Java Thread Dump in Bash

[Randall R Schulz]

Frank-Michael,

At 12:09 2004-02-17, Frank-Michael Moser wrote:
>Unfortunately neither Ctrl-Break nor Ctrl-Scroll produce a thread dump 
>not in pure bash and not in RXVT for me. I'm using latest cygwin and 
>tried JDK 1.4.2_02 and 1.5.0 beta.
>
>Also the java applications I tried do not read from standard input.
>
>What versions (cygwin/java) do you use.

Cygwin: Latest "kernel" and latest version of all packages.
Java: Latest (specifically, 1.4.2_03); If it might matter, I generally 
use the "java" and "javac" commands from the SDK bin, not the "jre" 
bin. That's the directory that includes the compiler and related 
development tools, where as the jre bin has only the JVM and other 
runtime resources.

By "pure BASH" I take it you mean BASH in a console window, in contrast 
to an RXVT window. I do _not_ use the "tty" option in the CYGWIN 
environment variable. Do you?


>Is it worth to send my cygcheck output attached? Should I really 
>expect Ctrl-Break to work - this would be great?

I expect it to work because it does work for me...

Cygcheck output isn't anything I can use for any purpose I can think of 
in resolving this discrepancy between how your system and mine behave.


>Frank-Michael

Randall Schulz


>Randall R Schulz wrote:
>
>>Frank-Michael,
>>
>>CTRL-BREAK produces a thread-dump using the latest Sun JVM on my 
>>system when launched from BASH. However, if the program is reading 
>>standard input from the unredirected console, it receives an 
>>end-of-file indication on that stream as well.
>>
>>Randall Schulz
>>
>>At 03:42 2004-02-17, Frank-Michael Moser wrote:
>>
>>>Searching the mailing list archive I found that there is an old 
>>>thread from Dec 2000 which exactly describes my problem:
>>>
>>>http://sources.redhat.com/ml/cygwin/2000-12/msg00490.html
>>>
>>>In short: Using Ctrl-Scroll does not cause a Java program to dump 
>>>threads as it does in cmd.exe. Unfortunately the thread ended up 
>>>with some personal strife.
>>>
>>>I understand that the signal problem could be by design. But now (3 
>>>years later) maybe there are news about this issue? Has someone a 
>>>way to work around this problem?
>>>
>>>Frank-Michael


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/