Byte Code Manipulation vs. Preprocessor

Hello,
I'm not sure if I'm asking this in the right forum or not, but if I'm not, please let me know and I'll change it asap.
I�d like to know what the advantages associated with byte code manipulation are.
For instance, if one had the option, why wouldn�t anyone prefer to use traditional pre-processor methods of altering java code as oppose to byte code manipulation (ie. with javassist or ASM)? -- in respects to porting J2ME apps over a variety of cellphones.
Thanks,
Mark

see I was just trying to use BCEL to learn about it. I
want to manipulate the class files. So no point comes
of a different solution.Why nobody is giving me the
solution that I want...Is it b'coz its not possible
with BCEL or what....Please help...
Kindly and please give the replay based on the BCEL.
Thanking you,Initially, you stated that your problem was how to debug production code that may or may not have debugging code present - to which you have been given an answer. If, however, your goal is a purely academic exercise (this has all the earmarks of one, especially since this won't do one bit of good in real life) then you are perhaps misunderstanding what BCEL is. You can't insert System.out.println(...) statements with it, since those are uncompiled java source lines - instead you need to insert compiled bytecode instructions, which will look more like this:
34 invokevirtual #9 <Method java.lang.String getQuery()>
37 invokevirtual #7 <Method void println(java.lang.String)>
40 getstatic #5 <Field java.io.PrintStream out>which I wouldn't think is very useful. I can't suggest strongly enough that you consider not running a binary classfile mutating application on production code - that is a very bad idea.
As I see it, you have two options: compile two different binaries, one for debug and one for release; or use a flexible logging system (which I have done in every case). A third option of course is to continue playing around with BCEL for the purpose of learning the java bytecode world (but if you really want to do that, I suggest you write your own classfile disassembler using the Sun VM spec - it isn't as hard as you think and extremely educational) - but don't inflict a production application with a poor design simply to satisfy a whim.

Similar Messages

Connecting to EMS fails with No mapping for the Unicode character exists in the target multi-byte code page

I am getting the following error when trying to connect to both my exchange servers.
New-PSSession : [ex2013-002.nafa.ca] Connecting to remote server ex2013-002.nafa.ca failed with the following error
message : No mapping for the Unicode character exists in the target multi-byte code page. For more information, see
the about_Remote_Troubleshooting Help topic.
At line:1 char:12
+ $Session = New-PSSession -ConfigurationName Microsoft.Exchange -ConnectionUri ht ...
+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ CategoryInfo : OpenError: (System.Manageme....RemoteRunspace:RemoteRunspace) [New-PSSession], PSRemotin
gTransportException
+ FullyQualifiedErrorId : 1113,PSSessionOpenFailed
EMS used to connect ok. I am not sure if there is any connection but Outlook was installed recently on the exchange server to enable mailbox level backups.
Any help would be appreciated.
Steve Hurst

Hello Steve,
Firstly, you cannot install Outlook with Exchange because they share certain dll files.
About the EMS question, I suggest we try rebuilding the powershell virtual directory. If it still does not work, check the application log for more referernce.
Thanks,
Please remember to mark the replies as answers if they help and unmark them if they provide no help. If you have feedback for TechNet Subscriber Support, contact
[email protected]
Simon Wu
TechNet Community Support

No mapping for the Unicode character exists in the target multi-byte code page

hi,
i have an issue with sharepoint 2013 and IE 10.
im using the sharepoint rest web service and make an ajax data call to retrive data from sharepoint lists, the call fail and return a server error: "No mapping for the Unicode character exists in the target multi-byte code page".
i have to say that everything works fine with chrome and firefox.
what can i do for fixing it?
Thanks a lot
alon

Hi,
From your description, I know you get an issue with IE 10 in SharePoint 2013 when you use SharePoint REST API to retrieve data from SharePoint list.
I am not quite sure what cause your issue. Could you provide your code, so I could test it in my environment and troubleshoot for you.
In addition, you could test your issue in another computer or another version of IE.
Best Regards
Vincent Han
TechNet Community Support

