Vm-dev April 2009

vm-dev@lists.squeakfoundation.org

22 participants
37 discussions

Can FFI be used with an Alien-aware VM?
by Jay Hardesty 16 Apr '09

16 Apr '09

Hello - new question: Can FFI be used in an Alien-aware VM? Having built a new Mac VM I get "Callout mechanism not available", though the FFI classes and plugin seem to be in place (SqueakFFIPrims.bundle in 'Sqeak VM Opt.app/Contents/Resource'). Also, Applescript generates an error on startup at ApplscriptInstance>>primOpenDefaultConfiguration: type subtype: subtype because the TestOSAPlugin is not defined (although this plugin was not defined in my previous 3.8-biuilt VM). Not sure if this might be related to Alien/FFI somehow.. Thanks once again for all the help (hopefully getting closer). Jay

2 1

current steps for Mac VM build?
by Jay Hardesty 15 Apr '09

15 Apr '09

Thanks very much for the info. But unfortunately now much more confused than before. I'm deploying on OSX 10.5.6 (intel) Following the steps Andreas Raab listed (and if he's confused, then I'm feeling hopeless): - Which "Speech" to download from Squeak-39-alpha (there are several Speech-* files there)? - "Load Freetype somehow"? Is this resolved? The instructions linked by John in that message are the same I was following for the 3.8-6665 image build with VMMaker-3.8.b6, which in turn recommends src-tree revision 1444. Questions raised: 1) Which (all?) of the three files do I install for VMMaker from SqueakSource? Does the order matter? 2) Which "Speech" files, and others besides Balloon3D, are also prerequisites? 2) If rev 1444 still the one to use with a 3.10 image? Or the latest 1963? 3) Which specialChangeSets need to be filed in for correct version of the above? (the files in the link don't match the specialChangeSets folder for either rev 1444 or 1967) 4) Is there a set of instructions more up to date than the link in the message? Thanks once more Jay > At 1:36 AM +0100 4/15/09, Jay Hardesty apparently wrote: >> >>Hi - I've built VM's numerous times in the past, but currently having >>no luck trying to build a Mac Carbon VM on a Intel Mac running 10.5.6, >>using Xcode 3.1.3. >> >>In short (more detail below), I get a link error: library not found >>for lcrt1.10.5.o >> >>Details for the 3.8-image-6665 src-rev-1444 attempt (leaving aside >>details of the previous attempt using a 3.10 image src-rev-1963 that >>ended up with sqVirtualMachine.o undefined's): >> >>Following the readme (platforms/Mac OS/vm/Documentation/readme.txt) I >>took a basic 3.8 image version 6665, installed Balloon3D, and >>VMMaker-3.8b6. >> >>Downloaded version 1444 of the svn source tree (based on the comment >>in the VMMaker 3.8b6 changeset). >> >>Filed in (removing line-feeds as needed): >>VMM38-gc-instrument-image.1.cs >>VMM38-64bit-imageUpdates.1.cs >>ArraysToGlobalStruct-JMM.1.cs >>Gnuifier.6.cs >>(JMM-fixBiasToGrow.1.cs.zip is not in the source tree, and don't see >>it in my change sets - any case left out) >> >>Generated sources (after setting plugins etc as directed), , renamed >>the result from "src32" to "src", and then gnuified. >> >>Placed SqueakVMUNIXPATHS.xcodeproj, resources, and Squeak VM >>Universal-Info.plist in the directory containing "src' and >>"platforms". >> >>Selected "Deployment" as target in xcode and compiled, with result: >> "library not found for lcrt1.10.5.o" >> >> >>Some questions: >>Which "Base SDK" setting: 10.3.9. 10.4, 10.5? (tried clean builds with >> each) >>Should I really (still) be using version 1444 of the source? >>Ditto for Squeak 3.8 6665 (versus a 3.9 or 3.9 image)? >> >> >>Thanks for any help, >>Jay > > > Latest VMMaker is on Squeaksource now. > See this message below from John McIntosh on April 11, 2009: > Later images should work fine as far as I know. I haven't built a VM > lately tho but did successfully awhile back with 3.10 dev images. > What OS X do you intend to deploy on? I'm guessing the base SDK would need > to match since it would need to reside on the target OS. > > Ken G. Brown > > > At 3:01 PM -0700 4/11/09, John M McIntosh apparently wrote: >>Original-recipient: rfc822;kbrown(a)mac.com >>X-Brightmail-Tracker: AAAAAA== >>Delivered-to: list-vm-dev(a)lists.squeakfoundation.org >>From: John M McIntosh <johnmci(a)smalltalkconsulting.com> >>To: