October 2011 - Pharo-dev - lists.pharo.org

Re: [Pharo-project] crash.dmp file

by John Toohey

Am now using the latest VM from Elliot. Looks good so far, but most of my users have left, so I'll continue to monitor it. Actually has to write a new script to get it to work, and run that directly from a bin directory containing the VM and the plugins. For some reason, even with the -plugins directive, the VM could not find the vm-display-null plugin. On Wed, Oct 19, 2011 at 11:44, Igor Stasenko <siguctua(a)gmail.com> wrote: > On 19 October 2011 16:04, John Toohey <jt(a)parspro.com> wrote: >> I just downloaded the latest COG Unix VM from Jenkins. Can I just copy >> the contents of the zip into the one-click image that I am running? >> (Into ../Contents/Linux) >> > > Just keep in mind that VMs built on Jenkins using different naming for > executable > i.e. CogVM instead of squeak. > So you may need to fix script(s) to invoke it properly. > >> >> >> On Wed, Oct 19, 2011 at 09:50, Lukas Renggli <renggli(a)gmail.com> wrote: >>> I am using VM.2496 from Eliot's builds for the build server and >>> TextLint. This version works quite well for me. >>> >>> Lukas >>> >>> On 19 October 2011 15:19, John Toohey <jt(a)parspro.com> wrote: >>>> Do you know of a VM that is in production somewhere? I've switched to >>>> the COG VMs (much faster than the old ones) , but they all seem very >>>> unstable. Maybe I'm taking the wrong ones from the build server, but I >>>> desperately need one that is stable enough to deploy. I'm using this >>>> one, at the moment, Â VM Version: Croquet Closure Cog VM [CoInterpreter >>>> VMMaker-oscog-IgorStasenko.123] >>>> >>>> On Wed, Oct 19, 2011 at 08:44, Lukas Renggli <renggli(a)gmail.com> wrote: >>>>> PharoDebug.log is created before the image opens a debugger. >>>>> >>>>> crash.dmp is created when the VM dies. >>>>> >>>>> Lukas >>>>> >>>>> On 19 October 2011 14:26, John Toohey <jt(a)parspro.com> wrote: >>>>>> One of my production servers was down this morning, but instead of a >>>>>> PharoDebug.log file, all I found was a file called crash.dmp, with a >>>>>> minimum stack trace and no information about what caused the crash. >>>>>> Can someone tell me what this file is, and why the usual log file was >>>>>> not created? >>>>>> >>>>>> >>>>>> -- >>>>>> ~JT >>>>>> >>>>>> >>>>> >>>>> >>>>> >>>>> -- >>>>> Lukas Renggli >>>>> www.lukas-renggli.ch >>>>> >>>>> >>>> >>>> >>>> >>>> -- >>>> ~JT >>>> >>>> >>> >>> >>> >>> -- >>> Lukas Renggli >>> www.lukas-renggli.ch >>> >>> >> >> >> >> -- >> ~JT >> >> > > > > -- > Best regards, > Igor Stasenko. > > -- ~JT

