, please see java.lang.Runtime.addShutdownHook(java.lang.Thread): http://java.sun.com/javase/6/docs/api/java/lang/Runtime.html#addShutdownHook(java.lang.Thread)

Oops
~tell paztulio about javadoc Runtime.addShutdownHook(*)

paztulio, please see java.lang.Runtime.addShutdownHook(java.lang.Thread): http://java.sun.com/javase/6/docs/api/java/lang/Runtime.html#addShutdownHook(java.lang.Thread)

Please ignore the above.

ok, thanks to all

~kancho DRMacIver
Please ignore the above. :-)

hi , can anyone tell me how i can load a new jar host into the classpath after the app has been started ?

Ok, done.

if you are fooling around with jars, you can write your own classpath host management using URLClassLoader, you can hack the system class loader, or you can use one of the frameworks that deals with plugins (JPF or OSGi)

i'm trying to load jars into tomcat6 hosting webapps after they have been deployed

that ... you're dealing with j2ee container classloaders ... unless tomcat has a section on doing it, I think it might be more trouble than it's worth

it has some implementations , but so far i'm in the dark , looks like everything is protected in the WebAppClassloader
but can't i add classes in a standard way to any application with say System Class Loader ?

I wonder if the Java 7 language improvements will do anything towards removing some of the JEE nonsense.

spring removed a lot of the jee nonsense already

No it didn't.
It just made a more palatable version of it.
(Admittedly a much more palatable version)
(In particular what I was thinking of is that the module system may well end up removing a lot of the class loader hell one experiences)

do you have experience loading jar's into the classpath of a servlet / ee container without redeploying the app ?

Nope.
At least not beyond hot redeploy capabilities provided by the app server

ok
I checked... i++ on int isn't atomic

http://rafb.net/p/1jtKNP62.html
can someone please tell me why this error is occurring ?
both the files are in the same package

okay

and if i remove the numbers in brackets, it gives error about uninitialized variables

rohan Complex c1(1,2);
Complex c2(2,3);
er
you can't do that

how did you check that?

Complex c1 = new Complex(1,2);

oh damn .. i guess C++ is hard to shrug off :-s

it's okay

and the reason is ? because memory is never allocated at compile time ?

I wrote a program, which starts T threads incrementing the same variable in a loop (N times). After the thread finishes, I print the current value of the variable. After the last one finishes, the result should be T * N. It isn't

just because. It's how the language is designed
Complex c1; // legal

if two classes are in the same package, do they still need the import statement ?

gotta go.
nope

no

and I tested it on a single processor, single core machine

allright, cool

ok, then after correcting the above mistake i am getting undefined symbol "Complex"

compile Complex first

don't use "java" as a package :-)

already done

hi

ouch, sorry

as an aside, you should also put "public Complex(int x, int y)"
and you changed them to Complex c1 = new Complex(1,2)?

i try to replicate a complex application state across two machines (using serialization). the problem is that i get some ClassNotFoundException in ObjectInputStream.resolveProxyClass, how can i solve this?

i made that change

(seems to be a problem with proxy deserialisation)

same error .. i changed the package to javatest

you know that your classpath has to include the root of your package hierarchy?

oh... i didn't know that :-s

it's worth reading up on classpaths, but if you have (now) source/javatest/Complex.java (etc) then one way to compile/run would be:
cd source; javac host -classpath . javatest/*.java ; java -classpath . javatest.Ctest

wow paulweb515 , thanks a lot
i was using CLASSPATH=.. java Ctest.java

i hate to ask this, but is there any way to launch web document from a standalone java app?
i know it's possible within an applet
oh, i mean, is there any standard and portable way (as opposed to Runtime.getRuntime().exec(...))

you can use CLASSPATH with some hand waving, but I find it more predictable to use the command line arguments

I think there's some standardish stuff out there which basically checks the OS and does an appropriate exec for each available option.
I think I saw one the other day. One second.

yeah, found some GPL library for doing exactly that
and it's pretty new, so i guess there's no way to do it in standard J2SE
well no problem, i don't really need it, it just would have been nice ^_^

Yeah, no way of oding it in J2SE that I know of.

public void tableChanged(TableModelEvent tableModelEvent)

http://sourceforge.net/projects/browserlaunch2/ is LGPL, which is more tolerable.

is this called when a row is removed?

you need to read up a bit more on threads

http://www.koders.com/java/fid078E17954F288533970E38C98D5500BDC996878F.aspx is a random class for doing it under a pretty liberal license. A bit old, but probably still works.
Hmm
A lot old. I rescind the "probably still works."

java doesn't support operator overloading, right ?

well, thanks in any case ^_~ perhaps we'll use it in the end, dunno

thankfully no

lol, why "thankfully" ?

because it's a trainwreck

i guess that Java is not for such things .. it's for GUI or atleast web based UI .. so operator overloading would really have no use

it's always fun to be debugging someone's code only to finally realize that they thought + was a great way to so some method that had nothing to do with what normal folks would even consider

yeah. operator overloading is overvalued.

it's like deciding to name your methods really odd names cause it's 'cool'

lol

it's unneeded and just causes code maintainability issues

But it has to be proper use, and it's very hard to do that within the constraints of operator overloading.
But it has to be proper use, and it's very hard to do that within the constraints of operator overloading.
The best example I can think of which fits inside operator overloading is overloading || for coalesce.
i.e. foo || bar = (foo == null) ? foo : bar

you do a lot of sql don't you?
8^)=

|| is string concatentation in SQL.
This comes from javascript.

ah!
8^)=

Or do you mean because I called it 'coalesce'?
(As it happens, I do do a fair bit of SQL. Less than I'd like though.)

yeah. coalesce gave you away.

yeah, because me being new to your code it's immediately obvious that || means coalesce
so much more readable when you have to go look up the definition
we should do that for all method names!
just have them short a1, a2 and a lookup table for what they do

Oh noes. It took you 30 seconds to look up a method.

