RE: compile error to source links
At 03:30 PM 7/12/01 -0400, Javier Lopez wrote: Where is compile.el. I checked the emacs lisp directory and xemacs-packages directory. I could not find it. I am using XEmacs 21 In the distribution of Emacs I am using is in emacs-root/lisp/progmodes/compile.el Depending on how emacs was installed, there may only be compile.elc and not compile.el. The .elc file is a compiled version of the .el file, and not human-readable. In that case, you would have to find an emacs distribution that includes the lisp sources. Troy Javier
Re: A func to insert Timer code?
I think a better template to use is the following:
// Original version
int getValue() {
return super.getValue() * 5;
}
// Instrumented version
int getValue() {
try {
Timer.startTimer(Test.getValue());
return super.getValue() * 5;
} finally {
Timer.stopTimer(Test.getValue());
}
}
This also handles correctly cases where there are multiple return
statements, and you get timing information even if there's an
exception. The implementation is fairly simple:
(beginning-of-defun)
(insert try {\n Timer.startTime( (get-function-name) );\n)
(end-of-defun)
(insert } finally {\n Timer.stopTimer( (get-function-name) );\n}\n)
This isn't quite right. At least on my emacs, the defun functions put you
on the wrong side of the braces. get-function-name is probably also not
the right function name, but there is some function that does that. (The
name shows up in my modeline, so probably semantic/senator knows it.)
Troy
At 09:02 PM 1/24/02 +0100, Paul Ebermann wrote:
Altmann, Michael skribis:
I have been using xemacs and jde for quite a while, but I am not much of an
elisp programmer. Could someone help me (or point me to the appropriate
online resource) write the following elisp function? I have a java class
called Timer that I use to time various chunks of code. It has methods
startTimer and stopTimer that should surround any code I want to time. I
would like an elisp function to insert calls to Timer for the current java
method. For example, given a java program that looks like
[...]
I would like code inserted like
[...]
void bar() {
Timer.startTimer(Test.bar);
int x=3;
Timer.stopTimer(Test.bar);
}
}
(Sorry for missing tabs, my newsreader ignores or deletes them.)
What should be done here?
---
int getValue() {
return super.getValue()*5;
}
---
The simple Version
---
int getValue() {
Timer.startTimer(Test.getValue());
Timer.stopTimer(Test.getValue());
return super.getValue() * 5;
}
---
is not really helpful (it does not time
anything time-consuming).
You may need something like
---
int getValue() {
Timer.startTimer(Test.getValue());
int temp = super.getValue() * 5;
Timer.stopTimer(Test.getValue());
return temp;
}
---
So your function has to choose a name for a
temporary variable, and using the right type ...
Sorry, this does not really help, but only
shows a possible problem.
Paul
Troy Daniels
[EMAIL PROTECTED]
781-273-3388 x218
Byte-compiling (Was RE: JDE config problem?)
At 09:07 AM 2/12/02 -0500, Paul Kinnucan wrote: Berndl, Klaus writes: I've run the JDE with eieio both compiled and uncompiled and I've never noticed any difference in performance. Further, I can't imagine why it should cause bad performance. eieio is not used to do any computationally intensive tasks. On a related but orthogonal note, do you have an idea how byte-compiling would affect memory usage? Most (all?) of the time when a JDE command is slow, I'm fairly certain that it's because I'm low on memory and the program is swapping. (Compiling with the compile server often takes less than a second, but can take more than 15.) Troy First of all not only JDE uses eieio but also semantic which in turn is used by JDE. Secondly macros defsubst functions are inlined if byte-compiled and eieio has a lot of them (if i remember right, Eric has told me this). BTW: This argument would also recommend, that we should semantic byte-compile! But here comes the original explanation from Eric, the author of eieio: Byte compilation converts the lisp program into a stack based program with integer IDs for all built-in symbols. The core of EIEIO uses almost all built in commands (like car, cdr, assoc, etc). In particular, the method dispatch mechanism does a lot of work. When EIEIO core is not byte compiled, the method dispatch runs between every single EIEIO method call. The byte compilation step speeds this up a lot. Therefore my strong recommendation: At least eieio and semantic should be byte-compiled. And for other packages it is IMHO at least a good way to detect some errors... If you have noticed bad performance, I'd like to know under what circumstances so that I can investigate. There may be ways to improve the performance. Yes, just byte-compile eieio and semantic :-) Klaus, I don't doubt that byte-compiling eieio improves performance. What I doubt is that byte-compiling eieo improves JDE performance perceptibly. The JDE does not use eieio for any computation-intensive tasks. A few