Managed system Configuration : The JMX metrics of this Host / Byte Code Adapter are missing

Hi Friends
While Configuring EP system with SOLMAN 7.1 SP12 it is giving below message in self Diagnosis Phase.
1. Solman system is already having ISAGENT version 8 and 9 installed (but Managed system is having 8)
2. I have followed the automatic Process as below
==========================================
Automated Installation of the Introscope Agent via SMD
The Solution Manager provides an application that performs the setup of the Introscope
byte code agent for Java automatically. This section explains the steps to run the setup.
Before you can run the Introscope agent setup, the setup wizard for the managed system
must have been executed.
1.
2.
3.
4. Launch the Introscope Setup application: SAP Solution Manager Configuration 
Managed Systems Configuration  Step 7 Configure Automatically  Automatic
Activity “Byte Code Adapter Installation”  “Open Java URL” (only available in edit
mode). Your screen will look similar to the image below:
5. Check the Enterprise Manager Settings at the top of the screen. These are the
connection parameters that will be used by the agent to connect to the Enterprise
Manager.
6. In the pane Introscope Agent Setttings, select the system that you want to instrument
with the agent. Check “select all” and click “Retrieve Current Settings”. This will display
the current status of the agent setup.
7. To initially setup or update the agents, click Setup Introscope Agent …. This will open
the setup dialog as shown below. Select the desired agent version (by default only one
available) and choose the profile. Next, check the desired instrumentation areas, and,
in the case of AIX, review the AIX settings to match your environment (J9 or classic
mode).
8. Finally, hit Apply to perform the necessary changes in the agent configuration. Agent
config files will be adapted and the Java VM parameters will be set as required for the
managed system.
================================================================
3. Restarted SAP but error still present.
Please help to resolve the issue.
Thanks

Found the solution :
1819577 - Workload Analysis data missing for Java systems when using ISAGENT 8.2.4.0
Cause
l This issue has been identified as product defect and documented in note 1273028
Resolution
l As mentioned in note 1273028, update to ISAGENT 8.2.4 Patch 1 timestamp: 2012-06-26, please proceed as described below)
1. Enable the Maintenance Mode in the Agent Administration UI (http://solmanserver:port/smd/AgentAdmin)
2. Deploy the file to the Solution Manager JAVA stack using the Software Deployment Manager (SDM)
3. Disable the Maintenance Mode in the Agent Administration
4. Execute again the activity 'Byte Code Adapter Installation' of the step 'Configure Automatically' in the 'Managed System Configuration' for the
affected systems. This activity will deploy the new version of the
Introscope JAVA agent in the managed system servers and reconfigure the JVM parameters to use the new version of Introscope JAVA
Agent;
5. Check from Introscope Agent Admin if "IS Agent State" contains value "Running but needs restart"
6. Restart the affected managed system(s)
Finally Solution is update your SOLMAN system patch ISAGENT 8.2.4.0 from patch 0 to 1 via SDM.
Thanks
Mukesh

Question about Byte Code

Hello...can anyone tell me how I write a java program, compile it and get my byte code written into another file??
Thanks in advance!

I am looking to write my bytecode into a .bc file and
open it and actually be able to read the bytecode
instruction...(stuff like iload_2, i2d etc.,)I see... that was not clear from your first question. What you want is a disassembler. The SDK comes with one called 'javap': http://java.sun.com/j2se/1.4.1/docs/tooldocs/windows/javap.html

Question about compiled byte code

Howdy folks,
I have a client who wants to run an application in both
on-line and off-line modes. The user could run the application
locally on their laptop making changes and such which would get
stored to local database files (probably FoxPro free tables just to
make it easier on me). Then when the user got back to their
internet connection they could run the application and it would
sync with the online tables (probably MySql tables at that point).
So the question is, if I compile Cold Fusion code into Java
byte code, will it be able to execute independantly of the Cold
Fusion Server? I realize that I could load ColdFusion on the user's
laptop, but I don't think I want to do that. I'm assuming that the
answer to my question will be "No. You can't do that. Cold Fusion
isn't meant to work like that." To which my next question would be,
"Well, what language would be best for the type of application I
have described above? Action Script, maybe?"
Any thoughts are welcome, especially if you've written an
application like the one I've described.
Thanks very much,
Chris