I don't buy it
I've dealt with the bad side of it way too much

If you don't save 30 seconds from the improved readability once you know what this meant, you were spending so little time with my code that I don't give a shit whether you can read it.

thankfully you have no say in adding such silliness to the language! :P
smarter minds have prevailed

so far...
8^)=

Anyway, || is perfectly readable in this context, seeing as || means or.

yeah. that *is* clear.
what happens when you say a || b when a and b are Booleans?
are you or'ing them? or coalescing them?
readability++

As it happens, I have a good answer for that. It just doesn't work in Java.

how do you ensure 'smart' usage of overloading?

well, you're proposing that this happens. how do you propose to make that work?

Sensible type signatures. (?Boolean) || (Boolean) is coalesce. (Boolean) || (Boolean) is or.

yay! complicating the spec++

Nope.
Follows from other stuff I already want.

well, we all know you're a nutter anyway. P^)=

True
But just because I'm crazy doesn't mean I'm wrong.

but can you be trusted? that's the question...
and anyone named after a TV hero probably can't be...

waz-: To answer your earlier question, I don't. Pretending that you can stop people from writing stupid code is a waste of time.
waz-: In the same way that I don't enforce at the language level that people don't name their methods foo, bar and baz.

but you *can* restrict usage of things like MI that are known problems.

hey DRMacIver didn't you say 4.2 was released?
http://www.jboss.com/downloads/index -- that's still showing 4.2 beta

waz-: It is my tentative experience with readability that if you use a type class and user definable operator overloading mechanism a lot of the problems go away.
waz-: So you can't arbitrarily overload +, but anything you declare to be a number can have + defined. Ditto and orderability and any other operators you like to define.

that was me
4.2.0.GA is the lastest
if the page is wrong I'll bug someone

dmlloyd, ah - but...
unless i'm missing the 'naming' problem... is this what they're calling the Enterprise Application Platform ?
the home page says 4.2-beta also. but they're throwing around 'middleware' a lot.

http://labs.jboss.com/jbossas/downloads

i'm working out. trying to lose my middleware.

~lart cheeser

tell eidolon about sensayuma

eidolon, ain't you got no sensayuma?

they may be referring to "application platform" which is sort of a different thing from "application server"

bah.
stooopid marketing weirdos.

http://www.jboss.com/products/platforms/application/faqs

ohhhh.
interesting.
blah
i hate when they draw lines like this.
particularly when there is so much overlap

Ok. I concede the point that || isn't perfect here. The point is that a coalesce operator is really handy, and if all you have is operator overloading rather than fullblown user definable operators, || is a relatively harmless choice.

hi guys, does anyone know if theres a way to get iSight cameras working in java?
using JMF i havn't had much luck, it doesnt seem to pick it up as a device

forgiving the ignorance, but what does the coalesce operator actually do? (and don't say coalesce

it is.

a != null ? a : b;

Which is what?

coalesce is handy

Ah, right.

so. i don't get this. is the JBoss Enterprise Application Platform free?

but it's a simple enough method to write, too.

or is only the Application Server free?

yes

what's sun.misc.Unsafe

a class you shouldn't care about

ah, the subscription is for support i guess? it's one of the FAQ's

Sure. I've written a version of it for my own usage. But I find foo || bar || baz (or equivalent) to be more readable than coalesce(foo, bar, baz).
bSure. I've written a version of it for my own usage. But I find foo || bar || baz (or equivalent) to be more readable than coalesce(foo, bar, baz)./b

no one knows about isights and java here i take it

i'll try just the appserver - see if that has enough JPA pieces in it (which is the only real plugin we use)

Infix notation is often a big win for readability. If it weren't we'd all just write lisp.

SinJax, if no one answered, probably not.

alas

I'm not saying "Doom and disaster! No one can read coalesce(foo, bar, baz)". Just that it would be nicer infixed.

yay Doom!

Hmm.
I have a sneaking suspicion that something is seroiusly broken in JBoss 4.2

don't say that DRMacIver
i'm downloading 4.2 now.

it's okay

hoompf... it seems that AtomicInteger is over 2x faster than Integer
(at least when I compare i++ of Integer with i.incrementAndGet() of AtomicInteger
)

i++ does autoboxing?

That's not terribly surprising.

that's why I wanted to see what's inside sun.misc.Unsafe
it does
yeah, I wouldn't say a thing if the reverse was true...

i++ on an Integer is Integer.valueOf(i.intValue() + 1), so it has to do a boxing operation which AtomicInteger doesn't.

hoompf..
hmm... but still.. I thought that using Integers does not introduce so much of the overhead

I try to run a java app by using command "java venturi.class" but it complaints: "java.lang.ClassNotFoundException: venturi.class not found in ..." what do I do wrong?

if i got a function that takes a "? extends Number" (generic) parameter, shouldnt i be able to put a Number in there then? or do i have to put an Integer/Long in? (Number doesnt EXTEND number..)

since they did something special with Integer (so that you can write i++ --- kind of operator overloading, but not quite), they could have done it without boxing... I mean, it could be just incrementation of the underlaying int value.

Heyas. i have a problem with jdbc where i don´t know why it happens. Calling the function createDB() and getting an SQLException but when i copy the sql command directly into phpmyadmin it works without problems. Any ideas? http://rafb.net/p/dlTOYT67.html (Code of the function/the sql function i am importing and the error)

(and not creating a new Integer object)

It depends what you mean by 'not much'. But Integers are immutable, so i++ needs to get a new Integer, not just mutate an int. valueOf is cached for small values of i, but for large it has to create a new Integer.

yeah, just thought of that

But they're both pretty damn fast. It's rare that this will be a bottleneck unless you're doing really scary high performance things.

naah... I just wrote this test to see if ++ is atomic
and I'm testing how different types behave

Ok I am new to Java and have what is probably a simple question. I am writing an applet and I need to fill a JTable from a database can I pass the JTable into the funtion and return the JTable?

(I've been discovering recently that although allocation is extremely fast, it's not too hard to beat it)

i++ is atomic on all primitive types other than double and long.

well.. you are wrong there

JLS must be wrong too.

