Hi, Is there anyone who recently was able to install VMmaker into Trunk? Any tips? I tried to load VMmaker 3.8b6 into the latest Squeak3.10.2- Trunk-091128 (that I first updated with 'load code updates').. I get the error 'Cannot compile -- stack including temps is too deep' Background: I would really like to create the latest Trunk and VMmaker without any install problems so I can start debugging the simulator. The simulator can no longer simulate images later than 3.7 and I need it to simulate the Trunk inside the Trunk. Cheers, - Merik Voswinkel ========================================== 13 December 2009 12:09:38.694 am VM: Mac OS - a SmalltalkImage Image: Squeak3.10.2 [latest update: #3805] SecurityManager state: Restricted: false FileAccess: true SocketAccess: true Working Dir /Users/merik/Desktop/Squeak/Squeak3.10.2-Trunk-091128 Trusted Dir /foobar/tooBar/forSqueak/bogus Untrusted Dir /Users/merik/Library/Preferences/Squeak/Internet/My Squeak CompiledMethod(Object)>>error: Receiver: (GeniePlugin>>nil "a CompiledMethod(3606)") Arguments and temporary variables: aString: 'Cannot compile -- stack including temps is too deep' Receiver's instance variables: (GeniePlugin>>nil "a CompiledMethod(3606)") CompiledMethod>>needsFrameSize: Receiver: (GeniePlugin>>nil "a CompiledMethod(3606)") Arguments and temporary variables: newFrameSize: 6 largeFrameBit: 131072 header: nil Receiver's instance variables: (GeniePlugin>>nil "a CompiledMethod(3606)") BytecodeAgnosticMethodNode>>generate: Receiver: primSameClassAbsoluteStrokeDistanceMyPoints: myPointsOop otherPoints: otherPointsOop myVec...etc... Arguments and temporary variables: trailer: #(0 0 0 0) primErrNode: nil blkSize: 1021 nLits: 44 literals: {#var:type: . #myPoints . 'int * ' . #otherPoints . #myVectors . #ot...etc... stack: a ParseStack at 52 of 57 method: (GeniePlugin>>nil "a CompiledMethod(3606)") locals: an OrderedCollection({myPointsOop} {otherPointsOop} {myVectorsOop} {oth...etc... lit: 45 Receiver's instance variables: comment: nil pc: nil selectorOrFalse: #primSameClassAbsoluteStrokeDistanceMyPoints:otherPoints:myVec...etc... precedence: 3 arguments: an OrderedCollection({myPointsOop} {otherPointsOop} {myVectorsOop} {...etc... block: {[:myPointsOop :otherPointsOop :myVectorsOop :otherVectorsOop :mySquared...etc... primitive: 0 encoder: {an EncoderForV3PlusClosures} temporaries: {{base} . {insertRemove} . {jLimiT} . {substBase} . {insert} . {re...etc... properties: an AdditionalMethodState sourceText: 'primSameClassAbsoluteStrokeDistanceMyPoints: myPointsOop otherPoin...etc... locationCounter: 3 localsPool: an IdentitySet({myVectorsOop} {otherPointsOop} {otherSquaredLengths...etc... CompiledMethodWithNode class>>generateMethodFromNode:trailer: Receiver: CompiledMethodWithNode Arguments and temporary variables: aMethodNode: primSameClassAbsoluteStrokeDistanceMyPoints: myPointsOop otherPoin...etc... bytes: #(0 0 0 0) Receiver's instance variables: superclass: Object methodDict: a MethodDictionary(#method- >(CompiledMethodWithNode>>#method "a Com...etc... format: 134 instanceVariables: #('node' 'method') organization: ('accessing' method node selector) ('private' method: node:) subclasses: nil name: #CompiledMethodWithNode classPool: nil sharedPools: nil environment: Smalltalk "a SystemDictionary with lots of globals" category: #'Compiler-Support' traitComposition: nil localSelectors: nil --- The full stack --- CompiledMethod(Object)>>error: CompiledMethod>>needsFrameSize: BytecodeAgnosticMethodNode>>generate: CompiledMethodWithNode class>>generateMethodFromNode:trailer: - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - GeniePlugin class(Behavior)>>compile:classified:notifying:trailer:ifFail: MethodAddition>>createCompiledMethod [] in [] in [] in MCPackageLoader>>basicLoad [] in [] in OrderedCollection(SequenceableCollection)>>do:displayingProgress: OrderedCollection(SequenceableCollection)>>withIndexDo: [] in OrderedCollection(SequenceableCollection)>>do:displayingProgress: [] in [] in ProgressInitiationException>>defaultMorphicAction BlockClosure>>on:do: [] in ProgressInitiationException>>defaultMorphicAction BlockClosure>>ensure: ProgressInitiationException>>defaultMorphicAction ProgressInitiationException>>defaultAction UndefinedObject>>handleSignal: MethodContext(ContextPart)>>handleSignal: MethodContext(ContextPart)>>handleSignal: MethodContext(ContextPart)>>handleSignal: ProgressInitiationException(Exception)>>signal ProgressInitiationException>>display:at:from:to:during: ProgressInitiationException class>>display:at:from:to:during: ByteString(String)>>displayProgressAt:from:to:during: OrderedCollection(SequenceableCollection)>>do:displayingProgress: [] in [] in MCPackageLoader>>basicLoad BlockClosure>>on:do: [] in MCPackageLoader>>basicLoad BlockClosure>>ensure: ...etc... |