Squeak Virtual Machine Development Discussion >> <vm-dev(a)lists.squeakfoundation.org> >>Subject: Re: [Vm-dev] VM team >>Date: Sat, 11 Apr 2009 15:01:44 -0700 >>X-BeenThere: vm-dev(a)lists.squeakfoundation.org >>Reply-to: johnmci(a)smalltalkconsulting.com, >> Squeak Virtual Machine Development Discussion >> <vm-dev(a)lists.squeakfoundation.org> >>List-Id: Squeak Virtual Machine Development Discussion >> <vm-dev.lists.squeakfoundation.org> >>List-Unsubscribe: >> <http://lists.squeakfoundation.org/mailman/listinfo/vm-dev>, >> <mailto:vm-dev-request@lists.squeakfoundation.org?subject=unsubscribe> >>List-Archive: <http://lists.squeakfoundation.org/pipermail/vm-dev> >>List-Post: <mailto:vm-dev@lists.squeakfoundation.org> >>List-Help: >> <mailto:vm-dev-request@lists.squeakfoundation.org?subject=help> >>List-Subscribe: >> <http://lists.squeakfoundation.org/mailman/listinfo/vm-dev>, >> <mailto:vm-dev-request@lists.squeakfoundation.org?subject=subscribe> >>Sender: vm-dev-bounces(a)lists.squeakfoundation.org >> >> >>On 11-Apr-09, at 10:56 AM, Andreas Raab wrote: >> >>>Sounds good to me. I see this more as a formalization of the arrangement >>> we already have, since you have been taking Tim's role lately with doing >>> most of the integration etc. So you've got my support here. >>> >>>In terms of what to do I think what we most need is a bit more clarity >>> for people. For example, how does one actually set up a VMMaker image? >>> You kinda need bits from the four corners of the world and I don't think >>> the process is described anywhere. The shorthand in my VMMaker image >>> says: >>>* Load Speech from http://source.squeakfoundation.org/39a >>>* Load FFI from Universes >>>* Load VMMaker from http://squeaksource.com/VMMaker >>>* Load Balloon3D-Constants, Balloon3D-Plugins from >>> http://squeaksource.com/Balloon3D >>>* Load Freetype somehow >>> >>>This could certainly be improved. The simulator is another piece of the >>> story where we should spend enough time that it can actually be used for >>> simulation of various aspects. And then maybe fix the cruel and unusal >>> UI for VMMaker ;-) >>> >>>Cheers, >>> - Andreas >> >> >> >>Well having pointers to everything on squeakvm.org would help. >>People still load a 3.8 vmmaker from squeaksource for example, so a bit >> more of where do I start if I have this platform would help >> >> >>mac carbon VM build instructions here >> >>http://squeakvm.org/cgi-bin/viewcvs.cgi/trunk/platforms/Mac%20OS/vm/Documen… >> >>-- >>=========================================================================== >>John M. McIntosh <johnmci(a)smalltalkconsulting.com> >>Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com >>=========================================================================== >

4 6

process state of core dump
by David Farber 15 Apr '09

15 Apr '09

I'm still on my quixotic quest to recover a working image from a core dump. Currently, the interpreter is crashing on my image because there is no suspended context. (from gnu-interp.c initializeInterpreter()) /* begin loadInitialContext */ printf("Begin loadInitialContext.\n"); sched = longAt(((longAt((foo->specialObjectsOop + BaseHeaderSize) + (SchedulerAssociation << ShiftForWord))) + BaseHeaderSize) + (ValueIndex << ShiftForWord)); printf("sched: %.8X.\n", sched); proc = longAt((sched + BaseHeaderSize) + (ActiveProcessIndex << ShiftForWord)); printf("proc: %.8X.\n", proc); foo->activeContext = longAt((proc + BaseHeaderSize) + (SuspendedContextIndex << ShiftForWord)); printf("activeContext: %.8X.\n", foo->activeContext); printf("About to beRootIfOld.\n"); if ((((usqInt) foo->activeContext)) < (((usqInt) foo- >youngStart))) { beRootIfOld(foo->activeContext); } As output, I get: memStart: 77DA6000. Reading image. Initializing interpreter. Begin loadInitialContext. sched: 77DA7658. proc: 7A1CEDA4. activeContext: 00000000. About to beRootIfOld. Segmentation fault If I am reading this data correctly, it is saying there is no suspended context. Which would make sense (to me) since this core dump was taken at some random point in the interpreter's machinations. So, what are my options? (Do I have any?) Can I fake or force a suspended context? Am I reading the data wrong? Thanks, David ---- David Farber Red River Software