I'm sorry, but if your programs depend on this "fact", you have some sleeples nights ahead of you

hi.. somebody can help me... I´d like to know if .. is possible cripto the password intro jboss datasource-ds.xml...

show me the place in JLS where it sais that it's atomic, will ya?

mebbe leading ';', mebbe table is already there, mebbe it doesnt like the quoted table name, why is the error on line 1?

It's either JLS or JVM spec. Hold on.

well.. I have sun vm and it's not atomic here

x = ..; x.add(new Integer(1), new Integer(2));.. shouldnt that work?

I couldn't find it in JLS

the database where i am importing it is empty no tables in there
trying it without quoted table names now

does anyone know if there is a function for servlets that will get the directory of the base of the application?

http://madbean.com/2003/mb2003-44 Oh - yeah. You're right.
ishi++

ishi has a karma level of 1, pr3d4t0r

hmm... what's karma?

Thread safety is totally overrated anyway.

it helps you get good parking spaces

As I think I mentioned the other day, I found a spectacularily unthreadsafe piece of code in one of our crucial features the other day. It's been there for months, and it's the only part of the code that hasn't fallen over.
(The code in that specific feature)

java -classpath . venturi
no suitable method `main' in class
-- what do I do?

show the stack trace in pastebin, the error on line 1 doesnt make sense

~tell ishi about wiki

http://www.javachannel.net

See the javabot FAQ

same without quots
http://rafb.net/p/63slix97.html - stacktrace

karma is basically a measure of your respect in the channel

well.. in the particular case of counting something, using i++ might get you results even 50% wrong (for 20 threads single cpu single core)

well the respect we have for you

I should probably clarify. I am in no way, shape or form advocating thread unsafe code.

well.. not true, really... when counting "nothing" it can... counting something which takes some time will probably work ok

It was a joke, partly made out of bitterness inspired by the fact that I'm having insane locking issues at the moment.

Anyone know how to get a venturi.class (from: http://www.ce.utexas.edu/prof/KINNAS/319LAB/Applets/Venturi/venturi.zip) running standalone in java command line? tried java venturi but it says: "no suitable method `main' in class"

hello!

http://i0005.photobucket.com/albums/0005/icanhascheezburger/2007/6/12/1/128261114448593750lolcatalignr.jpg

I have problem with conversion from Boolean to boolean

whadefu?

~tell PawciooS about javadoc Boolean.booleanValue()

PawciooS, please see java.lang.Boolean.booleanValue(): http://java.sun.com/javase/6/docs/api/java/lang/Boolean.html#booleanValue()

boolean bResult = Boolean.valueOf((String) o);

no `server' JVM at `CProgram Files\Java\jre1.5.0_11\bin\server\jvm.dll'.

but it doesn't work
why?
where o = Object

Are you using Java 1.4?

no, compiler is GWTCompiler

Ok. So you're using Java 1.4
(Or some subset thereof)
GWT doesn't support 1.5 syntax yet does it?

GWTCompiler use java 1.4 ? I didn't know

note the path, DRMacIver
oh, not me.

is there any solution for converting these data ?

See the javadoc I gave you a while ago.

see above what I wrote

javadoc.

ok,

Yes. What you wrote attempts to use autounboxing.
1.4 doesn't support this.
(You should be using parseBoolean anyway)

parseBoolean ?

Boolean.valueOf(String s)

ok

That returns a Boolean. parseBoolean returns a boolean.

oh
well I'd say Boolean.booleanValue is probably what he wants

This is a standard naming convention on all the primitive classes.
Not really, as he's parsing a string and trying to get a boolean.

not completely standard

I'm pretty sure they all have it.

there were some exceptions, that I had memorized at one point for scjp, but I've long since forgotten

ok
All the ones I've cared about have it anyway.

any other ideas?

boolean parseBoolean(String s)
actually parseBoolean is since 1.5
that's why i didn't know about it I guess
I gotta update my cert

wrr

hello guys, is there a way to get a String out of a ByteBuffer?

sure.
there's a couple ways.

Does toString not do the right thing?
(Genuine question. I don't know. It does for charBuffer)

no

dmlloyd, toString() method return the status, do I need to pass it to a CharBuffer and then user CharBuffer's toString method?

ok

no way to specify encoding

Ah

you can wrap it in a charbuffer, or you can get the bytes and use the string byte[] constructor
are you trying to get the whole buffer as a complete string? it won't be a partial buffer or anything like that?

dmlloyd, I think i like more the second
dmlloyd, yes, the complete buffer, since then I need to parse it, and String methods are nice for what i need

Surely the second is potentially very inefficient?

it's inefficient either way
you have to copy
well, probably one extra copy for getting a byte array

so it's better to use the charbuffer?

1,093 INFO [Server] JBoss (MX MicroKernel) [4.2.0.GA (build: SVNTag=JBoss_4_2_0_GA date=200705111440)] Started in

boolean bResult = Boolean.parseBoolean(sBool); but it return me: The method parseBoolean (String) is undefined for the type Boolean.

BEWARE, I LIVE!

~tell P4C0 about javadoc CharsetDecoder

P4C0, please see java.nio.charset.CharsetDecoder: http://java.sun.com/javase/6/docs/api/java/nio/charset/CharsetDecoder.html

that's the "right" way

That's what I meant (about the extra copy)

I must first declare Boolean as null next insert method parseBoolean ?

Actually, more serious problem. You can't always get an array out of a buffer.

the CharsetDecoder can also handle partial characters
indeed

dmlloyd, thanks

you don't want to use "asCharBuffer" because it just converts pairs of bytes to chars

thanks

which is no good unless you know for sure that your buffer is the same type of UTF-16 that the jvm uses
including byte order, etc

Gah. Stupid fucking computers. Nothing works.

they're just good at pointing out our human weaknesses

I'm fine when it's my weaknesses they're pointing out. I take criticism reasonably well.
But when I have to deal with the weaknesses in the n million lines of buggy third party code and arcane APIs I depend on, it really really gets to me.