Oct. 19, 2011

Re: [Pharo-project] crash.dmp file

by Igor Stasenko

On 19 October 2011 16:04, John Toohey <jt(a)parspro.com> wrote: > I just downloaded the latest COG Unix VM from Jenkins. Can I just copy > the contents of the zip into the one-click image that I am running? > (Into ../Contents/Linux) > Just keep in mind that VMs built on Jenkins using different naming for executable i.e. CogVM instead of squeak. So you may need to fix script(s) to invoke it properly. > > > On Wed, Oct 19, 2011 at 09:50, Lukas Renggli <renggli(a)gmail.com> wrote: >> I am using VM.2496 from Eliot's builds for the build server and >> TextLint. This version works quite well for me. >> >> Lukas >> >> On 19 October 2011 15:19, John Toohey <jt(a)parspro.com> wrote: >>> Do you know of a VM that is in production somewhere? I've switched to >>> the COG VMs (much faster than the old ones) , but they all seem very >>> unstable. Maybe I'm taking the wrong ones from the build server, but I >>> desperately need one that is stable enough to deploy. I'm using this >>> one, at the moment, Â VM Version: Croquet Closure Cog VM [CoInterpreter >>> VMMaker-oscog-IgorStasenko.123] >>> >>> On Wed, Oct 19, 2011 at 08:44, Lukas Renggli <renggli(a)gmail.com> wrote: >>>> PharoDebug.log is created before the image opens a debugger. >>>> >>>> crash.dmp is created when the VM dies. >>>> >>>> Lukas >>>> >>>> On 19 October 2011 14:26, John Toohey <jt(a)parspro.com> wrote: >>>>> One of my production servers was down this morning, but instead of a >>>>> PharoDebug.log file, all I found was a file called crash.dmp, with a >>>>> minimum stack trace and no information about what caused the crash. >>>>> Can someone tell me what this file is, and why the usual log file was >>>>> not created? >>>>> >>>>> >>>>> -- >>>>> ~JT >>>>> >>>>> >>>> >>>> >>>> >>>> -- >>>> Lukas Renggli >>>> www.lukas-renggli.ch >>>> >>>> >>> >>> >>> >>> -- >>> ~JT >>> >>> >> >> >> >> -- >> Lukas Renggli >> www.lukas-renggli.ch >> >> > > > > -- > ~JT > > -- Best regards, Igor Stasenko.

Oct. 19, 2011

Re: [Pharo-project] IS Smalltalk Source also an object?

by Hilaire Fernandes

Le 19/10/2011 15:36, Schwab,Wilhelm K a Ã©crit : > So far, the one "deployed executable" that I have from Pharo is about 60 MB image+changes. In fairness, I was happy to get it working (I use it pretty much every day), and have not attempted any tricks to reduce its size. > Take a look to the DrGeo set up I use to build a deployed executable from a core image. https://gforge.inria.fr/scm/viewvc.php/drgeo2/?root=istoa Hilaire -- Education 0.2 -- http://blog.ofset.org/hilaire

Oct. 19, 2011

Re: [Pharo-project] crash.dmp file

by Igor Stasenko

Err, sorry i cannot understand what is the question? On 19 October 2011 16:29, Torsten Bergmann <astares(a)gmx.de> wrote: > Did you try with one of Eliots Cog VMs > http://lists.squeakfoundation.org/pipermail/vm-dev/2011-October/009540.html > > Since Eliot fixes bugs there the most up to date > version (currently VM.r2499) may already include a bugfix. > R2499 is the revision in Eliots subversion repo > (see http://squeakvm.org/cgi-bin/viewcvs.cgi/branches/Cog/) > > AFAIK Igor later also includes these fixes into the Git-Managed > VM that is the source to build the VMs on Jenkins > > Bye > T. > > -- > Empfehlen Sie GMX DSL Ihren Freunden und Bekannten und wir > belohnen Sie mit bis zu 50,- Euro! https://freundschaftswerbung.gmx.de > > -- Best regards, Igor Stasenko.

Oct. 19, 2011

[Pharo-project] crash.dmp file

by Torsten Bergmann

Did you try with one of Eliots Cog VMs http://lists.squeakfoundation.org/pipermail/vm-dev/2011-October/009540.html Since Eliot fixes bugs there the most up to date version (currently VM.r2499) may already include a bugfix. R2499 is the revision in Eliots subversion repo (see http://squeakvm.org/cgi-bin/viewcvs.cgi/branches/Cog/) AFAIK Igor later also includes these fixes into the Git-Managed VM that is the source to build the VMs on Jenkins Bye T. -- Empfehlen Sie GMX DSL Ihren Freunden und Bekannten und wir belohnen Sie mit bis zu 50,- Euro! https://freundschaftswerbung.gmx.de

