Another classpath question

Hi,
I used to programm in java but then I took a break because of school issues for a long time. Now it seems like Ive forgotten everything.
Well I installed java again (because reinstalled windows meanwhile) and I successfully set my PATH variable, which means that compiling of *.java files works easily.
But now I have the problem of running my *.class file. If I type "java HellloWorldApp" (like written in the tutorial) I get the error:
"Exception in thread "main" java.lang.NoClassDefFoundError: HelloWorldApp "
It says in tutorial to try "set CLASSPATH=" which I did and it worked. But now I want to have that permanently. So what should I do? Should I change the CLASSPATH enviroment variable to an empty value in windows xp?
thanx
UsulSK
P.S.
It also sais in the tutorial about that problem: "For more information, consult the section 5. Update the PATH variable in the installation instructions.". I went there but there is nothing written about that problem and it is only explained how to set the PATH variable which I already managed to do.

CLASSPATH denotes the path for searching your class files. So obviously it cannot be set to empty :)
Since yours is WinXP, to set CP permanently, rightclick on My Computer and click on Properties. Choose Advanced tab and click on Environment variables. There you will find CLASSPATH. Click on Edit button to modify the path. Leave the path you see as it is and then append it with the following path of your lib directory
c:\<java_install_dir>\lib
where C: is the drive where your java is installed. If you are doing your programs on a different directory such as C:\UsulSK\Java, you can include the same in the classpath. Always separate different paths with semi colon.
Click on OK to confirm the modification. The new classpath takes into effect when u open a new cmd prompt window... .
HTH
PS: Read material on setting classpath from sun website

Similar Messages

Another newb question: multiple virtual servers

Hi, I have yet another ignorant question. I have several unrelated web projects that I am working on, and I would like to be able to set up a virtual server for each one for testing purposes, such as: http://project1, http://project2, http://project3. Can someone tell me if this is doable, and if there are any tutorials/resources on this for someone who has 0 experience running a web serer? Sorry for being so ignorant!

Yes, it is doable.
You can setup virtual server either by IP or by name.
If you have one IP, and want to set them up by name (ex. http://project1, http://project2, http://project3) you can do so easily with this type of configuration:
<virtual-server>
    <name>mydomain</name>
    <http-listener-name>http-listener-1</http-listener-name>
    <host>*.mydomain.com</host>
    <document-root>/www/domain</document-root>
</virtual-server>
<virtual-server>
    <name>myotherdomain</name>
    <http-listener-name>http-listener-1</http-listener-name>
    <host>*.myotherdomain.com</host>
    <document-root>/www/myotherdomain</document-root>
</virtual-server>
....The important part here is that
a) all virtual servers share the same HTTP listener
b) which virtual server serves the request depends on the $HOST request header send by the client. Sun Web Server does the matching for you. It will match $HOST vs. the virtual server's host attribute. Depending on which site you connect to the right virtual server will be used.
c) if the $HOST request header does not match any of the virtual servers, then the default virtual server defined in the HTTP listener will be used.
To create a virtual server, use the Admin GUI, access the configuration, and then add new virtual server. Or use the following CLI command.
wadm> create-virtual-server --config=myconfig --http-listener-name=http-listener-1 --document-root=/www/docs/myserver.com --host-pattern=myserver.com --log-file=../logs/myserver.com-error_logs myserverHost pattern will be used for matching. Some of this elements might be optional.
Hope that helps. And keep the questions coming :D
Edit: Also check the documentation
Using Virtual Servers in SJS Web Server 7.0

Another simple questions

Hello friends:
Another simple question: I need to learn things about Oracle on my desktop.
My machine runs Windows 98. Oracle has some desktop product of its Database Line?
For example: Oracle Personal?
And Oracle Lite? What's the main difference between Oracle Personal and
Oracle Lite?
Thank You
Gracias
Ing. Pablo Romero
CORDOBA ARGENTINA

1. I didn't know the answer to your first question, but I googled it and it says the item is the in-call audio boost.
http://forums.crackberry.com/f71/flag-icon-47659/
2. isn't this setting determined by the carrier? So it's not a setting in the phone, but when you call in to your voicemail you can change your options?