microseconds of difference in the execution of a debugger step command or the creation of a dialog buffer is not going to be perceptible to a human being with ordinary powers of perception. My experience running the JDE with eieio both compiled and uncompiled confirms this. On the other hand, incorrectly compiling eieo does make a PERCEPTIBLE difference in the response time of JDEE commands. It increases the response time to INFINITY and forces me to repeatedly having to tutor new users on the correct way to byte-compile sets of interdependent Lisp files. This takes up a significant portion of my time, which is why I try to discourage people, especially novices, from byte compiling the JDE. The only way you are going to convince me to change this policy is to cite a SPECIFIC JDE command that is PERCEPTIBLY improved by byte-compiling eieio. As for semantic, compiling semantic does shorten loading of large Java files but otherwise does not make a significant enough difference, in my opinion, to warrant encouraging new users to compile the JDE. I've used a lot of Emacs applications and the only one that I have used where compilation is absolutely necessary is psgml because compiling psgmal reduces parsing of the Docbook DTD from hours to a few seconds. - Paul Troy Daniels [EMAIL PROTECTED] 781-273-3388 x218
Re: Why JDEE ?
At 01:30 PM 2/25/02 -0500, Paul Kinnucan wrote: John Cobo writes: Paul, Am I correct in thinking that your JDEE allows a single set up of the JDK, Emacs, Tomcat, Web server, any shared JAR, WAR files for the development team, etc. on a shared server. Developers can then 'simply' access the environment through a window on to the server ? Do you mean use a shell window running on a PC to open an instance of Emacs on a Unix server and then interact with the Emacs running on Unix through the shell window running on the PC? Although I've never tried working this way, I'm pretty sure it would break some of the JDEE's features. A better bet would be to run X Servers on the PCs that would allow the PC users to start instances of Emacs on the Unix server that could then display on the PC in a full window that supports all JDEE features. You could use the server to maintain a shared source code repository from which developers created their own sandboxes on their PCs. At my company, the Unix disks can be mounted on the PC using samba, so that it looks like a network drive. I then run emacs on my PC, and can read and compile the files on the Unix disk. Actually running the code requires a window on the Unix machine*. This may work better or worse, depending on (at least) the availability of samba, and whether the net lag is worse for interacting with emacs or for writting the class files. Troy * You could probably configure jde to execute an rsh command to run the code. Troy Daniels [EMAIL PROTECTED] 781-273-3388 x218
Re: Why JDEE ?
At 08:56 PM 2/25/02 +, Phillip Lord wrote: Troy == Troy Daniels [EMAIL PROTECTED] writes: Troy At 01:30 PM 2/25/02 -0500, Paul Kinnucan wrote: John Cobo writes: Paul, Am I correct in thinking that your JDEE allows a single set up of the JDK, Emacs, Tomcat, Web server, any shared JAR, WAR files for the development team, etc. on a shared server. Developers can then 'simply' access the environment through a window on to the server ? Troy At my company, the Unix disks can be mounted on the PC using Troy samba, so that it looks like a network drive. I then run Troy emacs on my PC, and can read and compile the files on the Unix Troy disk. Actually running the code requires a window on the Unix Troy machine*. This may work better or worse, depending on (at Troy least) the availability of samba, and whether the net lag is Troy worse for interacting with emacs or for writting the class Troy files. Another possibility would be to have a go with TRAMP, which is ange-ftp for the new millenium. I use it across ssh, which works very well. I've not tried it on windows though, and I don't know how much JDE functionality it would break. I'd be interested to know. I expect that it breaks compiling and running. Assuming it uses ange-ftp-style names, c:\jdk1.3.1\javac \\user@host:\dir\file.java is unlikely to succeed. :-) Troy Phil Troy Daniels [EMAIL PROTECTED] 781-273-3388 x218
JDEbug and LinkedList
Hello,
There appears to be a bug in JDEbug regarding circular references. I was
trying to look at a java.util.LinkedList. The implementation in my JDK
uses a class like the following:
class LLNode {
Object element;
LLNode next;
LLNode previous;
}
where next and previous point to the next and previous elements of the
list. Looking at the first element was fine. I clicked on the reference
to the first LLNode, and expanded element. When I tried to look at the
second element by expanding next, the mechanism to presever the expanded
state imploded. It expanded the LLNode for next, then next.previous, then
next.previous.next, and so on, looping between those two two objects until
it hit some iteration limit.