Oct. 19, 2011

Re: [Pharo-project] crash.dmp file

by John Toohey

I just downloaded the latest COG Unix VM from Jenkins. Can I just copy the contents of the zip into the one-click image that I am running? (Into ../Contents/Linux) On Wed, Oct 19, 2011 at 09:50, Lukas Renggli <renggli(a)gmail.com> wrote: > I am using VM.2496 from Eliot's builds for the build server and > TextLint. This version works quite well for me. > > Lukas > > On 19 October 2011 15:19, John Toohey <jt(a)parspro.com> wrote: >> Do you know of a VM that is in production somewhere? I've switched to >> the COG VMs (much faster than the old ones) , but they all seem very >> unstable. Maybe I'm taking the wrong ones from the build server, but I >> desperately need one that is stable enough to deploy. I'm using this >> one, at the moment, Â VM Version: Croquet Closure Cog VM [CoInterpreter >> VMMaker-oscog-IgorStasenko.123] >> >> On Wed, Oct 19, 2011 at 08:44, Lukas Renggli <renggli(a)gmail.com> wrote: >>> PharoDebug.log is created before the image opens a debugger. >>> >>> crash.dmp is created when the VM dies. >>> >>> Lukas >>> >>> On 19 October 2011 14:26, John Toohey <jt(a)parspro.com> wrote: >>>> One of my production servers was down this morning, but instead of a >>>> PharoDebug.log file, all I found was a file called crash.dmp, with a >>>> minimum stack trace and no information about what caused the crash. >>>> Can someone tell me what this file is, and why the usual log file was >>>> not created? >>>> >>>> >>>> -- >>>> ~JT >>>> >>>> >>> >>> >>> >>> -- >>> Lukas Renggli >>> www.lukas-renggli.ch >>> >>> >> >> >> >> -- >> ~JT >> >> > > > > -- > Lukas Renggli > www.lukas-renggli.ch > > -- ~JT

Oct. 19, 2011

Re: [Pharo-project] crash.dmp file

by Lukas Renggli

I am using VM.2496 from Eliot's builds for the build server and TextLint. This version works quite well for me. Lukas On 19 October 2011 15:19, John Toohey <jt(a)parspro.com> wrote: > Do you know of a VM that is in production somewhere? I've switched to > the COG VMs (much faster than the old ones) , but they all seem very > unstable. Maybe I'm taking the wrong ones from the build server, but I > desperately need one that is stable enough to deploy. I'm using this > one, at the moment, Â VM Version: Croquet Closure Cog VM [CoInterpreter > VMMaker-oscog-IgorStasenko.123] > > On Wed, Oct 19, 2011 at 08:44, Lukas Renggli <renggli(a)gmail.com> wrote: >> PharoDebug.log is created before the image opens a debugger. >> >> crash.dmp is created when the VM dies. >> >> Lukas >> >> On 19 October 2011 14:26, John Toohey <jt(a)parspro.com> wrote: >>> One of my production servers was down this morning, but instead of a >>> PharoDebug.log file, all I found was a file called crash.dmp, with a >>> minimum stack trace and no information about what caused the crash. >>> Can someone tell me what this file is, and why the usual log file was >>> not created? >>> >>> >>> -- >>> ~JT >>> >>> >> >> >> >> -- >> Lukas Renggli >> www.lukas-renggli.ch >> >> > > > > -- > ~JT > > -- Lukas Renggli www.lukas-renggli.ch