On Sun, Dec 13, 2009 at 12:22:53AM +0100, Merik Voswinkel wrote:
> > Hi, > > Is there anyone who recently was able to install VMmaker into Trunk? > Any tips? > I tried to load VMmaker 3.8b6 into the latest Squeak3.10.2- > Trunk-091128 (that I first updated with 'load code updates').. I get > the error 'Cannot compile -- stack including temps is too deep' > > Background: I would really like to create the latest Trunk and VMmaker > without any install problems so I can start debugging the simulator. > The simulator can no longer simulate images later than 3.7 and I need > it to simulate the Trunk inside the Trunk. Merik, Please load the latest version of VMMaker from http://squeaksource.com/VMMaker. The 3.8b6 version that you may have gotten from SqueakMap or Universes is now quite out of date. If you load the latest version from SqueakSource, then evaluate "VMMaker updateFromServer", you will have the latest VMMaker along with updates for some other plugins and utilities. Dave |
On Sunday 13 December 2009 08:20:14 am David T. Lewis wrote:
> Please load the latest version of VMMaker from > http://squeaksource.com/VMMaker. The 3.8b6 version that you may have > gotten from SqueakMap or Universes is now quite out of date. If you load > the latest version from SqueakSource, then evaluate "VMMaker > updateFromServer", you will have the latest VMMaker along with updates for > some other plugins and utilities. This gives a warning about missing dependencies on KlattResonatorIndices and FFIPlugin. Proceeding completes the load without any further issues (may need to start squeak with more memory to avoid running out of memory). Not sure if the broken dependencies of FFIPlugin may create problems down the road. BTW, there is a ready-made image with VMMaker at squeakvm.org/unix/ (search for unix-*-vmm.tar.gz Subbu |
In reply to this post by David T. Lewis
Hi,
My VMMaker do not understand "updateFromServer". Anything I miss out? Thanks. Ang Beepeng
|
In reply to this post by K. K. Subramaniam
Hi,
> On Sunday 13 December 2009 08:20:14 am David T. Lewis wrote: >> Please load the latest version of VMMaker from >> http://squeaksource.com/VMMaker. The 3.8b6 version that you may have >> gotten from SqueakMap or Universes is now quite out of date. If you >> load >> the latest version from SqueakSource, then evaluate "VMMaker >> updateFromServer", you will have the latest VMMaker along with >> updates for >> some other plugins and utilities. Thank you David, that worked this time (it did not load a few weeks ago). It would be helpful if there was a comment somewhere in the description that you only need to load VMMaker to also get slangbrowser etc. Loading all the code you find there will not work. > On Dec 13, 2009, at 7:30 AM, K. K. Subramaniam wrote: > This gives a warning about missing dependencies on > KlattResonatorIndices and > FFIPlugin. Proceeding completes the load without any further issues > (may need > to start squeak with more memory to avoid running out of memory). > Not sure if > the broken dependencies of FFIPlugin may create problems down the > road. Yes they did... > > BTW, there is a ready-made image with VMMaker at squeakvm.org/unix/ > (search > for unix-*-vmm.tar.gz Thank you Subbu, that hint helped a lot. I hoped there was one ready-made image (this one is from september 12th 2009) > > Subbu > Cheers, Merik |
In reply to this post by Ang BeePeng
On Sunday 13 December 2009 01:54:55 pm Ang BeePeng wrote:
> My VMMaker do not understand "updateFromServer". Anything I miss out? This method has a timestamp "dtl 11/14/2009 20:38 VMMaker class". What is your VMMaker version? Subbu |
In reply to this post by Merik Voswinkel-3
I did it two days ago.