As a result, I couldn't look at any local variables. It apparently even
failed to set up the properties so that I could close the objects. This
persisted through exiting the debugger and starting a new run.
In general, the mechanism is good and user full, but is there a way to turn
if off in cases like this?
Troy
Troy Daniels
[EMAIL PROTECTED]
781-273-3388 x218
Re: cc-mode indenting
At 11:42 AM 2/6/2003 -0500, Le Wang wrote: Hi, This is the current indentation style: code gmo.options = MQC.MQGMO_WAIT | MQC.MQGMO_FAIL_IF_QUIESCING | MQC.MQGMO_SYNCPOINT | MQC.MQGMO_CONVERT; /code I would like: code gmo.options = MQC.MQGMO_WAIT | MQC.MQGMO_FAIL_IF_QUIESCING | MQC.MQGMO_SYNCPOINT | MQC.MQGMO_CONVERT; /code What I do in this case is rewrite it as gmo.options = (MQC.MQGMO_WAIT | MQC.MQGMO_FAIL_IF_QUIESCING | MQC.MQGMO_SYNCPOINT | MQC.MQGMO_CONVERT); The parentheses change the context, so that it formats it like you want. Probably not the best solution, but it's one that works. Troy Troy Daniels [EMAIL PROTECTED] 781-273-3388 x218
Excluding specific jars from classpath
While answering a question about classpaths, At 10:51 AM 4/3/2004 +0200, Petter Måhlén wrote: What you should do is something like this: 1. M-x customize-variable RET jde-global-classpath RET 2. Add the directories/jars that you want, one in each list entry (note that if you add a directory, JDE will pick up any jars in that directory automatically) Is there a way to exclude a particular jar when JDE does this? In our project, we have a jars directory that contains all of the jars we use. Our build script also creates a jar here with all the classes from our project. I'd like to exclude that jar for several reasons. 1. Some of the classes in it might be out of date if I've compiled a file in JDE without rerunning out build script. 2. If the beanshell is running and the project jar is in it's classpath, our build script fails because it's unable to delete the jar file, since it's in use. This results in me repeatedly exiting and restarting the beanshell. If this can't be done, I'll try to modify the appropriate function to add it as an option. Troy Troy Daniels [EMAIL PROTECTED] 781-273-3388 x218
Re: Error when running an application within JDE
At 01:09 PM 11/10/2004 -0500, Paul Kinnucan wrote: Troy Daniels writes: To: [EMAIL PROTECTED] Subject: --text follows this line-- Please enter the details of your bug report here Hi Troy, It's possible that win32-start-process-show-window was not defined when jde-run.elc was compiled. I'll investigate. Meanwhile, workarounds are to run jde-run.el or try using the jde-compile-jde command to compile the JDEE. That should ensure that the macro is defined when jde-run.el is compiled. I had to delete all the .elc files first, but jde-compile-jde worked. To avoid this problem, you or Arthur might want to change line 343-345 from cd ../$JDE_FILE/lisp $EMACS --no-site-file -batch -l ../../compile-script -f \ batch-byte-compile *.el emacs_c.log to cd ../$JDE_FILE/lisp $EMACS --no-site-file -batch -l ../../compile-script -f \ --execute (require 'jde) -f jde-compile-jde emacs_c.log Looking at the output, I think there might be a bug in jde-run-etrace-show-at-mouse. The byte-compiler complains that Compiling file d:/Programs/emacs-21.2/site-lisp/jde/lisp/jde-run.el at Wed Nov 10 13:55:48 2004 While compiling jde-run-etrace-show-at-mouse: ** reference to free variable pos The code for that function is (defun jde-run-etrace-show-at-mouse (event) Jump to the stack position at the mouse click. Click anywhere on the line with the stack reference. (interactive e) (if jde-xemacsp (set-marker (car jde-run-etrace-current-marker) (event-point pos) (window-buffer (event-window pos))) (let ((pos (event-start event))) (set-marker (car