At 06:36 AM 11/7/00 -0800, you wrote:
>To: [EMAIL PROTECTED]
>Subject:
>--text follows this line--
>I am having trouble getting jdk1.3 to do basic debugging on my simple test
>program (which I've included after the JDE state dump.
>
>First I have problems launching the process. The first time I get a time
>out; the second attempt gets me connected.
There is apparently some threading issue with setting up the app standard
I/O socket that occurs on some users' machines but unfortunately has NEVER
occurred on my Windows 95, Windows NT, or Solaris machines (!!!!) and is
thus very difficult for me to debug. I intend to to keep hammering away at
the problem but it would be useful if people who have this problem would
look into the JDEbug code and try to figure out what the problem might be.
>I suspend the process, set a
>breakpoint at the first line, and continue.
This is NOT the way I recommend setting the initial breakpoint. See the
JDEbug user's guide.
Some points.
* The Process->Launch Process command launches the process vm and suspends
the process before it enters the process's main method. Thus, there is NO
NEED FOR YOU TO SUSPEND the process. It is already suspended at startup.
You simply need to set the breakpoint and then execute the Continue command
to enter the process's main method.
* AFAIK, the continue command does not resume suspended threads.The only
way I know for sure to resume a suspended process is to issue a
Process->Resume command. Thus, the following sequence of commands:
suspend
continue
which you used, leaves the thread in a suspended state (AFAIK). The correct
sequence would be
suspend
resume
continue
I could be wrong on this but it wouldn't hurt to use this sequence. In any
case, I have found that it is not usually necessary to suspend a process to
set breakpoints.
- Paul