Another classpath-package question

I have two classes residing in a directory called Record, and both classes declarepackage Record;as their first line of code. The RecordObject class instantiates the DBConnector class, and will not compile unless I strip the package declarations off both classes. The compile error for RecordObject (when both classes have the package directive) is:
C:\jswdk-1.0.1\examples\WEB-INF\servlets\Record\RecordObj.java:9: cannot resolve symbol
symbol : class DBConnector
location: class Record.RecordObj
static DBConnector dbc;
^
C:\jswdk-1.0.1\examples\WEB-INF\servlets\Record\RecordObj.java:27: cannot resolve symbol
symbol : variable DBConnector
location: class Record.RecordObj
connection = DBConnector.getInstance().getConnection();
when only the RecordObject class has the package directive, the error is:
C:\jswdk-1.0.1\examples\WEB-INF\servlets\Record\RecordObj.java:9: cannot access DBConnector
bad class file: .\DBConnector.class
class file contains wrong class: Record.DBConnector
Please remove or make sure it appears in the correct subdirectory of the classpath.
static DBConnector dbc;
Following the recent classpath posts, I just changed my classpath, but this did not help the compilation:
C:\>set CLASSPATH=%CLASSPATH%;C:\jswdk-1.0.1\examples\WEB-INF\servlets;
^

This is even stranger:
When I enter "set" in the dos window that I set the classpath in, I see my setting. However, if I open a separate dos window and enter "set", there is no CLASSPATH entry at all.
When I try to compile my RecordObj class from the command line of the dos window with the classpath, all is fine. I had been compiling from Textpad before. But when I try to compile from the other dos shell, I get the same errors that TextPad showed me.
Could somebody explain how setting the CLASSPATH only sets it in one dos window, and not in the total environment? Thanks.

Manifest Classpath Question oc4j.jar

I need some clarification on how the MANIFEST.MF in the oc4j.jar library is influencing classlading. Below are my specific questions:
My group is using a configuration management tool which stores all java libraries in a central repository. During builds the tool downloads all dependent libraries in basically a flat directory format and specifies a classpath to java when it runs its builds (and starts up OC4J via java com.evermind.server.OC4JServer). Is there anyway to override the Manifest.MF with the classpath constructed by our configuration management tool?
In many cases libraries are entered in twice but are from different locations. i.e.
../../lib/xmlparserv2.jar
../../oracle/lib/xmlparserv2.jar
Is there a difference between these two libraries?
In the classloading sequence when does the classloader refer to the Manifest.MF for locations of libraries?
Thanks in advance!
-Mark

Hi Mark,
I'm sure I haven't understood your question, so the following may be useless to you (please excuse me if it is).
I feel there is probably a better way to do what you are trying to do, but without more details, I can only guess.
Do you know about the java extension mechanism?
The tooldocs Web page has details of how the JVM locates classes.
There are also some whitepapers and other documents that you can access from the OC4J home page at Oracle's Technet Web site.
Good Luck,
Avi.

Another BW question about filter in the query definition

Hi ,
I have another question about filter in the query definition.
<u>Question:</u>
Which of the following objects can be filtered in the query definition?
A. characteristic
B. Key figure
C. Structure
D. Units
E. Hierarchy
My answer is A,B,D,E.
The answer in the book is A,B,D
Can anybody tell me which one is wrong?
Thanks in advance,
Liu Jia

HI
1.The answer could be A,B,D.you can not apply filters on hierarchies
2.Whenever there is a change in Attribute data(like Location,Num etc...)that has to be reflected in all aggregates of the cube.Thsi can be done by running "ATTRIBUTE CHANGE RUN".
sri

Yet another standby question...