I have java.lang.ClassCastException in this code:
String str = (String) o;
if (str.equals("true"))

so o isn't a String

no ?

if the VM complains about it, no it isn't.

I thought that it is properly, compiler doesn't show warnings

the compiler may not know.

how can I do it?

do what? i have no idea what you're trying to do

I have variable o of type Object
and I'd like to convert into as boolean
boolean bResult = Boolean.parseBoolean((String) o);

what type is it actually?

type of "o" ? Object

you said o = new Object() ?

type parameters of PacketTypePacketType cannot be determined; no unique maximal instance exists for type variable PacketType with upper bounds PacketType,com.echomine.xmpp.IStanzaPacket, however in eclipse the project is building without any error

no, I got value o from parameters of method

source and target levels are set to 1.5 in ant buildfile

public void onSuccess(Object o)

what else could cause such error?

so what's the actual type?

actual type is Object

How do I get an Ant task to invoke Ant on a bunch of build files that's in a folder? Is fileset supposed to be for this kind of purpose? It seems like it's for classpath kinda stuff
How do I get an Ant task to invoke Ant on a bunch of build files that's in a folder? Is fileset supposed to be for this kind of purpose? It seems like it's for classpath kinda stuff

no it's not.
that's the *declared* type of the parameter.

I must check it - how ?

but then how do I reference those collected *.xml files?

well, from where is that method getting called?

task

okay, can i use it as like arguments or something?

it is in: AsyncCallback

to echo out all the names, how can I do that?

GWT, eh?

If I need to write my own Ant task, that's fine, but just wonderin' if it's possible without.

it implements 2 methods: public void onFailure(..)
yes

how about you just pastebin the code?

ok, momento

I'm reading Java Security, how come it tells us "A lot" about ClassLoader, I mean why do we need to care about it, we don't need to care about loading classes right?

sometimes you do.

wrr, pastebin.com doesn't work

try #ant channel

yeah, ##ant's kinda dead though :x

well, looks like javac can't compile the class while Eclipse can
is it bug in Eclipse ot javac?

http://rafb.net/p/uGlQGd46.html
a href="http://rafb.net/p/uGlQGd46.html"http://rafb.net/p/uGlQGd46.html/a

depends on the structure i guess :x

open a bug against https://bugs.ecliipse.org/bugs/ Eclipse JDT Core with your testcase ... they'll investigate

I'll try asking in ##ant anyway.

code is good ?

well, it seems NOT be a bug in eclipse, I guess
but a bug in javac from Sun

what seems?

http://forum.java.sun.com/thread.jspa?messageID=4196171
I have similar case, which compiles well under Eclipse, bit isn't compiled by javac

Hello folks

hi

http://rafb.net/p/66iEFx89.html

humm decode method is returning underflow

my GWT experience is limited, but you control what gets passed to onSuccess, so just pass in a Boolean and cast it appropriately in that method.

Ok I am new to Java and have what is probably a simple question. I am writing an applet and I need to fill a JTable from a database can I pass the JTable into the funtion and return the JTable?

but converting to a String and then comparing to "true" is just retarded.
1. probably. 2. we don't discuss applets here. see the topic.

I thank you but this isn't just about applets this is passing the control from a class form to a class.

i know.

but there isn't method for convert into boolean :/

to convert what to a boolean?
the Object?

yes

Booealn b = (Boolean)o;
if o is actually a Boolean, that'll work just fine.

ok, next code is?

Are you simply trying to do something like JTable myFunc(JTable t) { /* do stuff */ return t; }?

next code what?

what do you write code

tell PawciooS about english

PawciooS, english is http://eugeneciurana.com/galereya/view_photo.php?set_albumName=Humor&id=Spelling_English_Jules

because I want boolean
not Boolean

why?

Exactly. I keep getting an error "cannont find symbol" I have tried imports javax.swing but that doesn't work either.

you import classes not packages.

because I want boolean for checking of values in next methods

ahhh so package javax.swing should fix it?

boolean?

no

No, you need "import javax.swing.*;"

smack joev
* imports

cheeser, * imports is http://en.wikibooks.org/wiki/Java_Programming/StarImports

you forgot that GWTCompiler is works on java 1.4 ?

I'm not maintaining his code :^)

you're giving stupid advice

Thank you for your help. I knew it was simple and I was just missing it.

I know that there is method from Boolean into boolean, but in version 1.4 it doesn't have

wtf do you think Boolean.booleaValue() is?

What is the right answer then?

I don't actually use "star imports" in any of my stuff, for what it's worth.

import the JTable class

I tried it, and it doesn't work :/

so don't advise newbies to do that.

Thank you.

tell PawciooS about doesn't work

PawciooS, doesn't work is useless. Tell us what it is, what you want it to do, and what it is doing. Consider putting some code and any errors on a pastebin. (use ~pastebin for suggestions)

I realized from joev response that was the correct thing to do.

Time to lobby Sun to kill star imports, then

but cheeser thank you for bringing it to my attention.

ok, please wait, I'll do code

I am in need of a way to programmatically gather all open file handles on a win32 machine....I fully suspect the for JNI, however, I do not know how to grap open file handles....any suggestions?

yes, you'll need JNI for that. for the details on how to do that in the win32 API, try a windows forum.

I fully expected an answer like that, but was hoping someboyd on the java forum had done it in one of their previous JNI/COM lives

i haven't personally
but i haven't done jni in years

d'oh... I have something strange here... Say, I have one Runnable object created, and start 20 threads with this object (only one instance). Now, in method run(), I have synchronized (this) {...} I thought that only one thread will be able to enter this synchronized section...

so what's your question?

lol, now it's works ok
I don't know, why it doesn't work ealier
thx for help hyhy

well.. it's not true... all threads entered my synchronized section

sure

but I already know why

it's probably not a good idea to pass the same Runnable to multiple threads

why?
why?)

you should think of Runnables *as* threads.
putting a Runnable in more than one thread seems ... wrong somehow.
i'm not a threading expert but it definitely raises some flags in my mind.

ok, if there is no other reason

I often write Runnables which are basically just immutable 'function' objects.

it works for me- let me paste.

right. if they have no state.