jde-run-etrace-current-marker) (posn-point pos) (window-buffer (posn-window pos) (jde-run-etrace-goto)) Note that if jde-xemacsp, then pos is used but not defined. I think you want this instead. Unless you use the Xemacs, you won't notice the problem. (defun jde-run-etrace-show-at-mouse (event) Jump to the stack position at the mouse click. Click anywhere on the line with the stack reference. (interactive e) (let ((pos (event-start event))) ;; this line moved here (if jde-xemacsp (set-marker (car jde-run-etrace-current-marker) (event-point pos) (window-buffer (event-window pos))) (set-marker (car jde-run-etrace-current-marker) (posn-point pos) (window-buffer (posn-window pos) (jde-run-etrace-goto)) Troy Troy Daniels [EMAIL PROTECTED] 781-273-3388 x218
Synchronization issues in JDEbug
Hello, It looks like JDEbug needs to have some methods synchronized. I tried setting a breakpoint while the process was running, and got the following in the Message window: Setting breakpoint at line 85 in d:/jaguar_te1.2/KBR_ISD_VOB/Jaguar_Plan_Generator_Component/source/PhysicalNetworkBuilder/com/alphatech/kbr/jaguar/networkbuilder/SpatialSpec.java. Error: evaluating debugger output caused a Lisp error. See *messages* buffer for details. Error: evaluating output from the debugger caused a Lisp error. Debugger output: it.framework.TestSuite TestSuite.java 208 runTest) (list 25 junit.framework.TestSuite TestSuite.java 203 run). Lisp error: (void-variable it.framework.TestSuite) Breakpoint set at line 85 in class d:/jaguar_te1.2/KBR_ISD_VOB/Jaguar_Plan_Generator_Component/source/PhysicalNetworkBuilder/com/alphatech/kbr/jaguar/networkbuilder/SpatialSpec.java. It also appears that, at least with JDK 1.4.2, there is a lot of useless traffic going from Java to emacs. My *JDEBug* buffer is full of output that looks like (jde-dbo-event-set 1 none (list Thread 1 main runnable suspended by debugger (list stack traces of several threads ))) Since jde-dbo-event-set is a no-op if the optional fourth arg is omitted, this slows down the debugger considerably. If the java side checked for the fourth argument, is could avoid a lot of toString calls, bandwidth to Emacs and parsing of the string by Emacs. Troy Troy Daniels [EMAIL PROTECTED] 781-273-3388 x218
Re[3]: jde and emacs memory usage
At 11:02 PM 11/18/2004 -0500, Eric M. Ludlam wrote: Raul Acevedo [EMAIL PROTECTED] seems to think that: Wow, that chart thing is pretty cool. :) There are several random chart functions, plus a chart menu item in the Senator menu. Is there a reason chart-space-usage explicitly calls csh with call-process-region? My cygwin installation doesn't have csh, only bash and sh. At the very least, you should use shell-file-name. Troy Troy Daniels [EMAIL PROTECTED] 781-273-3388 x218
Problems using jde-ant-build with spaces in ant home
Please enter the details of your bug report here
I'm having problems running ant with jde-ant-build. If I set
jde-ant-invocation-method to Script, it works fine. The head of the
compilation buffer is below. Note that the value of ant.home is quoted.
cd
d:/Jaguar/dev/logging_reorg/KBR_ISD_VOB/Jaguar_Plan_Generator_Component/source/Utilities/com/alphatech/kbr/jaguar/logging/
ant -Dant.home=d:/Program Files/apache-ant-1.6.2 -buildfile
'd:/Jaguar/dev/logging_reorg/KBR_ISD_VOB/Jaguar_Plan_Generator_Component/source/Utilities/build.xml'
-emacs
If I set it to Ant Server, I get the following in the compilation
buffer. Note that ant.home is not quoted. Looking at the lisp files,
it appears that quotes are first added then removed.
AntServer output:
-Dant.home=d:/Program Files/apache-ant-1.6.2 -buildfile
d:/tdaniels_view/EBO/build.xml -emacs
Buildfile: d:\tdaniels_view\EBO\build.xml
BUILD FAILED
Target `Files/apache-ant-1.6.2' does not exist in this project.