Hi gurus,
Please help to resolve the quiz. I have completed manual physical standby installation. It is up and running. I then decided to install Grid Control and to use Data Guard for the standby creation. I've successfully installed GC, deployed all agents, I can see my previous standbys, life is good in short. Now, when I am trying to create the second standby for my production database using the Data Guard I stuck on a logical question:
On the step 3 of Data Guard standby creation, there is a note, saying:
The instance name (also referred to as the SID) must be unique on the standby host.
Yeah, fair enough. But i can't understand what should I do next? I have:
db_name YELLOW - primary prod
db_name YELLOW - standby prod 1st
and i need db_name YELLOW - standby 2nd
But I can't create it on the same server where the first standby resides.. What options do I have? Do I need to create the second standby on another server? Do I need to kill 1st standby? Is there any way to avoid this conflict? I am confused. I have no equipment for TWO standby servers... I thought I can install there up to 9 standbys.. But it seems to be a problem.. And definitely, I cannot give my second standby other name than YELLOW as all my applications are strongly tied on db_name...
Friends, please clarify this if you can. I am kinda lost in here..
Thanks as usual,
Maria

well, that's a good question! :)
because I don't have spare server for the standby. the question is not about: 'should I or should not put more than one standbys in one server' , but more about: 'is there any way to put two standbys with the identical db_names in one server'. I know that I probably sound unprofessional, but I just wanna know.. If the answer is 'no way', I will kill the first standby then and recreate the second one using DG with the correct name.
Thanks for writing to me,
M.

Another packing question.

Hi,
How do I pack EJB:s that I reside on another server?
I have read Tyler Jewell's excellent articles on this
subject on OnJava, but I still can't get it to work.
My ear contains the following files:
app.jar non ejb classes used by app-ejb.jar
app-util.jar non ejb classes used by app.jar
app-ejb.jar session beans that live in WL
other-ejb.jar session beans that live in another app server
other-util.jar non ejb classses used by other-ejb.jar
The Class-Path: (manifest file) in app.jar points to
app-util.jar
other-ejb.jar
other-util.jar
The Class-Path: (manifest file) in app-ejb.jar points to
app.jar
app-util.jar
From my client i call a session bean in app-ejb on the WL server.
That session bean in it's turn uses classes in app.jar, that
in is't turn call a session bean in other-ejb.jar on the second
app server.
WL complains about not being able to narrow the second session
bean to the correct class.
But, if I put other-ejb.jar and other-util.jar in the CLASSPATH
in the startWL.cmd file it works fine...
BTW, I'm running WL 6.1 sp2, jdk1.3.1_02, WinXP.
Any help would be appreciated.
/ Jörgen Lundberg

public class Clock{
int hours;
boolean isTicking;
public Clock(Clock c){
this.hours=c.hours;
this.isTicking=c.isTicking;
HTH
Chandru.MThis isn't entirely correct code, since the hours and isTicking variables belong to the entire Clock class, rather than one instance of it. This is what you want:public class Clock {
Integer hours;
Boolean isTicking;
public Clock(Clock c) {
    // these statements make them instance variables
    hours instanceof Clock;
    isTicking instanceof Clock;
    hours = c.hours;
    isTicking = c.isTicking;
}Hope this is helpful to you.

Fundamental j2ee classpath question

hi folks --
I have a j2ee project with
- a WAR (which has among other things a servlet in it)
- a ejb project (which has a stateless session bean in it)
I need to call the session bean from the servlet and my servlet code gets a reference to the bean via jndi as follows
...skipping jndi stuff ...
LoadTimerTaskHome loadTimerTaskHome = (LoadTimerTaskHome)PortableRemoteObject.narrow( timerHomeObject, LoadTimerTaskHome.class);
However, at compile time, LoadTimerTaskHome (the home interface of the bean) can't be resolved in the servlet class. Which makes sense cuz how would the servlet class know about the bean class?
And that's the question: At compile time, how can i make classes in my WAR aware of classes in my EJB proj, which are deployed to the same EAR? There must be a very simple J2EE way of doing this...
I'm not an EJB person, so i'm just confused about what is probably extremely basic....
thx VERY much for any insights...

