 |
AppletTalk.com Java discussions newsgroups
|
| View previous topic :: View next topic |
| Author |
Message |
server Guest
|
Posted: Tue Jan 24, 2006 3:46 pm Post subject: -Xmx100m works vs -Xmx300m |
|
|
message unavailable |
|
| Back to top |
|
 |
neo Guest
|
Posted: Tue Jan 24, 2006 3:46 pm Post subject: Re: -Xmx100m works vs -Xmx300m |
|
|
There are several system threads which run as daemon ( finalizer, gc,
hotspot detector etc, which depends on wich jvm you are using). Why
dont you remove (or reduce) -Xss parameter.
Nirav |
|
| Back to top |
|
 |
Simon Guest
|
Posted: Fri Feb 03, 2006 8:23 pm Post subject: Re: Changing the System.getProperty("os.name") value |
|
|
ktdreyer (AT) gmail (DOT) com schrieb:
| Quote: | Thank you, but this is not my applet and I do not have access to the
source code. Is there another way to accomplish this?
- Ken
|
You can extend the applet's main class, override init() to make the call to
System.setProperty(), and possibly add the new class to the applet's jar. |
|
| Back to top |
|
 |
Roedy Green Guest
|
Posted: Fri Feb 03, 2006 8:48 pm Post subject: Re: Changing the System.getProperty("os.name") value |
|
|
On Fri, 03 Feb 2006 16:23:41 +0100, Simon <count.numbers (AT) web (DOT) de>
wrote, quoted or indirectly quoted someone who said :
| Quote: | Thank you, but this is not my applet and I do not have access to the
source code. Is there another way to accomplish this?
|
You can always disassemble or decompile them. The code is remarkably
like the original.
See http://mindprod.com/jgloss/decompiler.html
--
Canadian Mind Products, Roedy Green.
http://mindprod.com Java custom programming, consulting and coaching. |
|
| Back to top |
|
 |
Thomas Hawtin Guest
|
Posted: Sat Feb 04, 2006 3:09 pm Post subject: Re: Changing the System.getProperty("os.name") value |
|
|
ktdreyer (AT) gmail (DOT) com wrote:
| Quote: | Thank you, but this is not my applet and I do not have access to the
source code. Is there another way to accomplish this?
|
If you run it in the appletviewer, then you can use:
-J-Dos.name="Mac OS"
Note, all the other properties will remain the same, as will various
implementations. It wont start looking like Aqua.
Tom Hawtin
--
Unemployed English Java programmer
http://jroller.com/page/tackline/ |
|
| Back to top |
|
 |
Roedy Green Guest
|
Posted: Mon Feb 06, 2006 5:20 am Post subject: Re: Changing the System.getProperty("os.name") value |
|
|
On Fri, 03 Feb 2006 16:23:41 +0100, Simon <count.numbers (AT) web (DOT) de>
wrote, quoted or indirectly quoted someone who said :
| Quote: | You can extend the applet's main class, override init() to make the call to
System.setProperty(), and possibly add the new class to the applet's jar.
|
And I hope it goes without saying , calls super.init().
--
Canadian Mind Products, Roedy Green.
http://mindprod.com Java custom programming, consulting and coaching. |
|
| Back to top |
|
 |
Thomas Hawtin Guest
|
Posted: Tue Jun 20, 2006 6:23 pm Post subject: Re: Which Java version to target? |
|
|
Mike Amling wrote:
| Quote: | wizofaus (AT) hotmail (DOT) com wrote:
I would strongly suggest using 1.4x
That's my current intention. It means no generics, which as a C++
programmer is frustrating, but I can live with it.
That's odd. I haven't used it myself, but I'd read that Java's
generics facility involves no changes to the bytecode language.
|
The bytecode instructions themselves remain the same, but there are
extra attributes added to the class files.
| Quote: | What
happens if you try to compile with
javac -source 1.5 -target 1.4
|
Try it. It wont.
1.5 has the undocumented and unsupported -source jsr14 (IIRC).
Does anyone have any figures for which versions browsers have installed?
The online installer for Windows should have updated most to 1.5.
Windows/Solaris/Linux users that have used offline installers in the
last year and three quarters will presumably mostly installed 1.5. No
idea about Mac users, other than Apple is usually around a year behind.
Tom Hawtin
--
Unemployed English Java programmer
http://jroller.com/page/tackline/ |
|
| Back to top |
|
 |
Thomas Hawtin Guest
|
Posted: Wed Jun 21, 2006 5:10 pm Post subject: Re: Which Java version to target? |
|
|
wizofaus (AT) hotmail (DOT) com wrote:
| Quote: |
Most of the Windows systems I've tried it on only have 1.4.2_03, so
|
Seems a bit old. 1.4.2 is up to at least 1.4.2_11 (some of those
security updates).
| Quote: | that's my target at this point. But apparently even that's not enough
- there are things that fail mysteriously under that version that work
fine when you have the 1.5 JRE installed (e.g. JFrame.add() vs
JFrame.setContentPane()), so you absolutely have to test on a system
that only has 1.4.2 installed.
|
I'm sure you can develop it on a machine that has both 1.5 and 1.4.2
installed.
| Quote: | I'm just getting used to having my code littered with casts. The fact
that it took till version 1.5 of Java until there were typed resizable
collections in the language is a little baffling to me.
|
It was dragged out, but it isn't so easy to retrofit such a feature.