Lastly, if I set jde-ant-invocation-method to Java, I get this.
Here, we have a mixture of single and double quotes. Somehow, the
double quote and space between ant.home's value and the Main class is
interpreted as a space but not a break between paraemeters.2
c:/j2sdk1.4.1_03/bin/java.exe -classpath 'd:/Program
Files/apache-ant-1.6.2/lib;c:/j2sdk1.4.1_03/lib/tools.jar'
-Dant.home=d:/Program Files/apache-ant-1.6.2 org.apache.tools.ant.Main
-buildfile
'd:/Jaguar/dev/logging_reorg/KBR_ISD_VOB/Jaguar_Plan_Generator_Component/source/Utilities/build.xml'
-emacs
java.lang.NoClassDefFoundError: Files/apache-ant-1/6/2
org/apache/tools/ant/Main
If it matters, I'm running with the CVS version, not the beta listed
below. I launch emacs from a cygwin bash shell.
Emacs : GNU Emacs 21.2.1 (i386-msvc-nt5.0.2195)
of 2002-03-19 on buffy
Package: JDE version 2.3.4beta6
Required packages: cedet-1.0beta2
current state:
==
(setq
jde-gen-session-bean-template '((jde-import-insert-imports-into-buffer
(list \javax.ejb.*\
\java.rmi.RemoteException\)) '
(jde-wiz-update-implements-clause \SessionBean\) '
(jde-gen-method-signature \public\ \void\ \ejbActivate\
nil \RemoteException\ ) ' (if jde-gen-kr () 'n)
\{\''n \}\''n 'n (jde-gen-method-signature \public\
\void\ \ejbPassivate\ nil \RemoteException\ ) '
(if jde-gen-kr () 'n) \{\''n \}\''n 'n
(jde-gen-method-signature \public\ \void\ \ejbRemove\
nil \RemoteException\ ) ' (if jde-gen-kr () 'n)
\{\''n \}\''n 'n (jde-gen-method-signature \public\
\void\ \setSessionContext\ \SessionContext ctx\
\RemoteException\ ) ' (if jde-gen-kr () 'n)
\{\''n \}\''n 'n (jde-gen-method-signature \public\
\void\ \unsetSessionContext\ nil \RemoteException\
) ' (if jde-gen-kr () 'n) \{\''n \}\''n 'n ')
jde-gen-beep '((end-of-line) '
\Toolkit.getDefaultToolkit().beep();\''n')
jde-complete-signature-display '(Eldoc)
jde-project-name default
jde-wiz-toString-code-EOL '(Unix)
jde-which-method-format '([ jde-which-method-current ])
jde-run-classic-mode-vm nil
jde-complete-unique-method-names nil
jde-find-granularity '(Character)
jde-which-method-max-length 20
jde-javadoc-gen-nodeprecatedlist nil
jde-imenu-include-classdef t
jde-javadoc-gen-link-online nil
jde-complete-display-result-type t
jde-gen-code-templates '((Get Set Pair . jde-gen-get-set)
(main method . jde-gen-main-method)
(toString method . jde-gen-to-string-method)
(Deep clone . jde-gen-deep-clone)
(Action Listener . jde-gen-action-listener)
(Change Listener . jde-gen-change-listener)
(Window Listener . jde-gen-window-listener)
(Mouse Listener . jde-gen-mouse-listener)
(Mouse Motion Listener .
jde-gen-mouse-motion-listener)
(Inner Class . jde-gen-inner-class)
(println . jde-gen-println)
(beep . jde-gen-beep)
(property change support .