Oct. 19, 2011

Re: [Pharo-project] crash.dmp file

by John Toohey

All I have it the following. I don't know when it crashs, it just goes down when users are logged in. Segmentation fault Mon Oct 17 15:57:44 2011 C stack backtrace: ./Contents/Linux/pharo[0x8096919] ./Contents/Linux/pharo[0x8096a8c] [0x667410] [0x777fe1ed] [0x777e00f4] [0x777dfce4] [0x777dff95] [0x777de41f] [0x777de520] [0x77806621] [0x777d7021] [0x778065a7] [0x777de698] [0x777de7ec] [0x777d47c8] [0x777dd34d] [0x777dd29b] [0x777dd220] [0x777dd607] [0x777dd34d] [0x777dd29b] [0x777dd220] [0x777dd49d] [0x777dd3e6] [0x777dd582] [0x777d47c8] [0x7936a714] [0x7936ac1d] [0x777dd34d] [0x777dd29b] [0x777dd220] [0x777dd607] [0x777dd34d] [0x777dd29b] [0x777dd220] [0x777dd49d] [0x777dd3e6] [0x777dd582] [0x777d47c8] [0x777d4668] Smalltalk stack dump: 0xbf94bb90 M [] in JulietDataStore>activeBookmakers 2060043820: a(n) JulietDataStore 0xbf94bbac M BlockClosure>asGlorpExpressionOn:in: 2063483052: a(n) BlockClosure 0xbf94bbcc M [] in SimpleQuery(AbstractReadQuery)>setUpExpressions 2063482800: a(n) SimpleQuery On Wed, Oct 19, 2011 at 09:34, Sven Van Caekenberghe <sven(a)beta9.be> wrote: > > On 19 Oct 2011, at 15:19, John Toohey wrote: > >> But I desperately need one that is stable enough to deploy. I'm using this >> one, at the moment, Â VM Version: Croquet Closure Cog VM [CoInterpreter >> VMMaker-oscog-IgorStasenko.123] > > AFAIK that VM is OK on Linux, it works for me, and for the build servers. > > Maybe you need to investigate a bit more, or provide more detail, no ? > > Sven > > > -- ~JT

Oct. 19, 2011

Re: [Pharo-project] IS Smalltalk Source also an object?

by Schwab,Wilhelm K

In graduate school, I would have been hard pressed to store my current/typical Smalltalk image, let alone move it from place to place, so think about disk space too. Overall, we probably still benefit from the efficiency that limited resources demanded. With very little effort, Dolphin produced deployed executables of 1 (with work) to 4 (being sloppy about it) MB. There is of course a fair amount of support code in the form of DLLs, both from Windows and the VM, but the part that one had to deliver for a hot fix was very manageable. So far, the one "deployed executable" that I have from Pharo is about 60 MB image+changes. In fairness, I was happy to get it working (I use it pretty much every day), and have not attempted any tricks to reduce its size. Bill ________________________________________ From: pharo-project-bounces(a)lists.gforge.inria.fr [pharo-project-bounces(a)lists.gforge.inria.fr] On Behalf Of Marcus Denker [marcus.denker(a)inria.fr] Sent: Wednesday, October 19, 2011 2:49 AM To: Pharo-project(a)lists.gforge.inria.fr Subject: Re: [Pharo-project] IS Smalltalk Source also an object? On Oct 18, 2011, at 9:31 PM, Nicolas Cellier wrote: > 2011/10/18 Schwab,Wilhelm K <bschwab(a)anest.ufl.edu>: >> Maybe I've played too much football without a helmet or did too much C programming, > > Yes, and yes, that's why you missed Eliot's answer ;) > http://permalink.gmane.org/gmane.comp.lang.smalltalk.pharo.devel/55024 > > Nicolas > > > but I see two questions popping out of this: (1) is source an object > in Smalltalk; So yes, I have been convinced that it is an object (as even a string is an object). (yet, we should think if with the resources we have today as opposed to 1978, can't we do even better? I always wonder how smalltalk would look like if they would have had 4GB of RAM instead of 256KB...) Marcus -- Marcus Denker -- http://marcusdenker.de

Oct. 19, 2011

Re: [Pharo-project] crash.dmp file

by Sven Van Caekenberghe

On 19 Oct 2011, at 15:19, John Toohey wrote: > But I desperately need one that is stable enough to deploy. I'm using this > one, at the moment, VM Version: Croquet Closure Cog VM [CoInterpreter > VMMaker-oscog-IgorStasenko.123] AFAIK that VM is OK on Linux, it works for me, and for the build servers. Maybe you need to investigate a bit more, or provide more detail, no ? Sven

Oct. 19, 2011