So there's no reason in principle it couldn't be passed to multiple threads.

I wanted the threads to share the same set of data, yes, I could create a "DTO class"

But I agree it seems a little strange.

d'oh... which? the sychronized?

It's possibly also useful in thread pooling circumstances? Not sure.

I said, that I know why many threads enter my sync. section --- I used wait() there... I didn't know that *any* method is able to affect the synchronized thing

http://eugeneciurana.com/pastebin/pastebin.php?show=2386
wait gives up the lock.

yeah... I said, I know why

"The thread releases ownership of this monitor and waits until another thread notifies threads waiting on this object's monitor to wake up either through a call to the notify method or the notifyAll method."
ah
I missed that part while I was whipping up an example.

next problem:

OH MY GOD ITS A MIRACLE. jboss 4.2 -actually- included their web console in the distro.
bOH MY GOD ITS A MIRACLE. jboss 4.2 -actually- included their web console in the distro./b

what's name of methods for create String for field encrypted via sha1 ? I want create string the same as exists in DB mysql for field sha1('password').

~tell PawciooS about javadoc MessageDigest

PawciooS, please see java.security.MessageDigest: http://java.sun.com/javase/6/docs/api/java/security/MessageDigest.html

big thx

I guess the factoid 'really big index' might be appropriate:
PawciooS, really big index is http://java.sun.com/docs/books/tutorial/reallybigindex.html

lol, you are know-all person

PawciooS, I have no idea what lol, you are know-all person is.

hey does the industry still used a lot of EJB, is it worth me learning this? I'm familiar already with hibernate...
i know EJBs run as part of an app server and use java persistence api

wtf, I have an american IP but I'm prevented from using netbeans only because it resolves to a .ir (iranian) resitered host

al qaida spy!

lol

it's only for fun
albeit it is terrorists.ir

Hello all... I have a simple question... is the SUN JVM the only one that uses the PermSpace concept? As in, is it the only JVM that requires the -XX:MaxPermSize argument to increase space for class creation?
Basically, I am asking if we switched to the IBM JVM or BEA's JRockit, should I see the same OOM exceptions regarding PermSpace?

probably not since each VM is probably tuned/configured differently

So, if I understand it correctly, switching to a different JVM COULD help us around the known CGLIB problem in JBoss/Tomcat/Hibernate issue?

MadOtis, what version of jboss are you using?
there was a permgen problem in jboss with jboss 4.0.x that was (in theory) fixed with 4.1 and later.
i'm evaling 4.2 now

our "standard" is 4.0.4, but I am experimenting using 4.2.0 with JRockit

4.0.4 absolutely has a permgen problem.
PermGen space

Cool, that's what I read. I am trying to build a case to move to 4.2.0 and wanted to make sure it still wasn't a JVM issue-jboss combination

this is ahuge one

Yes, it sure is.

my understanding is 4.0.x leaks deployment descriptors like a sieve. so if you're deploying a lot, it'll die a horrible death soon.

I've never had 4.2.0 run out on me, now that I think of it
4.0.x did all the time

Since we're under development, and constantly deploying, yes, it is a severe issue for us.

4.2.0 on JDK 6 seems to tickle a JVM bug on linux though
so I'd recommend using the latest java 5

i haven't tried 1.6 yet at all.

We can't use 6 for some interoperability reason... at least, that is what I was told.

though i just got a interesting deployment error with 4.2.

I ran across the bug about twice in the last 6 months, so it's not major, but I wouldn't put it into production yet

http://pastebin.stonekeep.com/518

I'm trying to vie for a case of using jboss 4.2.0 with JRockit... it seems to outperform Sun hands-down on our linux development server.

i think i hosed a JMS listener. *fiddles*

MadOtis, on what basis? I/O, or just plain execution speed?

never used jrockit.

I haven't profiled it yet, but, using Sun's 1.5.0_11, we see about 1.5 seconds per page load time. It's just under a second (SOTP) using JRockit

interesting
I wonder how it would compete with 1.6
which seems a bit faster to me, just observationally

hum. i may need to update jboss-ide.
it's not finding my jboss4.2 distro

re
how's it going?

okay... we locked down db4o. it's all working properly now.

excellent

about 80% my fiddling, and one last bit by the other programmer.
(she was storing a DSO object into db4o)
It No WorkSo Well Like That.

send me the mbean again, if you would? to email?

still in the pastebin, i think. sec.
http://pastebin.stonekeep.com/514

grazi
I'm going to embed it into glassfish too
you have a gmail account?

me? i do, but i never use it. why?

google docs uses your gmail account id

actually, no it doesn't.
i log into google docs with dbs@stonekeep.com

it should!
Ah, you have two accounts then

i have many

we should write up something on embedding services like this

hrm.

thou dost not agreeth?

no, not that. i think the jboss-ide plugin in eclipse, which i'm showing as the latest version, is not cmopatable with jboss 4.2-GA. the plugin doesn't look like it's been updated in 7 months.
so i can't add a 4.2 server.

ah
what about the exadel stuff?

gr.

alternately, what about keeping a connection up?

GRRRRRR.
yeah well
i'm also tired of jboss.org's fucking website having broken links left and right.

they have a porn website TOO?!
Wow

don't make me kill you, man.

JBoss porn/
?

Someone has to

i'll do it!

^Migs^: "jboss.org's fucking website..."

ooh! ooh! mr. kotter! mr. kotter!

Hey

ah okay

wow. i can't beleive you just made a reference to a mid-70's tv show.

~mid-70s++

okay...that's....enough

I don't know who that is.

can you pass a class type to a function? as opposed to a class instance

of course you can

yay. jboss-ide hasn't been updated since 10-11-2006

this assumes the method's expecting one, of course

well that's just DUCKY.

exadel...

well, yes, more specifically, how can I make a member function accept a class type?
I just need a syntax example

exadel?

exadel studio is going to be the next jboss ide

that's why there's no new updates?

I think so
not a bad move for jboss though, exadel is nice

how stable is exadel now?