1 0

current steps for Mac VM build?
by Jay Hardesty 15 Apr '09

15 Apr '09

Hi - I've built VM's numerous times in the past, but currently having no luck trying to build a Mac Carbon VM on a Intel Mac running 10.5.6, using Xcode 3.1.3. In short (more detail below), I get a link error: library not found for lcrt1.10.5.o Details for the 3.8-image-6665 src-rev-1444 attempt (leaving aside details of the previous attempt using a 3.10 image src-rev-1963 that ended up with sqVirtualMachine.o undefined's): Following the readme (platforms/Mac OS/vm/Documentation/readme.txt) I took a basic 3.8 image version 6665, installed Balloon3D, and VMMaker-3.8b6. Downloaded version 1444 of the svn source tree (based on the comment in the VMMaker 3.8b6 changeset). Filed in (removing line-feeds as needed): VMM38-gc-instrument-image.1.cs VMM38-64bit-imageUpdates.1.cs ArraysToGlobalStruct-JMM.1.cs Gnuifier.6.cs (JMM-fixBiasToGrow.1.cs.zip is not in the source tree, and don't see it in my change sets - any case left out) Generated sources (after setting plugins etc as directed), , renamed the result from "src32" to "src", and then gnuified. Placed SqueakVMUNIXPATHS.xcodeproj, resources, and Squeak VM Universal-Info.plist in the directory containing "src' and "platforms". Selected "Deployment" as target in xcode and compiled, with result: "library not found for lcrt1.10.5.o" Some questions: Which "Base SDK" setting: 10.3.9. 10.4, 10.5? (tried clean builds with each) Should I really (still) be using version 1444 of the source? Ditto for Squeak 3.8 6665 (versus a 3.9 or 3.9 image)? Thanks for any help, Jay

2 1

Re: [Pharo-project] [Vm-dev] hampering the desire of the VM and image to visit every object at startup time (multiple times)
by John M McIntosh 14 Apr '09

14 Apr '09

Well a PermSpace would be helpful since it would then avoid a GC which causes the header bits to change making all pages as written, and if the pages are written then the virtual memory system can't free the page under stress. However the real benefit here was that I actually never touch 6MB of objects, so they are never moved from Flash into dynamic RAM. So I do not need to find 1,600 pages of ram to hold pages which the application doesn't need. If I did an allInstances it would transverse PermSpace and bring all those 1600 pages into RAM, really want to avoid that. For example if the iPhone receives a phone call, it starts up a daemon to display a dialog asking if you want to answer or ignore, memory for that has to come from somewhere, or a double-click of the home key can bring up the iPod dialog prompt if the iPod audio app is running, again pages have to come from somewhere. With 700K of old objects, and about 5MB for young space it means the 4.5MB of image objects that is faulted in but not written, can be freed since a page fault will re-read them from the image file. On 14-Apr-09, at 11:02 AM, Nicolas Cellier wrote: > 2009/4/14 John M McIntosh <johnmci(a)smalltalkconsulting.com>: >> Yes, so WikiServer ( http://www.mobilewikiserver.com ) is an image >> file of 10.5 MB. >> >> As startup only 4.5 MB of OOPS memory is faulted in, and about 700K >> of >> memory is altered, which reduces initial memory use by 6MB and >> reduces >> the startup time by 3+ seconds. >> Given the slow speed of the iPhone, and the fact I've a 64MB limit, >> 6MB is a lot, and 3 seconds is welcome. Unfortunately a full GC will >> fault all that 10.5 MB in, however by doing some GC tuning one >> can avoid a full GC until things are quite stressed. >> > > Wouldn't a VW-like PermSpace reduce the burden? > > Nicolas -- = = = ======================================================================== John M. McIntosh <johnmci(a)smalltalkconsulting.com> Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com = = = ========================================================================