Tom Hawtin
--
Unemployed English Java programmer
http://jroller.com/page/tackline/ |
|
| Back to top |
|
 |
Philippe Amarenco Guest
|
Posted: Wed Nov 29, 2006 4:49 am Post subject: Re: Realtime Java - Getting Started |
|
|
"John" <javacc1 (AT) gmail (DOT) com> writes:
| Quote: | busysteve wrote:
I've written realtime software in C/C++ and I've written system
applications in Java. What I haven't been able to do is write realtime
software in Java (use of hardware interrupts and such). I've searched
for embedded development kits and APIs. I found javax.realtime but no
JVMs that support it. I've also found various kits that say they
support java but not to what degree. This is for fun so I don't want
to get crazy on cost. If you have done this can you point me in the
right direction?
I really doubt if Java really can fit in embedded systems development.
What types of systems applications you wrote in Java? I thought Java is
for high-level application development.
|
actually, the javax.realtime (www.rtsj.org) is an API quite close to
hardware and ressemble that of a realtime kernel. It gives great
control over memory allocation and object lifetime so it is much more
predictible.
by the way,
I have started a (free/open-source) project (initialy for school)
which aims to implement a JVM + CLDC + RTSJ.
http://sourceforge.net/projects/joker-vm
Anyone interrested in contributing is welcome to do so. (you'll have
to checkout/browse the svn rep', no tarball has been released yet).
--
Philippe Amarenco, aka Phix
epita 2007 - GISTR - ACU - EpX |
|
| Back to top |
|
 |
Guest
|
Posted: Thu Dec 14, 2006 9:59 pm Post subject: Re: Realtime Java - Getting Started |
|
|
Philippe Amarenco wrote:
Hmm, the project was started 4 years ago. |
|
| Back to top |
|
 |
Joshua Cranmer Guest
|
Posted: Fri Mar 02, 2007 4:44 am Post subject: Re: CMS collector ignoring -XX:CMSInitiatingOccupancyFractio |
|
|
M wrote:
| Quote: | I have a Java service that appears to ignore the
XX:CMSInitiatingOccupancyFraction flag completely. Here're the GC
related settings:
-XX:+PrintGCDetails
-XX:+PrintGCTimeStamps
-XX:+PrintTenuringDistribution
-XX:+UseConcMarkSweepGC
-XX:+CMSIncrementalMode
-XX:CMSInitiatingOccupancyFraction=50
-XX:+UseCMSInitiatingOccupancyOnly
-Xmx2048m -Xms2048m
-XX:NewSize=100m
-XX:MaxNewSize=100m
Despite setting CMSInitiatingOccupancyFraction to 50% and setting
UseCMSInitiatingOccupancyOnly to true, the CMS collector continues to
wait until the tenured generation grows *way* above 50% before it
kicks in.
The reason for 50% threshold is this app being very sensitive to
pauses and I want to avoid pause if at all possible, at the expense of
memory and CPU. However because CMS doesn't start early as it should
(should start at 50% tenured occupancy), I occasionally get YJ
promotions failures and results a STW collection pause, probably due
to fragmentation.
This is JRE1.6.0 on Linux. What's going on??
|
This is a virtual machine question, and so belongs on c.l.j.machine. |
|
| Back to top |
|
 |
Lew Guest
|
Posted: Sat Mar 03, 2007 8:21 pm Post subject: Re: CMS collector ignoring -XX:CMSInitiatingOccupancyFractio |
|
|
Joshua Cranmer wrote:
| Quote: | M wrote:
I have a Java service that appears to ignore the
XX:CMSInitiatingOccupancyFraction flag completely. Here're the GC
related settings:
-XX:+PrintGCDetails
-XX:+PrintGCTimeStamps
-XX:+PrintTenuringDistribution
-XX:+UseConcMarkSweepGC
-XX:+CMSIncrementalMode
-XX:CMSInitiatingOccupancyFraction=50
-XX:+UseCMSInitiatingOccupancyOnly
-Xmx2048m -Xms2048m
-XX:NewSize=100m
-XX:MaxNewSize=100m
Despite setting CMSInitiatingOccupancyFraction to 50% and setting
UseCMSInitiatingOccupancyOnly to true, the CMS collector continues to
wait until the tenured generation grows *way* above 50% before it
kicks in.
The reason for 50% threshold is this app being very sensitive to
pauses and I want to avoid pause if at all possible, at the expense of
memory and CPU. However because CMS doesn't start early as it should
(should start at 50% tenured occupancy), I occasionally get YJ
promotions failures and results a STW collection pause, probably due
to fragmentation.
This is JRE1.6.0 on Linux. What's going on??
This is a virtual machine question, and so belongs on c.l.j.machine.
|
It fits just fine here and I would like to know the answer, too.
-- Lew |
|
| Back to top |
|
 |
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
|