Eliot,   Running on the newest cogwin vmI could find (http://www.mirandabanda.org/files/Cog/VM/VM.r2434/cogwin.zip), I still get the lockup in the VM (stack dump below) when I run all tests. When I turn off Zn* and run all other tests, I get a crash (snapshot and dmp attached). VM version info: Cog VM 4.0.0 (release) from Jun 17 2011 Compiler: gcc 3.4.4 (cygming special, gdc 0.12, using dmd 0.125) -Brian  0x90cf1c M Semaphore>critical: 708356544: a(n) Semaphore  0x90cf38 M DelayWaitTimeout(Delay)>unschedule 729655500: a(n) DelayWaitTimeout  0x90cf50 M [] in DelayWaitTimeout>wait 729655500: a(n) DelayWaitTimeout  0x90cf70 M BlockClosure>ensure: 729655620: a(n) BlockClosure  0x90cf8c M DelayWaitTimeout>wait 729655500: a(n) DelayWaitTimeout  0x90cfa8 M Semaphore>waitTimeoutMSecs: 729655068: a(n) Semaphore  0x90cfd4 M Socket>waitForConnectionFor:ifTimedOut: 729655044: a(n) Socket  0x90cffc I Socket>waitForAcceptFor: 729655044: a(n) Socket  0x90d028 I ZnMultiThreadedServer>serveConnectionsOn: 729654372: a(n) ZnMultiThreadedServer  0x90d04c I [] in ZnMultiThreadedServer>listenLoop 729654372: a(n) ZnMultiThreadedServer  0x90d06c M BlockClosure>ifCurtailed: 729655324: a(n) BlockClosure  0x90d090 I ZnMultiThreadedServer>listenLoop 729654372: a(n) ZnMultiThreadedServer  0x90d0b0 I [] in ZnMultiThreadedServer(ZnSingleThreadedServer)>start 729654372: a(n) ZnMultiThreadedServer  0x90d0d0 I [] in BlockClosure>newProcess 729654652: a(n) BlockClosure ________________________________ From: Eliot Miranda <eliot.miranda@gmail.com> To: Pharo-project@lists.gforge.inria.fr; Brian Tabone <brian_tabone@yahoo.com> Sent: Tuesday, June 28, 2011 3:25 PM Subject: Re: [Pharo-project] Re Please test Pharo 1.3! On Tue, Jun 28, 2011 at 1:11 PM, Brian Tabone <brian_tabone@yahoo.com> wrote: Marcus,
   I have run all tests against http://www.mirandabanda.org/files/Cog/VM/VM.r2437/nsvmwin.zip, crashed the VM.
Why are you using the Newspeak VM to run Pharo? Â I tried running Zn* and 151 ran and passed. I then tested all but the Zn* and got further along but had another VM crash. crash2.dmp and Pharo 1.3 VM crash 2 are from this last run.
VM info:
Cog VM 4.0.0 (release) from Jun 19, 2011 Compiler: gcc 3.4.4 (cygming special, gdc 0.12, using dmd 0.125)
-Brian
________________________________ From: Marcus Denker <marcus.denker@inria.fr> To: Pharo-project@lists.gforge.inria.fr; Brian Tabone <brian_tabone@yahoo.com> Sent: Tuesday, June 28, 2011 2:27 PM
Subject: Re: [Pharo-project] Re Please test Pharo 1.3!
On Jun 28, 2011, at 9:20 PM, Brian Tabone wrote:
Sven,
  I have now tried ZnUserAgentTests>>#testDelete by itself and it ran and succeeded, I then ran all ZnUserAgentTests and it passed. I then ran all (and only) Zn* tests and 151 ran and 151 passed. I then ran all tests except Zn* and it ran much further up until the VM crashed. crash snapshots and the small dmp file attached.
The VM seems to be very old... we need to do these tests with the current code base...
-Brian
From: Sven Van Caekenberghe <sven@beta9.be> To: Pharo-project@lists.gforge.inria.fr Sent: Tuesday, June 28, 2011 1:48 PM Subject: Re: [Pharo-project] Re Please test Pharo 1.3!
On 28 Jun 2011, at 19:54, Marcus Denker wrote:
On Jun 28, 2011, at 7:50 PM, Sven Van Caekenberghe wrote:
On 28 Jun 2011, at 18:35, Brian Tabone wrote:
I ran the full test suite on Pharo 1.3 (one click image) on Win 7 64, and it appears that the image hangs up on a socket read. Below is some information from the F2 console window, I dumped stacks, processes, primitives, and socket info. The image does not respond to alt - . (no user interrupt window ever displays). Attached is a screenshot of the image where it hangs. The VM reports its version as:
Teleplace VM 1.0.15 (release) from Oct 4, 2010 Compiler: gcc 3.4.4 (cygming special, gdc 0.12, using dmd 0.125)
Sorry if this is a duplicate or known issue, just
figured I should be on the safe side and report it.
It seems that ZnUserAgentTests>>#testDelete fails for you.
It works for me and for the regular integration/build servers, does the windows slave pass this test ?
The windows slave does not yet run tests.. we will add that ASAP now that it is up and running.
OK, Marcus, I don't want to push, you'll let us know when it is ready.
Brian,
Does the ZnUserAgentTests>>#testDelete succeed when you run it in isolation ?
Does the whole test suite always fail at the same test ?
Sven
<Pharo 1.3 test without Zn.JPG><Pharo 1.3 VM crash.JPG><crash.dmp>
-- Marcus Denker -- http://www.marcusdenker.de INRIA Lille -- Nord Europe. Team RMoD.
-- best, Eliot
