Smalltalk › Squeak › Squeak - Dev

Updating from trunk gets stuck forever at "Initializing ReleaseBuilder-cmm.51

‹ Previous Topic Next Topic ›
 
classic Classic list List threaded Threaded
♦
♦
 7 messages Options
CdAB63
Reply | Threaded
Open this post in threaded view
♦
♦
|

Updating from trunk gets stuck forever at "Initializing ReleaseBuilder-cmm.51

CdAB63
Ok, it's generating a SequeakDebug.log (though no debug browser opens) with:

[] in WorldState>>displayWorldSafeles:  subscript is out of bounds: 513'
BlockClosure>>ermultiCanvas:    nil669hicAlarmQueue()oryAt: -> a PluggableSystemWindow(2972))(a Plugg...etc...
[] in BlReceiverouterContext: riWorldState>>displayWorldSafely:fely:ounds: 513
<<error Receiverreceiver: il: ri[closure]sure>>#on:do: "a CompiledMethod(1880)")
MethodCoReceiversender: nce variBlockClosure>>ifEis out of bounds: 513al:handleSignal:
Error(ExArgumentsignalerText: y 'subscript i bounds: 513:do: 513'orldSafely:0)")
Bitmap(OReect)>>at:put:ance variables: es: is out of bounds: 513'
SortedCoReceiver: a SortedCollection(0.0->Color white 0.0->(lor r: 1.0 g: 1.0 b: 1.0) . 1.0->(Color ...etc...
<<error Receivercormal: n:  variniles: es: n GradientFillStyle>>computePixelRampOfSize:olor r: 1...etc...
[] in GrReceiverorganization: le('access#('colorRamp' 'pixelRamp' 'radial' 'isTranslucent')ientFill...etc...
LRUCacheArgumentelement: 0.0->Co{{0.0->(Color r: 1.0 g: 1.0 b: 1.0) . 1.0->() . 1.0->(Color r: 1....etc......
GradientReceiverisTran: p:  varinil0->Color white . 0.0->(Color r: 1.0 g: 1.0 b: 1.0) . 1.0->(Color ...etc......
GraCache>>at:tFillStyle class>>initPixelRampCache
[] in PluggableSystemWindow(Morph)>>flStyle:borderStyle:or:ransform:
FormCanvas>>roundCornersOf:ind:submorphs:easOn:easOn:easOn:erWidth:corners:
[] in BlockClosure>>newProcessrocess:ding:

An interesting thing is that squeak keeps building & destroying SqueakDebug.log (meaning, when I follow the file it gets emptied & filled again with same content).

-------- Mensagem original --------
Assunto: Updating from trunk gets stuck forever at "Initializing ReleaseBuilder-cmm.51
Data: Wed, 05 Jan 2011 16:41:26 -0200
De: Casimiro de Almeida Barreto [hidden email]
Para: The general-purpose Squeak developers list [hidden email]


I got a couple of trunk images about 3 weeks old. When I update both images get stuck at "Initializing ReleaseBuilder-cmm.51". No use doing Alt+.
Progress bar stuck at 0%
Processor usage (single core) > 90% almost all all the time.

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND           
27935 casimiro  20   0 1037m 143m 2884 R 96.4  7.3   7:00.37 squeakvm          
 1736 root      20   0  113m  40m  18m S  5.7  2.1   6:00.59 X                 
 2402 casimiro   9 -11  160m 5056 3976 S  3.8  0.3   3:38.41 pulseaudio




SqueakDebug.log (9K) Download Attachment
update-nice.169-squeak.image.log (9K) Download Attachment
laza
Reply | Threaded
Open this post in threaded view
♦
♦
|

Re: Updating from trunk gets stuck forever at "Initializing ReleaseBuilder-cmm.51

laza
Same here on Ubuntu 10.10. On Windows7 the same image updates without a hitch.