Before I'd like to thanks everybody help. Now, except for UUIDPlugin which code needs a fix for Fedora (#include <uuid.h> must be #include <uuid/uuid.h>) all plugins are compiling clear. Another little issue is that make install copies plugins as so.xxx to /usr/local/lib/squeak/3.11.3- and I had to link everything to correct names xxx.so: ============================================= # cd /usr/local/lib/squeak/3.11.3- # for FILES in so.* > do > ln $FILES `echo $FILES | cut -f2 -d.`.so > done ============================================= To have VMMaker in trunk: 1st: Install Speech from Monticello Browser (repository http://www.sourcecode.com/Speech 2nd: Install VMMaker from Monticello Browser (repository http://www.sourcecode.com/VMMaker I guess that would be enough to have VMMaker working. Best regards, CdAB Em 12-12-2009 21:22, Merik Voswinkel escreveu: > > Hi, > > Is there anyone who recently was able to install VMmaker into Trunk? > Any tips? > I tried to load VMmaker 3.8b6 into the latest > Squeak3.10.2-Trunk-091128 (that I first updated with 'load code > updates').. I get the error 'Cannot compile -- stack including temps > is too deep' > > Background: I would really like to create the latest Trunk and VMmaker > without any install problems so I can start debugging the simulator. > The simulator can no longer simulate images later than 3.7 and I need > it to simulate the Trunk inside the Trunk. > > Cheers, > - Merik Voswinkel > |
In reply to this post by Ang BeePeng
On Sun, Dec 13, 2009 at 12:24:55AM -0800, Ang BeePeng wrote:
> > David T. Lewis wrote: > > > > then evaluate "VMMaker updateFromServer", you will have the latest VMMaker > > along with updates for some other plugins and utilities. > > My VMMaker do not understand "updateFromServer". Anything I miss out? > The #updateFromServer method was added very recently (just last month), but I only mentioned it on the vm-dev list. http://lists.squeakfoundation.org/pipermail/vm-dev/2009-November/003444.html If you use Monticello to load the latest VMMaker from SqueakSource, you should have the update method. Dave |
In reply to this post by K. K. Subramaniam
On Sun, Dec 13, 2009 at 12:00:26PM +0530, K. K. Subramaniam wrote:
> On Sunday 13 December 2009 08:20:14 am David T. Lewis wrote: > > Please load the latest version of VMMaker from > > http://squeaksource.com/VMMaker. The 3.8b6 version that you may have > > gotten from SqueakMap or Universes is now quite out of date. If you load > > the latest version from SqueakSource, then evaluate "VMMaker > > updateFromServer", you will have the latest VMMaker along with updates for > > some other plugins and utilities. > This gives a warning about missing dependencies on KlattResonatorIndices and > FFIPlugin. Proceeding completes the load without any further issues (may need > to start squeak with more memory to avoid running out of memory). Not sure if > the broken dependencies of FFIPlugin may create problems down the road. I see what you mean. The #updateFromServer will load the missing packages that caused the warnings, but it looks like you need to manually re-load the latest VMMaker afterward in order to bring it up to date. After that, the #updateFromServer should take care of future updates to any of these packages. If anyone knows how to fix this in the Monticello configuration, please let me know. This was my first attempt at putting together a configuration map, and I don't really know what I'm doing. > BTW, there is a ready-made image with VMMaker at squeakvm.org/unix/ (search > for unix-*-vmm.tar.gz Thanks. I think Ian keeps copies of his build images for reference. Dave |
On 12/13/09 2:17 PM, "David T. Lewis" <[hidden email]> wrote: > On Sun, Dec 13, 2009 at 12:00:26PM +0530, K. K. Subramaniam wrote: >> On Sunday 13 December 2009 08:20:14 am David T. Lewis wrote: >>> Please load the latest version of VMMaker from >>> http://squeaksource.com/VMMaker. The 3.8b6 version that you may have >>> gotten from SqueakMap or Universes is now quite out of date. If you load >>> the latest version from SqueakSource, then evaluate "VMMaker >>> updateFromServer", you will have the latest VMMaker along with updates for >>> some other plugins and utilities. >> This gives a warning about missing dependencies on KlattResonatorIndices and >> FFIPlugin. Proceeding completes the load without any further issues (may need >> to start squeak with more memory to avoid running out of memory). Not sure if >> the broken dependencies of FFIPlugin may create problems down the road. > > I see what you mean. The #updateFromServer will load the missing packages > that caused the warnings, but it looks like you need to manually re-load > the latest VMMaker afterward in order to bring it up to date. After that, > the #updateFromServer should take care of future updates to any of these > packages. > > If anyone knows how to fix this in the Monticello configuration, please > let me know. This was my first attempt at putting together a configuration > map, and I don't really know what I'm doing. > >> BTW, there is a ready-made image with VMMaker at squeakvm.org/unix/ (search >> for unix-*-vmm.tar.gz > > Thanks. I think Ian keeps copies of his build images for reference. > > Dave For all involved in this , I could easily load Vmaker in the http://ftp.squeak.org/various_images/FunSqueak/FunSqueak3.11-8472-alpha.zip , following advice in this thread . As collateral revamp of old Games and Music, Speech and Talking Heads is inside, so VMMaker have the prerequisites. I call for packages like or use more so I give priority and next week we could have a stable bigger image on top of regular 3.11 alpha. Edgar |