Well, rats.
I wrote a nice reply to your message BKBK, but lost it
because, apparently, my session timed out.
The basic jist, was that I've been working on AJAX, and have
been implementing some AJAX-like techniques at some other clients
(using hidden iframes combined with innerHTML -- I know not a
standard, but darn handy otherwise), but I couldn't see how that
would solve my on-line/off-line problem (unless I stuck with the
cookies idea).
I also did some reading on cookies last night (obviously, I
don't use cookies very often if at all in my daily coding), and I'm
a bit put off by the different browser limitations. I'd hate my
client to be chugging along, entering appointments into the
"database" (read: data being stored as cookies to be sync'd later
when the user goes online), and then suddenly run into the cookie
limitation. On top of that, if I'm reading right, IE (my client's
most likely choise of browser), will not let you know that you've
reached this limit, but will just begin dropping the older cookies
in favor of the newer ones. If I could programmatically sense this
limitation and then write the cookies to some file before
continuing that'd be geat, but since JavaScript can't write files
(that I know of) this isn't feasable. Also, if I could write a file
like that, I wouldn't bother with the cookies.
I think I'm going to end up writing it in FoxPro since my
company has a bunch of copies of it (and it's licenced per
developer and not per copy), and there are lots of folks in my
company who can help me get up to speed. That also means that I'll
probably need to write a web version of the code for when my
client's client's (does that make sense? :-) ) connect to the app
via the internet.
Anyway, I'm really enjoying everyones comments on the
subject. Can anybody think of a technique for a way around the
cookie limitations? Or perhaps another language that this whole
thing could be written in?
I really wish that I could compile my ColdFusion code for use
independant of the CF server. I know, that's not the way it works
and typically not what scripting languages like this are used for.
I suppose I could always install the developer's version of CF on
the user's local machine, write the code in CF and then just detect
whether or not the user is online and behave accordingly.

Inheritance and byte-code...

My coworker and I were having a discussion today that I imagine someone here will be able to resolve for us: we were talking about how wonderful inheritance is (when used properly, of course), and the issue of the compiled byte-code came up.
Basically, it was my conclusion that a subclass contains only the byte code that differentiates it from the superclass; in other words, if a superclass is modified and recompiled, there is no need to recompile the subclass for the changes to take effect. My coworker disagreed, however, arguing that there would be linking problems between the two classes if the subclass was not recompiled (sign of an old assembly programmer, if you ask me).
Anyone know the answer to this?

I created a trivial Parent class and extended it with an
even more trivial Child class. I compiled both classes
and then made a change to the Parent class. After
compiling the Parent class only, the change took effect
in the Child class. This indicates that you are right and
your co-worker is wrong. Here are the classes if you
would like to show him.
import java.awt.*;
public class Parent
     Dimension d;
     Parent() {
          d = new Dimension(333, 777);
     public Dimension getDimension() { return d; }
import java.awt.*;
public class Child extends Parent
     Dimension d;
     public Child() {
          d = getDimension();
          System.out.println(d.height + " " + d.width);
     public static void main(String[] args) {
          new Child();
}Mark

Is there a way to convert JVM 1.1x byte code to JVM 1.3+ byte code

I dont have the source code from a program that is used in my company. The original writers went out of business... so we cant ask them.
I was thinking of decompiling the code and then recompiling it... but I was wondering if there wasnt just a byte code updater?
Or if anyone could come up with a better method.
-S. Bob.

er.. well I dont think you understand. I have been assigned to get it functional... I mean, its my task to come up with a solution. :[
Unless of course you have a credible documented source that states 'NS 6.1 is a hot steaming pile of crap' and that no one should use it. I could leverage off of that to help my claim that any support for this browser is pointless.
Do you have a link that talks about the applets and why they dont function in NS 6.1 vs 4.7x ?
That or does anyone know why or how to get NS 7.0 working on Windows 95. (yes I do understand what I ask here... I wish it werent true.)
-S. Bob
'life in the corporate wheel, grinding one day at a time'

Strange Error - 5005: Unknown error optimizing byte code.