Alex

2011/1/5 Casimiro de Almeida Barreto <[hidden email]>
Ok, it's generating a SequeakDebug.log (though no debug browser opens) with:

[] in WorldState>>displayWorldSafeles:  subscript is out of bounds: 513'
BlockClosure>>ermultiCanvas:    nil669hicAlarmQueue()oryAt: -> a PluggableSystemWindow(2972))(a Plugg...etc...
[] in BlReceiverouterContext: riWorldState>>displayWorldSafely:fely:ounds: 513
<<error Receiverreceiver: il: ri[closure]sure>>#on:do: "a CompiledMethod(1880)")
MethodCoReceiversender: nce variBlockClosure>>ifEis out of bounds: 513al:handleSignal:
Error(ExArgumentsignalerText: y 'subscript i bounds: 513:do: 513'orldSafely:0)")
Bitmap(OReect)>>at:put:ance variables: es: is out of bounds: 513'
SortedCoReceiver: a SortedCollection(0.0->Color white 0.0->(lor r: 1.0 g: 1.0 b: 1.0) . 1.0->(Color ...etc...
<<error Receivercormal: n:  variniles: es: n GradientFillStyle>>computePixelRampOfSize:olor r: 1...etc...
[] in GrReceiverorganization: le('access#('colorRamp' 'pixelRamp' 'radial' 'isTranslucent')ientFill...etc...
LRUCacheArgumentelement: 0.0->Co{{0.0->(Color r: 1.0 g: 1.0 b: 1.0) . 1.0->() . 1.0->(Color r: 1....etc......
GradientReceiverisTran: p:  varinil0->Color white . 0.0->(Color r: 1.0 g: 1.0 b: 1.0) . 1.0->(Color ...etc......
GraCache>>at:tFillStyle class>>initPixelRampCache
[] in PluggableSystemWindow(Morph)>>flStyle:borderStyle:or:ransform:
FormCanvas>>roundCornersOf:ind:submorphs:easOn:easOn:easOn:erWidth:corners:
[] in BlockClosure>>newProcessrocess:ding:

An interesting thing is that squeak keeps building & destroying SqueakDebug.log (meaning, when I follow the file it gets emptied & filled again with same content).

-------- Mensagem original --------
Assunto: Updating from trunk gets stuck forever at "Initializing ReleaseBuilder-cmm.51
Data: Wed, 05 Jan 2011 16:41:26 -0200
De: Casimiro de Almeida Barreto [hidden email]
Para: The general-purpose Squeak developers list [hidden email]


I got a couple of trunk images about 3 weeks old. When I update both images get stuck at "Initializing ReleaseBuilder-cmm.51". No use doing Alt+.
Progress bar stuck at 0%
Processor usage (single core) > 90% almost all all the time.

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND           
27935 casimiro  20   0 1037m 143m 2884 R 96.4  7.3   7:00.37 squeakvm          
 1736 root      20   0  113m  40m  18m S  5.7  2.1   6:00.59 X                 
 2402 casimiro   9 -11  160m 5056 3976 S  3.8  0.3   3:38.41 pulseaudio
Chris Muller-3
Reply | Threaded
Open this post in threaded view
♦
♦
|

Re: Updating from trunk gets stuck forever at "Initializing ReleaseBuilder-cmm.51

Chris Muller-3
In reply to this post by CdAB63
Well, it could be the new FloatMathPlugin changes.  It could be something bad in your image.  What VM are you using?

If you press Alt+. in the image, do you see a bunch of RedX morphs?  So it probably couldn't draw the Transcript or other windows you had open, why don't you close them and try again.

On Wed, Jan 5, 2011 at 2:08 PM, Casimiro de Almeida Barreto <[hidden email]> wrote:
Ok, it's generating a SequeakDebug.log (though no debug browser opens) with:

[] in WorldState>>displayWorldSafeles:  subscript is out of bounds: 513'
BlockClosure>>ermultiCanvas:    nil669hicAlarmQueue()oryAt: -> a PluggableSystemWindow(2972))(a Plugg...etc...
[] in BlReceiverouterContext: riWorldState>>displayWorldSafely:fely:ounds: 513
<<error Receiverreceiver: il: ri[closure]sure>>#on:do: "a CompiledMethod(1880)")
MethodCoReceiversender: nce variBlockClosure>>ifEis out of bounds: 513al:handleSignal:
Error(ExArgumentsignalerText: y 'subscript i bounds: 513:do: 513'orldSafely:0)")
Bitmap(OReect)>>at:put:ance variables: es: is out of bounds: 513'
SortedCoReceiver: a SortedCollection(0.0->Color white 0.0->(lor r: 1.0 g: 1.0 b: 1.0) . 1.0->(Color ...etc...
<<error Receivercormal: n:  variniles: es: n GradientFillStyle>>computePixelRampOfSize:olor r: 1...etc...
[] in GrReceiverorganization: le('access#('colorRamp' 'pixelRamp' 'radial' 'isTranslucent')ientFill...etc...
LRUCacheArgumentelement: 0.0->Co{{0.0->(Color r: 1.0 g: 1.0 b: 1.0) . 1.0->() . 1.0->(Color r: 1....etc......
GradientReceiverisTran: p:  varinil0->Color white . 0.0->(Color r: 1.0 g: 1.0 b: 1.0) . 1.0->(Color ...etc......
GraCache>>at:tFillStyle class>>initPixelRampCache
[] in PluggableSystemWindow(Morph)>>flStyle:borderStyle:or:ransform:
FormCanvas>>roundCornersOf:ind:submorphs:easOn:easOn:easOn:erWidth:corners:
[] in BlockClosure>>newProcessrocess:ding:

