[prev in list] [next in list] [prev in thread] [next in thread] 

List:       orocos-users
Subject:    [Orocos-users] another scripting problem
From:       nick.vanthienen () mech ! kuleuven ! be (Dominick Vanthienen)
Date:       2011-01-19 12:48:58
Message-ID: 4D36DDBA.4040301 () mech ! kuleuven ! be
[Download RAW message or body]





On 01/18/2011 09:17 PM, Peter Soetens wrote:
> On Tuesday 18 January 2011 18:23:42 Dominick Vanthienen wrote:
>> On 01/18/2011 03:58 PM, Peter Soetens wrote:
>>> On Tuesday 18 January 2011 15:09:33 Dominick Vanthienen wrote:
>>>>     Hi,
>>>>
>>>> I've this statemachine, where I try to invoke an operation.
>>>> When I do so, it hangs, in a similar way as in my previous mail.
>>>> (blocking, I can't type anything in the taskBrowser OR a crash (see
>>>> below)) He executes the operation when I type the command in the
>>>> taskBrowser.
>>> I'm assuming you're using the latest toolchain-2.2 ?
>> yes, latest pull this morning
> Which commit id ?
>
>>>> the state machine:
>>>>
>>>> state ControliTaSCScene{
>>>>
>>>>        entry{
>>>>
>>>>            print.ln("hello 3")
>>>>            Scene.CartesianMotion_Wy_global=1.0
>>>>            yield
>>>>            print.ln("hello 4") //he will print this line!
>>>>
>>>> CartesianTrajectoryGenerator.moveTo(KDL.Frame(KDL.Vector(0.1,0.0,0.1),KD
>>>> L.R otation()),5.0) //this is the line he doesn't want to execute
>>> Could it be that moveTo is executed (the C++ function returns), but the
>>> script does not continue ?
>> moveTo is not executed (the first line is to log something)
> That's really weird. This can only happen if the operation is an OwnThread
> kind, and that the EE of that component is blocked for some reason.
>
>>>>            print.ln("hello 5") //he never get to this line :(
>>>>
>>>>        }
>>>>        exit{
>>>>
>>>>            Scene.CartesianMotion_Wy_global=0.0
>>>>            CartesianTrajectoryGenerator.stop()
>>>>            Scene.stop()
>>>>            CartesianMotion.stop()
>>>>            KukaLWR.stop()
>>>>
>>>>          }
>>>>
>>>>        transition if stopRobot==true then select StopRobot
>>>>
>>>>      }
>>>>
>>>> expected behavior (this is the same state machine, but without the
>>>> operation call, I call it manually)
>>>> ------------------------------------------------------------------------
>>>> --- -------------------------------------------------------
>>>>
>>>> SuperVisor [R]>   Test.activate
>>>> hello 1
>>>>
>>>>     = true
>>>>
>>>> SuperVisor [R]>   Test.start
>>>>
>>>>     = true
>>>>
>>>> SuperVisor [R]>   hello 2
>>>> hello 3
>>>> hello 4
>>>> hello 5
>>>>
>>>> SuperVisor [R]>
>>>> CartesianTrajectoryGenerator.moveTo(KDL.Frame(KDL.Vector(0.1,0.0,0.1),KD
>>>> L.R otation()),5.0) = true
>>>>
>>>> behavior with the operation call in the script: (log in attachement)
>>>> -------------------------------------------------------------
>>>> SuperVisor [R]>   Test.activate
>>>> hello 1
>>>>
>>>>     = true
>>>>
>>>> SuperVisor [R]>   Test.start
>>>>
>>>> (when it doesn't crash, he prints hello 2, 3 and 4 here)
>>> I'll need the valgrind output and the RTT compiled with 'Debug' instead.
>>> Which Linux distribution + which architecture are you using ?
>> Linux pma-09-053 2.6.35-22-generic-pae #33-Ubuntu SMP Sun Sep 19
>> 22:14:14 UTC 2010 i686 GNU/Linux
>>
>> No LSB modules are available.
>> Distributor ID: Ubuntu
>> Description:    Ubuntu 10.10
>> Release:        10.10
>> Codename:       maverick
>>
>>> Could you try to compile the *first* test program suggested here:
>>> https://partner-bugzilla.redhat.com/show_bug.cgi?id=593396
>>> and run it for a few minutes ?
>>> g++ test-free.cpp -o test-free -g -lpthread
>> when I run this program, it doesn't show anything in my shell, even
>> after a while
>> what should it do?
>> do I have to run it in a certain folder?
> No. If it didn't segfault after a few minutes, the libc should be fine. I still
> don't understand then why this error is triggered, it doesn't make sense.
>
> Did valgrind show something ?
>
> Peter

SOLVED: thanks Ruben!
all my components had a periodicActivity => deadlocks
by changing it to Activity solved this problem!
(@ peter: an older mail of me with the requested info is blocked by the 
mailing list)

Nick

[prev in list] [next in list] [prev in thread] [next in thread]