1 0

hampering the desire of the VM and image to visit every object at startup time (multiple times)
by John M McIntosh 14 Apr '09

14 Apr '09

I created a pharo entry to track the problem the VM & image has in wanting to visit every smalltalk object multiple times at startup time. Athought this behavior is masked by Gigaherz processors, it's very evident as a problem on the iPhone. Fixing it results in reducing MB of RAM memory usage and saves actual "seconds* of clock time at startup. http://code.google.com/p/pharo/issues/detail?id=737&colspec=ID%20Type%20Sta… -- = = = ======================================================================== John M. McIntosh <johnmci(a)smalltalkconsulting.com> Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com = = = ========================================================================

3 3

Root bit on active context? Closure VMs and thoughts on flushExternalPrimitives.
by John M McIntosh 13 Apr '09

13 Apr '09

Ok, I'm confused about the comments in initialCleanup, since in testing with a Pharo image from feb it DOES call the flushExternalPrimitives, but the comments imply a 3.8 or later image shouldn't have the root bit set. Also then the flushExternalPrimitiveOf: seems to be invoked by the snapshotCleanUp. So is the comment in initialCleanup wrong? Could in the Closure VM could we skip flushExternalPrimitives for the new image type, since we *know* any closure image would have been processed by snapshotCleanUp, versus being a 3.2 image? initialCleanup "Images written by VMs earlier than 3.6/3.7 will wrongly have the root bit set on the active context. Besides clearing the root bit, we treat this as a marker that these images also lack a cleanup of external primitives (which has been introduced at the same time when the root bit problem was fixed). In this case, we merely flush them from here." ((self longAt: activeContext) bitAnd: RootBit) = 0 ifTrue:[^nil]. "root bit is clean" "Clean root bit of activeContext" self longAt: activeContext put: ((self longAt: activeContext) bitAnd: AllButRootBit). "Clean external primitives" self flushExternalPrimitives. then flushExternalPrimitives "Flush the references to external functions from plugin primitives. This will force a reload of those primitives when accessed next. Note: We must flush the method cache here so that any failed primitives are looked up again." | oop primIdx | oop := self firstObject. [self oop: oop isLessThan: endOfMemory] whileTrue: [(self isFreeObject: oop) ifFalse: [(self isCompiledMethod: oop) ifTrue: ["This is a compiled method" primIdx := self primitiveIndexOf: oop. primIdx = PrimitiveExternalCallIndex ifTrue: ["It's primitiveExternalCall" self flushExternalPrimitiveOf: oop]]]. oop := self objectAfter: oop]. self flushMethodCache. self flushObsoleteIndexedPrimitives. self flushExternalPrimitiveTable snapshotCleanUp "Clean up right before saving an image, sweeping memory and: * nilling out all fields of contexts above the stack pointer. * flushing external primitives * clearing the root bit of any object in the root table " | oop header fmt sz | oop := self firstObject. [self oop: oop isLessThan: endOfMemory] whileTrue: [(self isFreeObject: oop) ifFalse: [header := self longAt: oop. fmt := header >> 8 bitAnd: 15. "Clean out context" (fmt = 3 and: [self isContextHeader: header]) ifTrue: [sz := self sizeBitsOf: oop. (self lastPointerOf: oop) + BytesPerWord to: sz - BaseHeaderSize by: BytesPerWord do: [:i | self longAt: oop + i put: nilObj]]. "Clean out external functions" fmt >= 12 ifTrue: ["This is a compiled method" (self primitiveIndexOf: oop) = PrimitiveExternalCallIndex ifTrue: ["It's primitiveExternalCall" self flushExternalPrimitiveOf: oop]]]. oop := self objectAfter: oop]. self clearRootsTable -- = = = ======================================================================== John M. McIntosh <johnmci(a)smalltalkconsulting.com> Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com = = = ========================================================================