A brute force approach is to also package the ejb interface classes (and any classes needed by those interfaces etc.) in the .war. However, a cleaner solution is to put the shared classes within a utility .jar and package it within the .ear such they are visible by both .war and ejb-jar modules.
In Java EE 5, this is easily done by putting the shared .jar in "lib" directory at the top-level of the .ear. Java EE 5 guarantees that any .jars within that directory are visible to all component modules within the .ear.
Prior to Java EE 5, this was only possible by using the MANIFEST-CLASSPATH attribute within each module to refer to the utility .jar. That's very cumbersome and error-prone so I recommend the automatic "lib" directory approach.
--ken

Re: another build question! (sorry lol)

Hi all
right, after days and days of researching all the excellent articles on here ive had to write a post! So apologies for going over old ground
Im going to build a new CS5.5 rig (having had a guts full of Apple and their FCPX fiasco its back to pc!)
so although i appreciate the 990x o'c is prob best option, bang for buck is leading me down the sandybridge i7 2600k o'c option on p67 mobo with 16gb of ram (option to take to 32 down the line when the chips are out)
anyway, im sorted on chip mobo (msi big bang marshal p67), nivida 570 etc , its the drives im struggling on!! Im edit avchd video and some after effects, small amount of 3d, and in FCP i always transcoded everything to prores. Now on cs5.5 it looks like real time performance is possible with high end hardwear.
So mobo, ram, and chip aside, my HDD config im unsure on, ive not really got the cash to go crazy with raid controllors etc, but understand need for seperate drives, etc and tbh might even go downt he cineform route as a prores alternative (prob avoiding hassles of drive speed with avchd)
im thinking as the SSD's are now dropping in price and the ocz are producing these 500 mb r/w speed sata 6 120gb drives for a reasonable price, would 3 of these drives (one for os, one for media , one for scracth disk ) be a good set up or is it a waste of cash and should i raid 0 from bios/mobo?
I appreicate that 120 gb drives for media etc are small, but i would take project media from another much large backup drive and just use the 3x ssd set up as working disks for editing & os? once project over, clear out drives to larger back up and start new project!
Its either that or i go SSD as bootdrive, but some sort of cheap raid set up for my scratch disk, media drives? prob is if i do that, from what ive read (brain dead now) i would be best off with 2 x raid 0 as scratch disk and orginal media respectively when workign with avchd
HOWEVER, on my mobo there are only 4 sata 6 ports, so if i use one for ssd boot drive, and then im left with 3 x sata 6 ports and another 4 x sata 3 ports to raid on? how does this work? any point in getting the sata 6 drives as one would be stripped with a sata 6 drive plugged into a sata 3 port (this was my reasoning behind using 3 small sata 6 ssd's plugged to the sata 6 ports and rest of sata 3 ports as storage and backups!
confused lol!! I just want the overclocked sandybridge system with decent gpu card, as much ram as possible at present, but im thinking my bottleneck will be in the HDD config! any suggestions are much appreicated! im not that techy so whilst have read all the articles am more confused now (plus normally a mac user, so its out of the box configuration usually!) for what its worth looking to purchase something from scan uk in terms of parts! total build cost including a reasonable screen £2,000
many thanks

You have a limited budget, especially in the UK, but then don't we all?
Going for the 980X will triple the cost of the CPU, but even when editing AVCHD material the gains are not sizable enough IMO to justify that cost differential. Add to that you will need 24 GB instead of 16 GB and that carries an additional price tag. Both factors will easily move you out of budget range if you want to have a number of disks and possibly a raid controller.
Did you read my article To Raid or not to Raid, that is the question. It can be found under the Overview tab at the top of the page? (Currently responding from abroad on my notebook and not having the bookmarks available for easy linking). http://forums.adobe.com/thread/525263
With media and projects I would advise against a raid0, because of the lack of redundancy. For pagefile, media cache and previews (scratch disks) raid0 is quite OK. They will be recreated if needed. The performance gain from a raid0 for media and projects over a parity raid is easily offset by the time spent on making backups. For parity raids do not use WD Caviar Blacks, but look at the Hitachi 7K3000 line of disks.
The question of Sandy Bridge versus the old X58 platform is essentially one of 'which limitations are acceptable to me'.
The Sandy Bridge is a great processor and at least the equal to the old i7-9xx quad cores. However, the platform, the chipset, has its shortcomings in terms of PCIe lanes. Whether that is relevant to you, only you can decide. But hey, we would be in serious trouble if Intel did not manage some progress in two years time from the i7-920 to the i7-2600K. So of course the i7-2600K shows much more potential than the almost retired 920, it is the chipset for the Sandy Bridge that is 'flawed' in comparison to the X58, but that is no surprise, since the Sandy Bridge is a 'middle-of-the-road' platform and the X58 was a 'high performance' platform.
BFTB-wise I think that within your budget limits, you should look at the i7-2600K, but with the best disk setup you can afford.