jde-gen-property-change-support)
(EJB Entity Bean . jde-gen-entity-bean)
(EJB Session Bean . jde-gen-session-bean))
jde-gen-cflow-else '(' \else\
(jde-gen-electric-brace jde-gen-conditional-padding-3)
'r'n \}\ (if jde-gen-comments
'(l \ // end of else\)) ''n)
jde-jdk-registry '((1.3.1 . d:/Program Files/jdk1.3.1_06)
(1.4.1 . c:/j2sdk1.4.1_03)
(1.4.2 . c:/j2sdk1.4.2_04))
jde-javadoc-gen-destination-directory JavaDoc
jde-mode-line-format '(- mode-line-mule-info mode-line-modified
mode-line-frame-identification
mode-line-buffer-identification global-mode-string
%[( mode-name mode-line-process minor-mode-alist
%n )%]-- (line-number-mode L%l--)
(column-number-mode C%c--) (-3 . %p)
(jde-which-method-mode
(-- jde-which-method-format --))
-%-)
jde-mode-abbreviations '((ab . abstract) (bo . boolean)
(br . break) (by . byte) (byv . byvalue)
(cas . cast) (ca . catch) (ch . char)
(cl . class) (co . const)
(con . continue) (de . default)
(dou . double)
Increasing memory for JDEbug
Hello, Is it possible to allocate more memory for the debugger process? I'm running into a problem where jde.debugger.Main is running out of memory. One of my local variables is a very long string (about 10 Mb) and the debugger process runs out of memory when it tries to display it. Troy Troy Daniels [EMAIL PROTECTED] 781-273-3388 x218
Re: Increasing memory for JDEbug
At 03:59 PM 1/18/2005, Paul Kinnucan wrote: Troy Daniels writes: Hello, Is it possible to allocate more memory for the debugger process? I'm running into a problem where jde.debugger.Main is running out of memory. One of my local variables is a very long string (about 10 Mb) and the debugger process runs out of memory when it tries to display it. Hi Troy, Try increasing jde-db-option-heap-size. If this doesn't work, let me know. The maximum is already set at 1000 Mb. Looking at the *JDEbug* buffer, I don't think that's the right option for the problem I'm having. A partial excerpt of the *JDEbug* buffer: cd d:/dev/logging_reorg_import/KBR_ISD_VOB/Jaguar_Plan_Generator_Component/products/PGTest/test c:/j2sdk1.4.2_04/bin/javaw.exe -classpath d:/Programs/emacs-21.2/site-lisp/jde/java/lib/jde.jar;c:/j2sdk1.4.2_04/lib/tools.jar jde.debugger.Main JDE -1 1 launch 1 -vmexec javaw -classpath deleted -Xms400m -Xmx1000m junit.textui.TestRunner com.alphatech.kbr.jaguar.pgtest.test.NightlyTests I don't need to change the values in the launch command, I want to add them to the javaw ... jde.debugger.Main command. The program creates a StringBuffer for a 10 Mb string. That works fine. When I step a few lines further and stringBuffer.toString is assigned to a local String variable, I get an java.lang.OutOfMemory message in the JDEbug buffer and the local variable window starts to act funky. (It sometimes leaves portions drawn in solid grey.) Troy Paul Troy Troy Daniels [EMAIL PROTECTED] 781-273-3388 x218 Troy Daniels [EMAIL PROTECTED] 781-273-3388 x218
Re: Increasing memory for JDEbug
At 04:30 PM 1/18/2005, Paul Kinnucan wrote: Troy Daniels writes: At 03:59 PM 1/18/2005, Paul Kinnucan wrote: Troy Daniels writes: Hello, Is it possible to allocate more memory for the debugger process? I'm running into a problem where jde.debugger.Main is running out of memory. One of my local variables is a very long string (about 10 Mb) and the debugger process runs out of memory when it tries to display it. Hi Troy, Try increasing jde-db-option-heap-size. If this doesn't work, let me know. The maximum is already set at 1000 Mb. Looking at the *JDEbug* buffer, I don't think that's the right option for the problem I'm having. A partial excerpt of the *JDEbug* buffer: cd d:/dev/logging_reorg_import/KBR_ISD_VOB/Jaguar_Plan_Generator_Component/products/PGTest/test c:/j2sdk1.4.2_04/bin/javaw.exe -classpath d:/Programs/emacs-21.2/site-lisp/jde/java/lib/jde.jar;c:/j2sdk1.4.2_04/lib/tools.jar jde.debugger.Main JDE -1 1 launch 1 -vmexec javaw -classpath deleted -Xms400m -Xmx1000m junit.textui.TestRunner com.alphatech.kbr.jaguar.pgtest.test.NightlyTests I don't need to change the values in the launch command, I want to add them to the javaw ... jde.debugger.Main command. The program creates a StringBuffer for a 10 Mb string. That works fine. When I step a few lines further and stringBuffer.toString is assigned to a local String variable, I get an java.lang.OutOfMemory message in the JDEbug buffer and the local variable window starts to act funky. (It sometimes leaves