An interesting thing is that squeak keeps building & destroying SqueakDebug.log (meaning, when I follow the file it gets emptied & filled again with same content).

-------- Mensagem original --------
Assunto: Updating from trunk gets stuck forever at "Initializing ReleaseBuilder-cmm.51
Data: Wed, 05 Jan 2011 16:41:26 -0200
De: Casimiro de Almeida Barreto [hidden email]
Para: The general-purpose Squeak developers list [hidden email]


I got a couple of trunk images about 3 weeks old. When I update both images get stuck at "Initializing ReleaseBuilder-cmm.51". No use doing Alt+.
Progress bar stuck at 0%
Processor usage (single core) > 90% almost all all the time.

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND           
27935 casimiro  20   0 1037m 143m 2884 R 96.4  7.3   7:00.37 squeakvm          
 1736 root      20   0  113m  40m  18m S  5.7  2.1   6:00.59 X                 
 2402 casimiro   9 -11  160m 5056 3976 S  3.8  0.3   3:38.41 pulseaudio
CdAB63
Reply | Threaded
Open this post in threaded view
♦
♦
|

Re: Updating from trunk gets stuck forever at "Initializing ReleaseBuilder-cmm.51

CdAB63
Em 05-01-2011 21:42, Chris Muller escreveu:
Well, it could be the new FloatMathPlugin changes.  It could be something bad in your image.  What VM are you using?
VM: 4.0.3-2202
As downloaded from www.squeak.org...

If you press Alt+. in the image, do you see a bunch of RedX morphs?  So it probably couldn't draw the Transcript or other windows you had open, why don't you close them and try again.
Interesting enough, when gets lost squeakvm doesn't give a dime about Alt+. The only way of stopping it is to give a nice & old killall - KILL squeakvm and to give a look at the log files...

CdAB

On Wed, Jan 5, 2011 at 2:08 PM, Casimiro de Almeida Barreto <[hidden email]> wrote:
Ok, it's generating a SequeakDebug.log (though no debug browser opens) with:

[] in WorldState>>displayWorldSafeles:  subscript is out of bounds: 513'
BlockClosure>>ermultiCanvas:    nil669hicAlarmQueue()oryAt: -> a PluggableSystemWindow(2972))(a Plugg...etc...
[] in BlReceiverouterContext: riWorldState>>displayWorldSafely:fely:ounds: 513
<<error Receiverreceiver: il: ri[closure]sure>>#on:do: "a CompiledMethod(1880)")
MethodCoReceiversender: nce variBlockClosure>>ifEis out of bounds: 513al:handleSignal:
Error(ExArgumentsignalerText: y 'subscript i bounds: 513:do: 513'orldSafely:0)")
Bitmap(OReect)>>at:put:ance variables: es: is out of bounds: 513'
SortedCoReceiver: a SortedCollection(0.0->Color white 0.0->(lor r: 1.0 g: 1.0 b: 1.0) . 1.0->(Color ...etc...
<<error Receivercormal: n:  variniles: es: n GradientFillStyle>>computePixelRampOfSize:olor r: 1...etc...
[] in GrReceiverorganization: le('access#('colorRamp' 'pixelRamp' 'radial' 'isTranslucent')ientFill...etc...
LRUCacheArgumentelement: 0.0->Co{{0.0->(Color r: 1.0 g: 1.0 b: 1.0) . 1.0->() . 1.0->(Color r: 1....etc......
GradientReceiverisTran: p:  varinil0->Color white . 0.0->(Color r: 1.0 g: 1.0 b: 1.0) . 1.0->(Color ...etc......
GraCache>>at:tFillStyle class>>initPixelRampCache
[] in PluggableSystemWindow(Morph)>>flStyle:borderStyle:or:ransform:
FormCanvas>>roundCornersOf:ind:submorphs:easOn:easOn:easOn:erWidth:corners:
[] in BlockClosure>>newProcessrocess:ding:

An interesting thing is that squeak keeps building & destroying SqueakDebug.log (meaning, when I follow the file it gets emptied & filled again with same content).