I use it

well, since it's now impossible fro me to use jboss-ide, i think i'll have to give it a whirl.
looks nice though

yeah, it's pretty good

would be nice toa ctually be able to edit JSP / HTML live in a tomcat/jboss env.

doesn't hurt that the exadel chaps aren't cretins

i've heard good things.

hi all
we are going to program with java in the next semester

whoa cool!

congratulations!

and so i want to prepare myself

We finally get Java into a school somewhere!
yay, toasters all 'round

it only took 11 years.

it's cutting edge. brand new. might not find anyone who really knows it.

persistence++

can you recommemd a good book

eidoLaptop, why don't you just edit it live in the jboss environment then :p
eidoLaptop, you don't need an ide for that...

tell litb about tij

litb, tij is Thinking in Java by Bruce Eckel, see http://www.mindview.net/Books/TIJ/ for the 3rd edition (free download, published in 2002) or see http://www.mindview.net/Books/TIJ4 for the current edition which covers Java 5 features.

tell litb about books

litb, I guess the factoid 'free books' might be appropriate:
litb, free books is http://www.planetpdf.com

hrm.

pfn, because it's a pain in the ass to do that. we deplopy the entire EAR into the appserver to start and run it.

i already know c++ fairly well

eidoLaptop, yeah, but you can cherry pick files to deploy

so, perhaps not the most beginners book

eidoLaptop, or just unjar your jsp files into there

then TIJ is probably a good one

i'm not sure how i can set up jboss et al to use the edit 'source' directory as HTML / JSP sources.
my first java book was 'teach yourself java in 21 days'. was a good start.

eidoLaptop, e.g. I'm often doing something like jar xf file.war WEB-INF/jsp and /content

unpacking it where?

eidoLaptop, into the context root of tomcat
or whatever container you use
instead of using a full deploy cycle

weird. so you run your packaging to build your distro, then unpack part of it?

how do you do it?

eidoLaptop, that, or just copy the file directly from the project over

does tij cover all features of jav?

ew. that's not really a live edit
'all features'?

eidoLaptop, it's live enough

how do I do what?

that would require an entire library.
i'm used to PHP development. edit, save, refresh

i mean, templatrs, exceptions, perhaps the garbage collector and streams and such?

I want to pass any type that is a descendent of Item and have the method instantiate and do stuff with it

*templates

eidoLaptop, then just open the file directly in eclipse

I like how you say "the garbage collector" like there's only one

yah, see, that breaks something else - because the project definition, as coupled with SVN, gets hosed.

eidoLaptop, add the context root as a project and edit that way

because i'm not editing my 'source' files.

it's charming and slightly endearing

i know broce eckels has written a good book about c++ too.

what i really need to do is the other way around - tell jboss that my project source dir is the JSP / HTML source...
so a save to a JSP will trigger tomcat to rebuild.

eidoLaptop, create a subproject or something
eidoLaptop, well, tomcat can set docBase on contexts

*nod*...

that ^

i don't know anything about that stuff. so forgive me if i simple it down

methodname(Class clazz)

we do very little HTML and JSP dev right now, would like to do more. 99% of our functinality is in EJB's

then call methodname(foo.class)

(and first person to say 'yeah, but jsps are ejbs' gets a boot to the head)

who says something stupid like that...

but... you can use JSPs like EJBs!

who says jsps are ejbs?

hmm, I can do Object foo = new clazz(); then?

no... but you can do foo=clazz.newInstance()

well, jsps become servlets. servlets are just a form of ejb, right?

yes!

eidoLaptop, uh, no

Or something

:p

hmm, ok, thanks

glad we have consensus here.

servlets are part of j2ee/javaee, and EJB is part of j2ee/javaee

what's your point?

no point
sayin

there goes my client.

i fully agree with it, dmlloyd

the only confusing are those names. j2ee, jre jsdk jbeam and such . i don't see through the wall of terms

eidolon, damn dude, /msg nickserv ghost eidolon next time

heh.

but it seems like bruce is more like a c++ guru

how do i set heap size ? export JAVA_OPTS="-Xmx256m" ?

did you try that?

do you even know what the java heap is, exactly?

it's that steaming thing in the corner

usually, people who do that get less than half of what they expect

is it java heap or just heap, jottinger ?

tool docs

cheeser, tool docs is http://java.sun.com/javase/6/docs/technotes/tools/

read up on the 'java' command

thanks

and then look up the garbage collectors, and look up tuning java heap
and fear. Do much of le fearing.

mm, f33r.

jottinger well i have to set specific size, as specified by app vendor

wow
which app?

alfresco

huh, never looked at that before.
yay cms.

ahh
john newton & co

jottinger is it only garbage collector that uses heap ?

man, I know I shouldn't be, but I'm astonished that this app, written in jdk5 with 3 jdk5 dependencies, actually runs in 1.4 with jbossretro

heh

you five maroons!

because, you know, life isn't painful enough.

right
I want to ditch 1.4 compat, but my boss says no

here's a sharp pointy stick. jab it in your eye a few times to complete the experience.

can interface implement another interface?

implement? no... extend, yes

no, but it can extend

~tell psyko_x about interfaces

psyko_x, interfaces is http://java.sun.com/docs/books/tutorial/java/concepts/interface.html

ok thx

and i guess, as an ide you recommend eclipse?
or just a plain text editor?

newbie ide

newbies shouldn't use IDEs. http://weblogs.java.net/blog/editors/archives/2007/02/you_better_get.html and http://weblogs.java.net/blog/gsporar/archive/2007/02/tools_that_do_t.html

i mean, i usually program c++ using emacs
can i use it also for java? i does java need some special things?

see. i really don't agree with that factoid.

i do.
newbies should learn the environment before using an IDE.
i can't tell you how many newbies can't run their java app outside of eclipse or don't understand classpaths and packages becuase the IDE obscures all that.
i can't tell you how many newbies can't run their java app outside of eclipse or don't understand classpaths and packages becuase the IDE obscures all that.

i want to caveat this.