portions drawn in solid grey.) Hi Troy, Try setting jde-db-option-vm-args to the vm command line option for the heap size. Unfortunately, I don't remember the option name. I think this will set the heap for the debugger as opposed to the heap for the process being debugged. Let me know what happens. That's just an alternate way to specify jde-db-option-heap-size. The launch command now contains -Xms400m -Xmx1000m -Xmx600m which is a concatenation of the values in jde-db-option-heap-size and jde-db-option-vm-args. As a side note, the jde-db-option-vm-args looks like it was copied from jde-run-option-vm-args and not completely updated. Troy Paul Troy Paul Troy Troy Daniels [EMAIL PROTECTED] 781-273-3388 x218 Troy Daniels [EMAIL PROTECTED] 781-273-3388 x218 Troy Daniels [EMAIL PROTECTED] 781-273-3388 x218
Crashing during build
Hello, Recently, emacs has started to occasionally fail on jde-build. I'm not certain if the problem is even in JDE, emacs or bash. I type C-c C-v C-b and the following appears in my *compilation* buffer. cd d:/dev/refuel/KBR_ISD_VOB/Jaguar_Plan_Generator_Component/source/Refuel/com/alphatech/kbr/jaguar/refuel/ ant -Dant.home=d:/Program Files/apache-ant-1.6.2 -buildfile 'd:/dev/refuel/KBR_ISD_VOB/Jaguar_Plan_Generator_Component/source/Refuel/build.xml' -emacs 6 [main] bash 1904 handle_exceptions: Exception: STATUS_ACCESS_VIOLATION 68303 [main] bash 1904 open_stackdumpfile: Dumping stack trace to bash.exe.stackdump Compilation finished at Fri Feb 18 15:47:26 It works for a while, but once it starts to fail like this, it will always fail until I log out. Restarting emacs isn't sufficient to fix the problem. All the files exist, and the build works with either M-x compile (with ant -emacs -find build.xml as the build command) or from a bash shell in a different window. Has anyone seen this or have a good guess where the problem might be? I'm running cygwin on Windows for my OS. Troy Troy Daniels [EMAIL PROTECTED] 781-273-3388 x218
Re: Regexp on sourcepath possible?
At 12:58 AM 3/15/2005, Paul Kinnucan wrote: Ole Arndt writes: Hi Lars, But I'm curious as to what kind of a project structure would make regular expressions in source paths useful. The structure that I have at work is one such. We have multiple subprojects side by side: source project1 prj.el com/foo/project1 *.java project2 prj.el com/foo/project2 *.java ... project17 prj.el com/foo/project17 *.java products similar tree with class files Some projects depend on others, but the dependency graph is non-cyclic, so we build and test project 1, then project 2 and so on. Each prj.el sets jde-sourcepath to . and jde-global-classpath to include ../products/projectName/directoriesWithClasses. Which works, except when I try to jump to a source file in a different subproject. What I'd like to do is set jde-sourcepath to (. ../*) but instead, I have to manually do (. ../project1 ../project2 ... ../projectN-1). Which eventually gets built up in each prj.el. And we frequently add more projects, so even if the list is complete, that's only a temporary condition. Whether this is a good coding practice is debatable, but it is what we do. Finally, to handle a path specification like: */src/java the JDEE would have to cycle through every directory on your system, appending /src/java to it, testing whether such a directory exists, and then searching that directory for the source file? Seems like a pretty time-consuming operation that might make finding and opening source files a very slow process. I'd be happy with just something that expands * like a unix wildcard. So */src/java woudn't be the equivalent of a unix find command, but more like the equivalent of ls */src/java. Troy Paul Just be sure to never save your project files with jde-save-project, or you will lose your changes. I have written myself a small module for a related purpose: Read the dependencies from a maven project.xml file and build the jde-global-classpath from this information. My .prj.el files now typically look like this: (setq pom (pom-read-pom)) (jde-set-variables '(jde-global-classpath (pom-get-classpath pom)) '(jde-project-name (pom-get-project-id pom)) '(jde-sourcepath (quote (./src/java ./src/test I have, of course, also a prj.el file for all my default settings in my top level source directory. Ole -- Ole Arndt http://www.sugarshark.com Troy Daniels [EMAIL PROTECTED] 781-273-3388 x218
Re: Regexp on sourcepath possible?
