Community
Participate
Working Groups
I'm running some of the p2 tests where we spawn a new process to run Eclipse and then come back and analyze the results on disk. While running these tests I'm seeing the following in the console: STDERR: gogo: InterruptedException: sleep interrupted STDERR: java.lang.InterruptedException: sleep interrupted STDERR: at java.lang.Thread.sleep(Native Method) STDERR: at org.apache.felix.gogo.shell.Activator.run(Activator.java:72) STDERR: at java.lang.Thread.run(Thread.java:680) I seem to get one block of these messages for every time we call this code: Process process = Runtime.getRuntime().exec(commandArray, null, output); StreamProcessor.start(process.getErrorStream(), StreamProcessor.STDERR, true); StreamProcessor.start(process.getInputStream(), StreamProcessor.STDOUT, true); process.waitFor(); return process.exitValue();
Actually I can reproduce this by running eclipse with the reconciler application from the command-line. eclipse -application org.eclipse.equinox.p2.reconciler.application I'm using build eclipse-platform-3.8-I20111115-0800-macosx-cocoa-x86_64.tar.gz for my testing.
(In reply to comment #1) > Actually I can reproduce this by running eclipse with the reconciler > application from the command-line. > > eclipse -application org.eclipse.equinox.p2.reconciler.application > > I'm using build eclipse-platform-3.8-I20111115-0800-macosx-cocoa-x86_64.tar.gz > for my testing. I managed to reproduce this with eclipse SDK 3.8 M4, by running the reconciler application with run/debug configuration. It turned out that the activator of the Gogo shell bundle starts a thread. This thread's run() method first sleeps 100 ms. It happens that during this time the reconciler application finishes, the Framework tries to stop, so the Gogo shell's Activator.stop() gets called. The stop() method interrupts the thread, which is still sleeping. Since the interrupt is called by the stop() method, it is not a problem that the run() does not get actually executed. The exception is printed only informatively.
> > Since the interrupt is called by the stop() method, it is not a problem that > the run() does not get actually executed. The exception is printed only > informatively. Well ... it is not very informative! While trying to develop a simple Eclipse Application with Juno M5, I started with simple "println" to make sure worked. All fine, console log in ID appeared as expected. Next, I added a new package (bundle) to manifest.mf file, and a couple of non functioning variables in my "start method". I then tried to "run" again, just to make sure the new bundle could be found, etc., still got my test printline, then I got the message gogo: InterruptedException: null java.lang.InterruptedException at java.lang.Thread.sleep(Native Method) at java.lang.Thread.sleep(Thread.java:851) at org.apache.felix.gogo.shell.Activator.run(Activator.java:72) at java.lang.Thread.run(Thread.java:736) So ... I spent the next two hours trying to figure out what was wrong, and how to open a quality bug report. Then found this bug, which seems to say "it doesn't mean anything". Is it a different problem because it says "null" instead of "sleep interrupted"?
(In reply to comment #3) > Is it a different problem because it says "null" instead of "sleep > interrupted"? I suspect that is a difference in the VM used. DJ and I use that fancy Mac VM ;-) Lazar, did you happen to open an issue against felix gogo about this? If/when you do please reference it here.
Created attachment 210998 [details] Test project I can always reproduce the InterruptedException when running MyTest as a JUnit Plug-in Test (with Oracle jdk6_27). When I add something like "Thread.sleep(5);" to the test, I sometimes get different exceptions, e.g.: gogo: CommandNotFoundException: Command not found: gosh org.apache.felix.gogo.runtime.CommandNotFoundException: Command not found: gosh at org.apache.felix.gogo.runtime.Closure.executeCmd(Closure.java:466) at org.apache.felix.gogo.runtime.Closure.executeStatement(Closure.java:395) at org.apache.felix.gogo.runtime.Pipe.run(Pipe.java:108) at org.apache.felix.gogo.runtime.Closure.execute(Closure.java:183) at org.apache.felix.gogo.runtime.Closure.execute(Closure.java:120) at org.apache.felix.gogo.runtime.CommandSessionImpl.execute(CommandSessionImpl.java:89) at org.apache.felix.gogo.shell.Activator.run(Activator.java:75) at java.lang.Thread.run(Thread.java:662)
(In reply to comment #5) > Created attachment 210998 [details] > Test project > > I can always reproduce the InterruptedException when running MyTest as a > JUnit Plug-in Test (with Oracle jdk6_27). > > When I add something like "Thread.sleep(5);" to the test, I sometimes get > different exceptions, e.g.: > > gogo: CommandNotFoundException: Command not found: gosh > org.apache.felix.gogo.runtime.CommandNotFoundException: Command not found: gosh > at org.apache.felix.gogo.runtime.Closure.executeCmd(Closure.java:466) > at org.apache.felix.gogo.runtime.Closure.executeStatement(Closure.java:395) > at org.apache.felix.gogo.runtime.Pipe.run(Pipe.java:108) > at org.apache.felix.gogo.runtime.Closure.execute(Closure.java:183) > at org.apache.felix.gogo.runtime.Closure.execute(Closure.java:120) > at > org.apache.felix.gogo.runtime.CommandSessionImpl.execute(CommandSessionImpl.java:89) > at org.apache.felix.gogo.shell.Activator.run(Activator.java:75) > at java.lang.Thread.run(Thread.java:662) The reason for the exceptions is the same as for the interrupted exception - the closing of the framework is done while the gogo shell performs initialization actions. The exceptions are different due to timing issues. Actually, if you use a longer interval in Thread.sleep(), no exceptions occur because the shell gets initialized during this sleep.
> Lazar, did you happen to open an issue against felix gogo about this? If/when > you do please reference it here. https://issues.apache.org/jira/browse/FELIX-3354
*** Bug 399986 has been marked as a duplicate of this bug. ***
To Equinox maintainers: could you update your Apache Felix Gogo dependency so that a fix for this bug would be included in Equinox? Thanks. The issue was closed in the upstream more than a year ago: https://issues.apache.org/jira/browse/FELIX-3354 https://issues.apache.org/jira/browse/FELIX-3331 https://issues.apache.org/jira/browse/FELIX-5077
(In reply to Alexander Levsha from comment #9) > To Equinox maintainers: could you update your Apache Felix Gogo dependency > so that a fix for this bug would be included in Equinox? Thanks. > > The issue was closed in the upstream more than a year ago: > https://issues.apache.org/jira/browse/FELIX-3354 > https://issues.apache.org/jira/browse/FELIX-3331 > https://issues.apache.org/jira/browse/FELIX-5077 Yes, see bug 514338 *** This bug has been marked as a duplicate of bug 514338 ***