Hello flashcoders,
I am facing strange problem since long time. This error code
even doesn't exist in the list of error codes.
This is the exact error I am getting while I compile the FLA
from flash CS3.
Location : , Line 1
Description : 5005: Unknown error optimizing byte code.
Source :
I don't know exact reason about its generation, But here are
the some possible reasons / hints:
Case 1) Overloading:
1.1) Size of .fla is 10.5 MB and its document class contains
more than 60 classes to import and has more than 100 variables.
1.2) Even if I put In document class - only variable
initialization and class importing are there. Nothing in its
constructor + no other functions are defined. Still error is there.
1.3) If we import all classes and has all variables then it
gives this compile error. But if we remove some particular numbers
of variable, it's start working. In this we can remove any type of
variables.
1.4) After reducing variables, application starts working
till that it won't.
Case 2)
2.1) Size of .fla is 1.75 MB and its document class is same
as above one.
2.2) All assumptions are same as above.
2.3) Now this class contains all functions and have
initialization of all variables + classes.
2.4) In this If we remove 3-5 variables, it will start
functioning else it won't.
Its a huge application so I am even confused that what is the
cause of error and this error stopped our working for a week now.
Bit more information about the project that may help the team to
identify the reason.
1. Project development started with flash public alpha 3.
When we started using Flash CS3, we had some design problem if we
do open the FLA in CS3, so we completely redesigned the Movieclips
etc., in Flash
CS3 IDE.
2. Project contains approx 250 classes.
3. In main application, it imports 67 classes. (it works if I
keep 63 classes in document class).
4. In the case 3 above, if that works with 63 classes and If
I do add 3 frames in existing movieclip, it stops working.
5. In the case 3 above, if that works with 63 classes and If
I do add / declare few more variables, it stops working.
It would be humble appreciation if someone can come up with
some light in the dark tunnel.
Best Regards,
Ashvin Savani - "arckid"
Founder & CTO - Avinashi.com
Adobe Community Expert
We Never Give Up!

I've posted an article on this problem -
http://www.negush.net/blog/5005-unknown-error-optimizing-byte-code/
- and here are a few ideas on how to handle it (check out the
comments):
- try turning off the optimizer
- delete ASO files
- also it seems that changing the java vm heap memory size
could help (find he environment variables in the Windows computer
properties and search the JAVA_TOOL_OPTIONS variable)

Byte Code Instrumentation with -javaagent in JCAPS 5.1

Hello All,
Has anyone successfully used Byte Code Instrumentation using the -javaagent switch in JCAPS 5.1?
I have 3 instances of JCAPS and two are reporting metrics fine the third is Not!
I am trying to instrument but my agent code is not being started and I am not seeing metrics...

Tags in JCAPS Repository are primarily used with deployment profiles. They are not the traditional CVS or SVN Tags. As far as I know, you cannot export projects based on Tags. The only place they are useful is when you are creating deployment profiles and want to use an older Tagged version of the connectivity maps etc and the components used within them such as JCDs, eWays, BPs etc. The caveat here is that if you used any external JARs within your JCDs, earlier versions of those would not be references corresponding to the repository Tag.
Hope that helps.
Z.

Is Java Byte Code is 100% platform independent

Hi all
A my room fellow is disturbing me about the platform independency of JAVA.
Now i only wan to know that generated Byte Code will be same if it is generated on different platforms (Window,Linux,Unix etc)using their respected JVM
Thanks in advance

The "Platform Independency" of Java is a bit of a joke
to me. You need a JVM that is platform dependent to
run it. Now granted, you can write one application
and itll run on any platform, but that's only assuming
they have the JVM installed on them. Not to say that
Java isn't cool and all, but the "platform
independency" solution still has a reliance on
something that is platform dependent... It's like
driving an automatic and saying you can drive....Sure,
you are making the car go, but something is doing the
shifting for you....(C: I guess I'm just arguing for
the sake of arguing...Well if you want to argue that point then no language is platform independant. C code for instance needs a compiler for the platform you want to compile it to run on and that compiled code won't work on any other platform. Ultimately every program has something that ties it to a particular platform, that's just the nature of things in computers.
The idea of platform independance was the removal of the need to re-compile a program with a native compiler just to get it to run on that platform.
Instead Sun opted to have a compact JVM to interpert the byte code and have the byte code standard across all implementations of the JVM.