-------- Mensagem original --------
Assunto: Updating from trunk gets stuck forever at "Initializing ReleaseBuilder-cmm.51
Data: Wed, 05 Jan 2011 16:41:26 -0200
De: Casimiro de Almeida Barreto [hidden email]
Para: The general-purpose Squeak developers list [hidden email]


I got a couple of trunk images about 3 weeks old. When I update both images get stuck at "Initializing ReleaseBuilder-cmm.51". No use doing Alt+.
Progress bar stuck at 0%
Processor usage (single core) > 90% almost all all the time.

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND           
27935 casimiro  20   0 1037m 143m 2884 R 96.4  7.3   7:00.37 squeakvm          
 1736 root      20   0  113m  40m  18m S  5.7  2.1   6:00.59 X                 
 2402 casimiro   9 -11  160m 5056 3976 S  3.8  0.3   3:38.41 pulseaudio



laza
Reply | Threaded
Open this post in threaded view
♦
♦
|

Re: Updating from trunk gets stuck forever at "Initializing ReleaseBuilder-cmm.51

laza
2011/1/6 Casimiro de Almeida Barreto <[hidden email]>
VM: 4.0.3-2202
As downloaded from www.squeak.org...

If you are desperate you could install a cog VM for linux. Here [1] you will find the latest version.

You need to do a

rm coglinux/lib/squeak/3.9-7/UUIDPlugin

so the beloved UUID bug won't bite you.

After that you should be able to do a

coglinux/bin/squeak yourtrunk.image

and do a successful update.
CdAB63
Reply | Threaded
Open this post in threaded view
♦
♦
|

Re: Updating from trunk gets stuck forever at "Initializing ReleaseBuilder-cmm.51

CdAB63
Em 06-01-2011 13:36, Alexander Lazarević escreveu:
2011/1/6 Casimiro de Almeida Barreto <[hidden email]>
VM: 4.0.3-2202
As downloaded from www.squeak.org...

If you are desperate you could install a cog VM for linux. Here [1] you will find the latest version.
Yes, I'm using a CogVM and it works OK, except for things concerning OpenGL (when starting a Wonderland it says me that I don't have graphics acceleration & things'll take forever to run (which turns to be painfully true). When I use standard VM GL works all right. BTW, I'm using a NVIDIA GPU (7300) and NVIDIA device drivers (because nouveau trully sucks).

To avoid this problem I'm using Nicolay Suslov stuff. But then I cannot update without the risk of breaking things I don't understand very well...

BTW, if I use standard 4.1 image (as downloaded from www.squeak.org) it will cause a core dump (Fedora) in CogVM. It's necessary to update such image with standard VM and then use CogVM on the updated image.

You need to do a

rm coglinux/lib/squeak/3.9-7/UUIDPlugin

so the beloved UUID bug won't bite you.

After that you should be able to do a

coglinux/bin/squeak yourtrunk.image

and do a successful update.



CdAB63
Reply | Threaded
Open this post in threaded view
♦
♦
|

Re: Updating from trunk gets stuck forever at "Initializing ReleaseBuilder-cmm.51

CdAB63
In reply to this post by laza
Em 06-01-2011 13:36, Alexander Lazarević escreveu:
2011/1/6 Casimiro de Almeida Barreto <[hidden email]>
VM: 4.0.3-2202
As downloaded from www.squeak.org...

If you are desperate you could install a cog VM for linux. Here [1] you will find the latest version.
Just installed a 4.2beta image. Guess what: updates without a problem.

Anyways it would be interesting to know what the glitch is (to avoid it happening in the future).

I think the problem must be around ReleaseBuilder but as I'm unable to stop execution to iterate over the problem I have no ways of being sure. What I know is that squeak tries to open a debugger brower and fails over and over. IMHO, it should be a failsafe procedure in cases like this (moreover because Alt+. won't work).

CdAB

You need to do a

rm coglinux/lib/squeak/3.9-7/UUIDPlugin

so the beloved UUID bug won't bite you.

After that you should be able to do a

coglinux/bin/squeak yourtrunk.image

and do a successful update.



Free forum by Nabble Edit this page