At 12:58 AM 3/15/2005, Paul Kinnucan wrote: Ole Arndt writes: Hi Lars, Lars Degerstedt [EMAIL PROTECTED] writes: Also, I've never understood the big attraction of having to manually find and open a project file, edit it, and then save it, an error-prone and tedious process. Not my idea of a good time but, heck, to each his own. Why do we use JDEE at all instead of just using the customize interface to make emacs perform like we want to? Because there's no option that turns on all the functionality we want. I generally edit prj.el by using the customize interface, but I'll sometimes edit prj.el directly. There's a few cases where I do this, which probably indicates my wish list of prj.el enhancements. I'm copying a prj.el file to a new but related project. I want most of the settings to be the same, but there's some points where the project name is hard-coded. M-x replace-string is actually better than jde-customize-variable, because I know I won't skip a variable. I want to run the program with a different set of arguments, but be able to swap back to the old arguments. Generally, this involves changing the arguments passed to junit.textui.TestRunner from package.SmokeTestSuite to package.NightlyTestSuite. I copy the old value to a comment, and then (usually) uncomment the value from the last time I did this. I want to turn on profiling or remote debugging, so I want to comment/uncomment a fairly complex set of arguments and not have to remember how I set them. Of course, there's always the alternative of proposing a change to the JDEE to enhance the customization interface to support features that you need, e.g., providing the option to specify a function to compute jde-sourcepath or as in this case regular expressions. This is the better long-term solution, assuming somebody has the time and knowledge to implement the solution. Troy Troy Daniels [EMAIL PROTECTED] 781-273-3388 x218
Re: Java-1.5 font-lock support?
At 05:29 PM 4/6/2005, jeff peck wrote: Is there a version of Java font-lock that supports Java-1.5? To so something reasonable with Generics and varargs (Object... args) and other 1.5-isms? While we're discussing 1.5, any idea when Semantic will support it? Troy Troy Daniels [EMAIL PROTECTED] 781-273-3388 x218
bsh and jde-expand-classpath-p
Hello, There appears to be a bug in the use of jde-expand-classpath-p and the bsh. With this directory structure, root prj.el (defines jde-global-classpath) subdir prj.el (inherits jde-global-classpath) the classpath that is passed to the bsh expands jde-global-classpath with respect to subdir, not root. As a result, directories this jar files are passed in as the directory, not the set of jars in the directory. The patch below resolves this problem for me. It won't do the right thing if the classpath is being set from something other than jde-global-classpath (e.g., if it's set from CLASSPATH). *** jde.el.~1.357.~ Thu Dec 23 18:42:48 2004 --- jde.el Tue Aug 30 16:55:30 2005 *** *** 2372,2378 environment variable. (let* ((directory-sep-char ?/) ;; Override NT/XEmacs setting (classpath ! (jde-build-path-arg nil (jde-get-global-classpath) t))) (format jde.util.JdeUtilities.setProjectValues(\%s\, %s); jde-current-project classpath))) --- 2372,2379 environment variable. (let* ((directory-sep-char ?/) ;; Override NT/XEmacs setting (classpath ! (jde-build-path-arg nil (jde-get-global-classpath) ! t 'jde-global-classpath))) (format jde.util.JdeUtilities.setProjectValues(\%s\, %s); jde-current-project classpath))) Troy Daniels [EMAIL PROTECTED] 781-273-3388 x218
Re: JDEbug and static variables
At 10:06 PM 9/15/2005, John wrote: Hi, I recently installed JDEE 2.3.5 in a GNU Emacs 21.3 environment running on Windows XP Pro. I'm using Java J2SE 1.5.0_04. The debugger works fine, however, I can't seem to view static variables. Am I missing something obvious or does JDEbug not work with static variables. Thanks in advance. JDEbug shows static variables inside objects that are local variables (including this). Inside a static method, it won't show the static variables for the class. I'm not aware of a method to display static variables from other classes (e.g., BorderLayout.CENTER) unless you have an instance of that class. Troy Regards, John Troy Daniels [EMAIL PROTECTED] 781-273-3388 x218