Apache Commons Net deployment fails to resolve, JDev classpath question

I wrote a wrapper to apache's most recent version of commons-net package (version 1.4.1) to allow me to execute FTP services from an Oracle 9i2 database. After adding the common-net .jar file to the JDev libraries, and including this library in my project, I debugged the code in Jdeveloper, then deployed to the Oracle 9i2 server. However, many classes of the of the common-net (and my code) failed to resolve in the database.
After downloading open source product DependencyFinder, and running against the Commons-net .jar file, I found that some packages were dependent on the Apache ORO regexp package. The ORO package, as far as I can tell, is not in any classpath that my JDev project uses. There is listed in the JDev libraries, a library called Apache Regexp, but I did not include that in the project, either.
After I downloaded the ORO package from Apache, added it to JDev's libraries, included it in my project, and added the ORO library to the list of classes to deploy (using the loadjava deployment profile), deployment to Oracle 9i2 of the my classes, the commons-net classes, and the ORO classes properly resolved.
My question is: Why did my application work under JDeveloper? How and where did it find the ORO classes?
The version of Java under which JDeveloper is running, and the version being used as the compiler for this project, is a new installation of JDK1.4.2_10.

Since the commons net 1.4.1 is compatible with jdk 1.3.1, I should only need to use loadjava on the .jar file. In fact, that works, given that there is no failure on load due to incompatible .jar or .class files structure.
But, I've just solved this part of the problem. The commons net 1.4.1 is dependent on, but does not include the Apache ORO library. Once the ORO library was deployed to Oracle 9i2, all common net 1.4.1 classes were resolved.
The "problem" is with JDeveloper. Without including the ORO library in the JDeveloper application, or in the classpath, it worked under JDeveloper -- the more surprising since I had never downloaded the ORO library from Apache. Only after downloading ORO, explicitly defining ORO to JDeveloper, and including ORO as a library in my JDeveloper project, did the deployment of the application to Oracle 9i2 succeed (though the application always executed successfully under JDeveloper).

Another Noobie question - readLine()

Hi all
This is another one of the extreme noob questions. I wrote a small program to read form a text file
    String NetworkLine = new String();
    try{
        FileReader inputFile = new FileReader("Test.txt"); //
        BufferedReader in = new BufferedReader(inputFile);
       try{
           while (in.readLine() != null){
            NetworkLine = in.readLine();
            System.out.println(NetworkLine);
       }catch (IOException e){
            System.err.println("Error in reading the file");
    }catch (FileNotFoundException e){
        e.printStackTrace();
    }My Test.txt is
One
Two
Three
Four
Five
Six
without the *
The output I get is
Two
Four
Six
I cannot figure out why only even no. lines are being read. I inserted a breakpoint at the line
          System.out.println(NetworkLine);This shows me that only even no. lines are being read by readLine(). Any help will be appreaciated.
Thanks
Ayesha

