Paul Kinnucan wrote:
>
> At 04:01 PM 4/28/00 -0400, you wrote:
> >
> >I have this problem, and for me sometimes hitting return twice in the CLI.*
> >window will cause it to show up. I've reported it before, but no one else
> >seems to have seen it. My app runs a interpreter and the only way to even
> >see the prompt is two returns.
> >
>
> Okay, now you've reminded me that I have heard of similar problems before.
> In every case that I can recall, the problem turned out to be some problem
> with how the Java process is writing to standard out. I believe in JDK
> 1.1.x there was a bug with buffered output that caused a problem for the
> beanshell. This was solved, I believe, by fiddling with the output buffer
> on the Java side (reducing the buffer to one char?). Anyway, if you're
> having problems, it might be worth pursuing the idea that the output is
> being buffered on the Java side and the buffer is not filled up and so
> never gets written out to standard out, e.g., if you have a 100 character
> i/o buffer on the Java side and your initial user prompt is only two
> characters, e.g., "> " with no eol, you would need to flush the output
> buffer to get it to appear on the emacs side.
>
> - Paul
>
If I just run my program from an MS-DOS prompt using java I see the
System.out messages, so the problem doesn't seem to be buffering. The
version of JDK I'm using is 1.2.2.
I'm using emacs 20.6 and JDE2.1.6beta23. Perhaps the stability of one
of those systems comes into play?
I've noticed something else that seems unrelated. When I try to step
into a constructor in a different file I get a prompt for the file
location, and when I enter the file name the cursor goes to the wrong
spot. The debug session is confused after that. However, if I set a
breakpoint in the other file, the debugger gets there just fine. Could
this be related?
I'm assuming the output should show up in the main class process window,
is that correct? What's a CLI.* window?
Regards,
JDE newbie...
[EMAIL PROTECTED]