yes, i agree

i don'mt agree with the factoid in general. i do agree that eclipse is _NOT_ the place to learn java.

hell,l a lot of them can't even *build* their software outside of an IDE

no ides for newbies

that's my recommendation

brb

i know what a classpath is
it is a directory where java searches for .class and .jar files

lets say I have a java applet whcih depends upon a particualr library
would i need that library to be packegd with jar of applet?

wrong!
~topicsmite geoaxis

And the wrath of /TOPIC descended with terrible fury upon geoaxis. And all the people marvelled, saying, Behold, we too should read the /TOPIC, lest we be stricken. And all the people read the /TOPIC, and went away edified.

cheeser, i see
the problem is i have to have my java component within a webpage

HAYE thar!

hi MickMcMack

AVAST

:3

post the Cache class, eh

i swear i'm oing to just give you access to the svn server.

well, hey, an mbean like this would be great for a db4o external project...

http://pastebin.stonekeep.com/519
a href="http://pastebin.stonekeep.com/519"http://pastebin.stonekeep.com/519/a
a href="a href="http://pastebin.stonekeep.com/519"http://pastebin.stonekeep.com/519/a"http://pastebin.stonekeep.com/519"http://pastebin.stonekeep.com/519/a/a

no, it is
I wasn't sure which methods you overrode

quick question guys, if I want to use a class in an applet that I wrote, is there a way to statically include it or something?
just having in the same directory doesn't seem to give it to the client to find

List.remove, when passed a refrence to a object, will only remove exactly that object, correct?

gaillard_, you have to import it.
~tell gaillard_ about import

gaillard_, import is http://java.sun.com/docs/books/tutorial/java/interpack/usepkgs.html

I did but it says it can't find the class def
I am assuming because on the cleint they don't have in their class path my helper class, so is there a way to package it with the applet .class?

~topicsmite gaillard_

And the wrath of /TOPIC descended with terrible fury upon gaillard_. And all the people marvelled, saying, Behold, we too should read the /TOPIC, lest we be stricken. And all the people read the /TOPIC, and went away edified.

Why would intarwebs not work when IRC/MSN still work? ;_;

how is that off topic?

Read the /topic.
The channel wiki has answers to your questions.

~ping

http://www.nataliedee.com/071405/ping.jpg

Wtf?

hee hee.

It's usually a good idea to read the /topic on a channel you aren't a regular at

Must be a stupid DNS thing.

OKi, shower time.

Don't do anything I wouldn't! _

i am only asking if there is a way to import a whole class into another not just the names and stuff

but he has to wash...

gaillard_, the channel is very very opposed to applets in general, so there's a stigma on them.
are you trying to import something from an applet into a standalone app?
or is this something in an applet itself?

no just a class into another class

jottinger, but washing is so boring when you can do more fun activities, like play with the bubbles! :o

import will only import something it finds on the classpath.

but in a way that the class being imported wouldn't have to exist on the person running the code

if it's saying it can't find it, then it's a classpath issue.

i don't get compile errors
only when running it says it can't find a definition

then it's in the compiler environment but not in the runtime environment.

tell eidolon about busy

eidolon, http://tinyurl.com/yqlpmt

8^)=

yes, so can i package it?
like a c++ include does?

hee hee

java doesn't have includes

so there is no way to put two classes in one basically?

That means C++ is better, right?

something like that

if they were "in one" they'd be... one class, not two
i.e., go learn jack shit about classpath

so, i'll try this slower. "Your runtime environment is not finding the class. The class needs to be in the classpath of the JRE the app is running in" - further edification will require you to understand how classpaths work.
~tell gaillard_ about classpath

gaillard_, The class path tells Java or the compiler in which jar files and folders to look for classes. Use the -cp/-classpath run-time options to specify the class path. Also see http://java.sun.com/j2se/1.5.0/docs/tooldocs/solaris/classpath.html and http://mindprod.com/jgloss/classpath.html#ANACHRONISM

i know that i just didn't know there wasn't anything that didn't act like c++'s include

erm

now you know.

thanks

the fact that you, like, NEVER saw anything even remotely like #include didn't clue you in?
and the "differences between C++ and Java" saying "Java doesn't have #include" meant nothing?

didn't mean it didn't have something else similiar...

I only ask because I care

will the gc dump an object if one of it's functions is executing?

gc will dump anything that has no references to it.

crap
cause I have this method, that removes references to itself, and then adds new ones

*thinks a moment*

is there somewhere i can get help about applets?

~tell gaillard_ about applets

gaillard_, applets is a Java program that can be executed from a browser over a network (http://java.sun.com/applets). See http://javachannel.net/wiki/pmwiki.php/FAQ/Applets

thanks

and, in case you h aven't gotten it yet, please dont' use applets.
use webstart.

will that cause the gc to dump the object if I have unlucky timing?

yuriks, don't take my word on that. there's a situation where something is Runnable and it's a running thread. i don't know how gc treats that. have to ask others.

Applets still have their purposes when you want to execute something within a browser though yes? Web start can't do that as far as I know.

READ THE TOPIC and the other information off the wiki. NO APPLETS.

Consty, there are -very very very- few situations where an applet is actually what you want to use.
they do exist, but 99.9% of the folks using applets shouldn't.

Rightly so.

I think you guys are just biased against the poor applets among you
applets dint do nuffin to you!

jottinger, only because they SUCK SUCK SUCK SUCK SUCK!

~tell jottinger about STFU.

jottinger, http://eugeneciurana.com/galereya/view_photo.php?set_albumName=Humor&id=warning

Tee, hee, hee...

u ppl r just mean

smack jottinger

hmm, java seems to support templates

nope

can i also partial specialize those?

They even have slavery in here jottinger. JavaBot is a person and he just does whatever the rest of them want.

tell litb about generics

litb, generics is http://java.sun.com/j2se/1.5.0/docs/guide/language/generics.html Generics tutorial: http://java.sun.com/j2se/1.5/pdf/generics-tutorial.pdf FAQ: http://www.angelikalanger.com/GenericsFAQ/JavaGenericsFAQ.html

free javabot!

generics are definitely NOT templates
they just look kinda similar

i dont' know templates at all.