Executing tiny programming language compiled byte code with the JVM

Hello,
For a project, I am developing a simple programming language, with a syntax not unlike Pascal, which compiles to Java Byte Code. Currently I execute the compiled byte code using my own interpreter. This works OK, but what I'd like to do now is to execute the compiled byte code using the actual existing JVM. The compiled byte code resides in a text file upon compilation of a sample program written in my tiny programming language. Does anyone know how I may start to go about this? Thanks in advance.
adam

The only way of getting a standard Java virtual machine to execute your java byte code, is to wrap it up into a completely valid java class file.
Any constants you use, must be put into the classes constant pool.
Any local variables will have to go into the local variable table of the method which uses them... etc, etc.
Something like ASM, would let you dynamically create a Java class file.
http://asm.objectweb.org/
regards,
Owen

Problem-Loading Byte code from DB

Dear all,
have any one tried to load the bytecode read from the database. If so, please suggest me a way to do it. Iam developing a frame-work where nothing except a servlet would reside on the disk as a file. after designing the frame work, i've encountered this problem...plz suggest me a way.

Yes! You are correct. My Intention is Exactly the Same. If The byte code is residing in a database, with the classname as key, i would load all the class tree from the DB and execute the program....Please Help me
Thanks
Sasi

JDK 64k byte code limitation for methods

Concerning the JDK 64k byte code limitation for methods (actually 65535 bytes): While the Java compiler doesn�t care for that problem, the Java VM does. Hence, an violation of the 64k limit is dedected at run time, too late, of cource. Is it possible to detect the violation at (or immetiately after) compile time? Are there any tools that support the perception of the violations as early as possible?

Thanks to the authors of the answers. I agree to you as far as hand-written code is concerned. However, we generate our java classes, thousands of files, each java class with a lot of methods. We do have rules to keep things as smal as possible. Nevertheless, some methods become very large. Thus the problem is to detect such methods. Once the methods are known, we are able to modify our generator. We implemented a small program that looks for large methods in terms of source code, but there is just a week correlation between source code and byte code. Hence, again the question: Is there a tool that supports the detection of methods larger than 64k byte code before installation?

Can we invoke a method using BCEL(byte code engineering library)

Hi i am new in Java bcel,So can anyone guide me in this problem.
I am searching for some techniques in BCEL to invoke a method of some class just like we use to invoke methods using Java Reflection.I have also tried a Java Assembler named "Jasmin" that used to interact with Operand Stack,Local variable array & Constant Pool etc so i guess this is the same containers with which our BCEL also interact with so then i am wondering why not then there exists a way through which we can invoke a method using BCEL IF this will happen in any case then kindly point it out some direction of help.
thanks in advance

856989 wrote:
I think i have clearly mentioned that "Tell me any way in doing this invocation in BCEL not in Reflection."
Why i not want to use reflection because it have "invocation overhead"Wrong.
If I dynamically load a class and reference it via a interface there is absolutely no difference in that and if I referenced the class statically and used it via an interface. The load and use process is exactly the same.
And the first uses "reflection".
If you use a proxy interface via something like java.lang.reflect.Method
then of course there is some overhead because there is in fact more classes involved. Just as if you had any other layer in any other code.
And even so unless you have actually profiled the application and found an actual bottleneck (and not just a measured impact) then even looking at it is a waste of time.
If it was me I would be far more concerned that inserting byte codes was correct for all cases.

Byte Code Manipulation vs. Preprocessor

Similar Messages

Maybe you are looking for