2 4

Linux maintainers
by Bert Freudenberg 13 Apr '09

13 Apr '09

Hi VM-Team, Dave, what's your take in this? We somehow need to establish permanent contact with the maintainers of the "squeak-vm" package in various Linux distributions. They operate pretty much autonomously until now ... Even if we get a fix into svn and Ian did a new release, there is no established procedure to make sure all the packages get updates. - Bert - Begin forwarded message: > From: Jerome Peace <peace_the_dreamer(a)yahoo.com> > Date: 12. April 2009 23:10:31 MESZ > To: squeak-dev(a)lists.squeakfoundation.org > Subject: [squeak-dev] Re: Squeak soundless on Ubuntu. Who will fix > it? Hi Chris, > Reply-To: peace_the_dreamer(a)yahoo.com, The general-purpose Squeak > developers list <squeak-dev(a)lists.squeakfoundation.org> > > > > > Hi Chris, > Thank you for your thoughts. > > My question is "Who will fix it." > > The what will be fixed. And the how it will be fixed. Are close to > being known. > Sound works in etoys. > > What I am asking is WHO in our community will take on the task of > making the distribution that Ubuntu gets work with sound. > > Or who in our community will provide the financial resources to > commission it done. > > Without someone with the will to make a good distribution we will > have this problem forever. > > Bert says he can help with his knowledge and expertise, but cannot > be point man on the issue. So the rest of the community needs to > come up with a point man. > > This problem is blocking not only squeak developers but also the > scratch community. > > > So Community. Who will fix it? > > Yours in curiosity and service, --Jerome Peace > > > >

2 1

VM team
by David T. Lewis 12 Apr '09

12 Apr '09

The Squeak oversight board suggested that we formalize a VM team (http://squeakboard.wordpress.com/) and Bert asked if I would be willing to serve as team leader. I replied that I would be happy to do so. I'd like to start by asking the list, and specifically the major platform developers (Ian, John, Andreas) for your support - do you agree that this is a good idea, and are you willing to have me serve as team leader? We will need to set some team objectives, and then establish a plan to achieve them. For starters, here are some initial ideas for a set of VM team objectives. All comments, additions, deletions, objections, etc are welcome. - Community - Communicate status and announcements to the community - Periodically update team objectives based on community input - Ensure availability of known-good versions (SqueakMap, Universes, SqueakSource) - Identify where help is needed (e.g. Sugar, Ubuntu, sound support) - Where appropriate, inform Board of resource or funding needs (e.g. access to computers for building and testing VM distributions) - Support Squeak as an open system - Support educational goals, an accessible VM is part of Squeak as a learning tool - Write it in Squeak, run it in Squeak, debug it in Squeak (InterpreterSimulator should work) - Everyone should be able to explore and create their own Squeak VM - Ensure license integrity - Provide solid base to enable VM distributions - Support distribution developers (Ian, John, Andreas, Tim, others) - Primary support for Mac/Win/unix/iSqueak - Enablement for others including RiscOS and SqueakNOS, new projects - Identify resources to support distribution developers - Coordinate changes across the several platform source trees and VMMaker - Ensure owners and maintainers for plugins - Drive resolution of Mantis issues - Provide base to support new ideas for VM - Support and enable new development (Spoon, Exupery, Cog, Hydra, others) - Enable new image formats (including 64-bit object memory, Cog) - Support Squeak new directions - Support eToys, OLPC, SqueakNOS, Cuis, Pharo - Provide common VM where possible, encourage extensions as needed Dave

5 8

Re: [squeak-dev] Squeak soundless on Ubuntu. Who will fix it?
by Bert Freudenberg 11 Apr '09

11 Apr '09

On 11.04.2009, at 09:25, Jerome Peace wrote: > So is there someone in the community willing to take responsibility > for getting this fixed in the distribution? Please? It would be great indeed if someone who actually knows Squeak would help in the VM packaging efforts for the various Linux distributions. The packages are mostly outdated, not built correctly, missing plugins etc. - though I'm still grateful for the volunteer packagers' work done so far. They're Squeak's ambassadors to the larger open- source community and need our support. I'm here to help and give advice (both as Linux VM hacker and the Board's new VM team advisor) but I can't actually spend the time maintaining a distro package myself (I did so for the OLPC distro but it is switching to plain Fedora now). If you want to help please contact me. - Bert - > Maybe related to > https://bugs.launchpad.net/ubuntu/+source/squeak-vm/+bug/235146

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

Vm-dev April 2009