... without the explosion of :: and crap like that

then again, i don't know c++. or c. or even how to tie my shoes.

heh
eidolon, if you're going to learn C++, do it right and use bjarne stroustrup's book, otherwise don't bother

The process of tying your shoe is probably more complicated than any program anyone in here as written.

. although they have proven to be problematic

but don't learn object orientation from his book

believe it or not, I learned OOP from programming in a MOO

idiomatic C++ is really it's own kind of programming

good thing it wasn't a POO

nothing else quite like it
not even remotely similar to java

Ideally, learn OOP from smalltalk or some other language without braces.

gahh. i don't think i should try installing exadel into easyeclipse server version.

c++ is not similar to java
c++ has many more techniques than java

c++ is sloppy

this is funny

with good reason

http://typicalprogrammer.com/programming/abject-oriented/

try to solve the n-queen problem with java generics
is that possible?

is there a way to figure out if an object is made of other objects? as an example, is there a way to figure out that a String is made up of Characters?

i mean, at the time java byte-compiles the source, not at runtime

you can even solve the n-queen problem without generics in java

generics aren't C++ templates

There is always going to be some things that can't be made as conveniently in Java as in C++. That doesn't make it a better language.

why would you need to solve the n-queen problems with generics

Maybe reflectively look at all member field types

sounds like the wrong approach

String instanceof CharSequence i guess

ok i'll look into that
liesn the String example is just one example. i need this to work for any generic object

i have not said that c++ is better
only that it provides more possibilities to code

sure, having pointer arithmetic basically doubles your coding options
so?

sometimes have 15 ways to solve the same problem is less good :-)

I think the point which litb is making is that C++ can solve the n queens problem at compile time.

it can? heh
and that's desired?

guys, I have a String and a ByteBuffer , i'm putting the String into the bytebuffer using string.getBytes() method and bytebuffer.put, but I also need to add a \r\n to it, if I use put again will it overwrite or append?

i call it the power of choice

C++ templates are a (lousy) form of compile time metaprogramming whereas Java generics are a form of parametric polymorphism.

I call it write-only source code

you halso have parametric polymorphism with c++

It's sometimes useful. It's not really the intended effect, but things like boost, etc. (ab)use it to get some really cool effects.
Not really. You have templates. It's not the same thing.

my understanding is that ArrayLists are better for large amounts of values and LinkedLists for small amounts?
is that right?

Sortof. LinkedLists have better append properties.

no
it all depends on what you do with the lists

I randomly add, remove and iterate through it's values

probably better for linkedlist altogether then

For ridiculously large data sets an ArrayList will tend to be smaller. But your data is unlikely to be large enough for that to matter.

hmm, ok, thanks =]

is there a standard class, like Pair in c++ , that couples 2 objects

is ArrayList any faster when doing .toArray? (I guess yes, since it can just copy the internal array)

I don't think it makes a lot of difference.

yes, and it usesarrays internally VERY OFTEN

It's proabbly some constant factor faster.
HS^: No.

look up what ArrayList does when you remove()

Hello, what does return d1 != d2 ? 1 : 0; mean? (where d1 and d2 are doubles; http://java.sun.com/docs/books/tutorial/java/nutsandbolts/op2.html explains only the case where there is a = instead of != and where d1 is a boolean..)

remove = arraycopy down
and setting last element to null

if (d1 == d2) return 1; else return 0;

so i have a ImageIcon and a (selfmade) Coordinates object, i want to place many in a datastructure//
What would be the best datastructure for this? MultiDimensional arrayList
?

Consty, thanks

when you use a string interator to iterate does the iterator automatically traverse over each character?
~iterators

psyko_x, I have no idea what iterators is.

~iterator

psyko_x, I have no idea what iterator is.

~Iterator

HS^, I have no idea what Iterator is.

javabot is getting molested .

Invalid factoid name

What do you mean by 'automatically'? It will iterate over the things you make it iterate over.

like if i try to print out what the iterator is looking at, will it print out characters? it seems like if u define a string iterator it should traverse over a string at a time, not a char at a time
nvm i got it

so ttheres no conenience class in java to prevent multidimensional arraylists?

i mean CRTP

Is there a reason why you must have the "main method" in a class? Or is it just made that way?

HS^: it isn't needed

Curiously Recurring Template Pattern

dmlloyd.. a lot is not needed

Ah. I have no idea what CRTP is.

it allows you to have parametric polymorphism

myListRox = new ArrayListListListListString();

done

ok. I'll believe you.

in java, templates aren't generally needed since everything is done by reference

also c++ allows you over a template parameter to chose the base class. and thus, allows you to chose whether you have a virtual method or you have not

all methods in java are virtual

ah

by the C++ definition

Except when they're not, as a compiler optimisation.
But that's not visible at the language level.

but the user will never see that

Right.

opa
algum brasileiro ?

try irc-hispano

brazil is portugese
if that matters

[andresgr]: isso mesmo

dacordo
eu son galego, se che vale
portugués e galego moi parecidos

in the www.canaljava.com.br invit for here
isso é verdade
is true

one thing that bothers me is the distinction of interface and base class
it looks logical to be if these two are the same

an interface is a pure virtual class

well, they aren't
an interface defines that, while a base class defines behavior

so, no implementations

hey.. is there a way to refresh the display when adding swing elements to a frame?

usually that happen automatically...

urgh, eidolon used jboss jmx stuff

vikingr, you may want to call invalidate() to hint that the layout needs to be refreshed

but you can revalidate the container if you have to

pfn, ernimril thanks, i'll try that

Most people use HTTP as the transport for web services. Anyone using JMS for that matter?

I don't see the point of using soap over wsdl... services aren't discoverable like they are via the web...
s/the web/http
not in a "standard" fashion at least
I think http soap services often append ?wsdl to nab the wsdl, isn't that true....
I think http soap services often append ?wsdl to nab the wsdl, isn't that true....
s/soap over wsdl/soap over jms

hehe