The problem is still there in 3.5 final. The fix button
in the debugger toolbar is not enabled when I change
some code. I'm pretty sure I can reproduce this problem
on every GUI file written using the form builder.

And to complete the context specification, the file I'm working on are
on a "view over filesystem"

*** Issue 34464 has been marked as a duplicate of this issue. ***

Changing priority, because the 'Fix & Continue' feature does not work
at all.

Please could you evaluate those P1 prio bugs. Thanks.

I agree, please do something to fix this issue. This bug is the *only*
reason for me not to use 3.5. I have reverted to 3.4 but I miss a lot
of the fixed bugs (especially better code completion) from 3.4 .
I am using the 3.5 release on 1.4.1_01 and Redhat 8.0

Evaluation: I disagree with Martins statement, the feature does not
work at all. For me it worked sometimes during debugging, but after
having fixed a couple of classes (the button *was* enabled and it
worked), the only thing that happened when i tried to fix another
class was that according to the debugger output the classes i had
fixed earlier were fixed again, although the source had not changed
again. But until the next debugging session i could not manage to fix
any further classes. Unfortunately i cannot reproduce this behavior
anymore because now i am experiencing the same problem as the OPs :
the button is not enabled anymore.
Therefore i think there is a problem with the state keeping and
propagation of "dirty" sources.

Hope this helps and thanks in advance to anybody who tries to fix this
truely severe issue.

*** Issue 34535 has been marked as a duplicate of this issue. ***

Excuse me if this is a dumb question, but is there anyone out there
using the Netbeans 3.5 Debugger successfully at all?! I have seen 3
installations of Netbeans 3.5 (2 on Unix and 1 on Windows), the fix
button never worked. Am I the only one experiencing these problems? If
not, why isn't this issue addressed by anyone. The 3.5 release version
is useless without this feature compared to 3.4 for anyone who uses
the debugger seriously (which should be most of the users, otherwise
they should continue to use emacs or notepad). 
Sorry if this sounds a bit rude, but I am really disappointed. I
thought that this would be fixed in no time, but the community seems
to be too small to address even the most important bugs (this one was
introduced between the last two release candidates! it used to work
perfectly!).
I know it is not fair blaming other people; I tried to fix it myself,
but this was the first time I took a look at the debugger module's
source code and it was just too much code. If I had to fix it it would
take me weeks of work to understand the current codebase. There must
be someone in the community who is familiar with the code.

Please, someone who is familiar with the codebase, evaluate this
problem, you could make a lot of people very happy users.

Looking at the target milestone is just too frustrating...

No.
You are NOT the only one. I've experienced debugger freeze, not
showing right values in watches, ...
Debugger freeze is described in issue 34693.

It used to work until 3.5 RC1 or maybe even RC2, but I 
can't find the release candidates in the web site. Any clue?

It seems that the functionality broke between release35 RC2 and RC3.

In nevada release it broke between RC5 and RC6.

This regression was probably introduced by fix of #27024.

fixed in main trunk

Good to hear. Where is the diff?

first of all thank you very much for fixing this issue.

@Jan Chalupa:
just check out the module and create a path for the current
jpdaDebugger, it's really easy. I had never done this before, but the
documentation for doing this is very good. (Looking at the code the
diff should be in JPDADebugger.java (the 
very last line of code seemed to have been commented out during the
fix of #27024 - don't ask me why!)):
this.setDataObjectToFix();

@all:
I think there is still one subtle difference to the previous (3.4)
behavior:
when I fixed a class the button is still enabled (which is no problem
but it would be clearer if it was disabled again) but pressing the
button again fixes the class once again although it is not needed.
this wouldn't be too much of a problem unless your are debugging a
method and the debugger is paused, because pressing the fix button
brings up the dialog which asks you whether you want to pop the call
or leave it on the stack, which can be quite annonying.

I really hope that an upgrade with be soon available on the
Netbeans update center because for many people this feature is
important and we all know how much time was required to get 3.4.1 out
(that is, I think that waiting a couple of months to see this fixed in
a 3.5.x is too much). Just an opinion... otherwise in a short time
I'll have to either go back to 3.4.1 or to use a Q-build...

Please, create an AutoUpdate emergency fix, or put avaliable the
correct .jar to fix this. Not all people (like me) has knowledge to
compile NetBeans to make things working...

NetBeans is the best IDE avaliable, put no debug is boring!

My2c.

Verified in dev-200307090100. I'll file another issue for the 'Fix'
button beeing still enabled.

Sorry to ask again, but the fix doens't seem to be included in the 3.5.1 
RC... please someone prepare an update center module with the fix :-( 

FYI the fix button sometimes works for me, but there 2 very annoying
problem.

1- you need to put the focus in a code window before attach, or you
will to do it and reattach.

2- sometimes you cannot fix more than one class per debug session,
again you need to reattach to make it work.

I have added this fix to release35_fixes branch.
It should be available on autoupdate soon.

Still broker in 3.5.1 final release?

Any estimates as when the fix will be available?  In which forms is it
downloadable? Or how can I use autoupdate to download the fix?

Fixed also in Nevada Patch 1 and in Arrow.