In reply to this post by David T. Lewis
I would like to ask David and Tim a few more questions to - Since you started maintaining VMMaker, did you make any changes to the InterpreterSimulator? I can of course go through all the updates and find it in your notes there, but it appears to me you did not work on InterpreterSimulator? - are you familiar with the methods coerceTo:sim: in class ByteSymbol and the missing #swapBytesFrom:to: in class Bitmap? -There is a methods coerceTo:sim:in class Bitmap though, maybe they where mixed up? Thank you for the help, Merik Voswinkel On Dec 13, 2009, at 5:17 PM, David T. Lewis wrote: > On Sun, Dec 13, 2009 at 12:00:26PM +0530, K. K. Subramaniam wrote: >> On Sunday 13 December 2009 08:20:14 am David T. Lewis wrote: >>> Please load the latest version of VMMaker from >>> http://squeaksource.com/VMMaker. The 3.8b6 version that you may have >>> gotten from SqueakMap or Universes is now quite out of date. If >>> you load >>> the latest version from SqueakSource, then evaluate "VMMaker >>> updateFromServer", you will have the latest VMMaker along with >>> updates for >>> some other plugins and utilities. >> This gives a warning about missing dependencies on >> KlattResonatorIndices and >> FFIPlugin. Proceeding completes the load without any further issues >> (may need >> to start squeak with more memory to avoid running out of memory). >> Not sure if >> the broken dependencies of FFIPlugin may create problems down the >> road. > > I see what you mean. The #updateFromServer will load the missing > packages > that caused the warnings, but it looks like you need to manually re- > load > the latest VMMaker afterward in order to bring it up to date. After > that, > the #updateFromServer should take care of future updates to any of > these > packages. > > If anyone knows how to fix this in the Monticello configuration, > please > let me know. This was my first attempt at putting together a > configuration > map, and I don't really know what I'm doing. > >> BTW, there is a ready-made image with VMMaker at squeakvm.org/unix/ >> (search >> for unix-*-vmm.tar.gz > > Thanks. I think Ian keeps copies of his build images for reference. > > Dave > > |
On Sun, Dec 13, 2009 at 07:24:42PM +0100, Merik Voswinkel wrote:
> > I would like to ask David and Tim a few more questions to > > - Since you started maintaining VMMaker, did you make any changes to > the InterpreterSimulator? I can of course go through all the updates > and find it in your notes there, but it appears to me you did not work > on InterpreterSimulator? Very little work has been on on InterpreterSimulator, so it is very possible that updates are needed here. The interpreter itself has not changed much over the years, with the notable exception of Eliot's changes to add support for block closures. > - are you familiar with the methods coerceTo:sim: in class ByteSymbol > and the missing #swapBytesFrom:to: in class Bitmap? > > -There is a methods coerceTo:sim:in class Bitmap though, maybe they > where mixed up? > I believe that there was some recent discussion of this on the list, although I do not recall the issue right now. These are methods that are meaningful in the translation to C but that would do nothing in simulation mode. Can you explain the issue that you are having? Dave |
On Dec 14, 2009, at 12:33 PM, David T. Lewis wrote: > On Sun, Dec 13, 2009 at 07:24:42PM +0100, Merik Voswinkel wrote: >> >> I would like to ask David and Tim a few more questions to >> >> - Since you started maintaining VMMaker, did you make any changes to >> the InterpreterSimulator? I can of course go through all the updates >> and find it in your notes there, but it appears to me you did not >> work >> on InterpreterSimulator? > > Very little work has been on on InterpreterSimulator, so it is very > possible that updates are needed here. The interpreter itself has not > changed much over the years, with the notable exception of Eliot's > changes to add support for block closures. That is the only change I found as well. Thanks. > >> - are you familiar with the methods coerceTo:sim: in class ByteSymbol >> and the missing #swapBytesFrom:to: in class Bitmap? >> >> -There is a methods coerceTo:sim:in class Bitmap though, maybe they >> where mixed up? >> > > I believe that there was some recent discussion of this on the list, > although I do not recall the issue right now. That was on Januari 22nd 2009, not about a specific issue. > These are methods that > are meaningful in the translation to C but that would do nothing in > simulation mode. Can you explain the issue that you are having? The issues are objects created as ByteSymbol>>coerceTo:sim: that should have been BitMap objects type instead. And a missing BitMap>> #swapBytesFrom:to; method. I think the errors are related to images that are not the same version as the Trunk image (possibly because that is now a closure type image). I'll debug some more and then file a bug in Mantis. Thank you for the info, I saved me a lot of time figuring it out. Cheers, Merik |
Free forum by Nabble | Edit this page |