    String NetworkLine = new String();
    try{
        FileReader inputFile = new FileReader("Test.txt"); //
        BufferedReader in = new BufferedReader(inputFile);
       try{
           while (in.readLine() != null){ // <-- This reads a line, checks it for null ,and drops it
            NetworkLine = in.readLine(); // <-- this reads the next line
            System.out.println(NetworkLine);
       }catch (IOException e){
            System.err.println("Error in reading the file");
    }catch (FileNotFoundException e){
        e.printStackTrace();
    }So the upshot is that you print every other line.
Corrected:
    String networkLine = new String();
    try{
        FileReader inputFile = new FileReader("Test.txt"); //
        BufferedReader in = new BufferedReader(inputFile);
       try{
           while ((networkLine = in.readLine()) != null){
            System.out.println(networkLine);
       }catch (IOException e){
            System.err.println("Error in reading the file");
    }catch (FileNotFoundException e){
        e.printStackTrace();
    }

Another beginner question about targeting functions

I've read what I thought was an excellent thread about scope in EA, but I'm still not getting some of the basics (btw the thread was http://forums.adobe.com/message/5287753#5287753).
I have "slides" in the symbol library and in the main composition compositionReady script I've set up a generic script for playing back sound. That script works fine when I bind it within the compositionReady script to a symbol and click, so I know the funciton itself is fine. My question, I'm programmatically loading the slides from the library as I go along and want to use the generic function I created in the compositionReady script, but I can't seem to create the proper syntax to call the script. Here's some that I have tried. What is the correct syntax? The script is on frame inside the slide symbol.
//sym.getComposition().getStage().playSound();
//sym.getComposition().playSound();
//Edge.getComposition("CER_1").playSound();

Thanks for having a look. Yea, it's getting to the script fine but it doesn't like the syntax. Console always indicates javascript error, which is what happens when Edge throws an exception because it doesn't know what to do. The article that Elaine posted on above gives me the sense that it might be best to change the function to a variable, then it would be in scope to call but I have parameters I'd like to pass with it and I don't think I can pass a parameter with to a variabalized (word?) function. I know it's just me moving from ActionScript to Edge JavaScript and scope. Whereas I could always find a homebase in ActionScript with the stage, it seems that the stage in Edge is just another symbol and calling a function within the stage is never getting there. Elaine hints at putting outside the stage closure and it would be accessible, which, of course, has led me to start readying JavaScript books and learning more about scope etc. It seems like that is life as a developer. I dont' get that, oh, better take a day or two to understand that concept, oh, that function works now, next problem. ;-)

Another excel question

hi,
I read an excel file like this:
CREATE OBJECT gv_excel 'Excel.Application'.
SET PROPERTY OF gv_excel 'VISIBLE' = 0.
CALL METHOD OF gv_excel 'Workbooks' = gv_books.
CALL METHOD OF gv_books 'Open'
    EXPORTING
      #1 = p_file
      #2 = 1.
FREE OBJECT gv_cell.
FREE OBJECT gv_excel.
FREE OBJECT gv_books.
FREE OBJECT gv_sheet.
CALL METHOD OF gv_excel 'QUIT'.
the question is, why the excel process is not kill in the task manager?
joseph

Hi ,
i think Rich is right.
another problem could be if you've used an object to
save your data:
like :
CALL METHOD OF H_EXCEL 'Save' = h_save.
so you've to free this too:
FREE OBJECT H_save.
regards Andreas

Windows 7 Another Preorder Question :)

Ok bestbuy.com has taken out the 108$ deposit and now the order is confirmed so does
bestbuy.com hold the Funds and use them when it ships or can i use the 108$ for something
else now since the order is done and i Can just put another 108$ in there for them to take it
out on 10/22/09 the release date or when it ships
i need to know if i can do this thanks

Well you can purchase either an upgrade or a full copy. Difference is, one has to be upgraded from a previous operating system (XP or Vista) and the other can do a clean install without the need to have a previous operating system. The Windows 7 Upgrade program provides you with an upgrade copy of Windows 7.
Adam
Best Buy Community - Retail, Americas
Forum Guidelines | Terms & Conditions | Community Guidelines | Blogging Guidelines
*Remember to mark your questions solved and click the star under the user's name to show your thanks!

Another classpath question

Similar Messages

Maybe you are looking for