with manuel oriol we started to play with the idea of a randomTester in Pharo. It poses a lot of questions.
Now doing the first experiments, I found an instance of point (nil@nil) (besides reciprocal broken on Point (0@0) and (Point allInstances select: [:each | each x isNil]) first pointersTo points me to recreateSpecialObjectsArray I looked into the literals and others and I could not find. Each time I execute Smalltalk recreateSpecialObjectsArray I get a new nil@nil objects.... but forcing a GC garbage them. But I still have this nil@nil object around. Does anybody have an idea from where nil@nil could come from? Stef |
of course it is there
newArray at: 34 put: Point new. now could not we have a well initialized point? What is the purpose of this one? Stef On Apr 21, 2011, at 6:39 PM, Stéphane Ducasse wrote: > with manuel oriol we started to play with the idea of a randomTester in Pharo. It poses a lot of questions. > > Now doing the first experiments, I found an instance of point (nil@nil) (besides reciprocal broken on Point (0@0) and > > (Point allInstances select: [:each | each x isNil]) first pointersTo > > points me to recreateSpecialObjectsArray > > I looked into the literals and others and I could not find. > > Each time I execute > Smalltalk recreateSpecialObjectsArray > > I get a new nil@nil objects.... but forcing a GC garbage them. > But I still have this nil@nil object around. > > Does anybody have an idea from where nil@nil could come from? > > Stef > > > > |
From can I can see with me newbie search in the VM, it is not used.
splObj: 33 ObjectMemory class >> initializeSpecialObjectIndices TheInterruptSemaphore := 30. SelectorCannotInterpret := 34. it jumps from 30 to 34. I search places doing "splObj: 33" but I didn't find anything. Maybe the VM guys can confirm or reject this. Cheers Mariano
On Thu, Apr 21, 2011 at 6:42 PM, Stéphane Ducasse <[hidden email]> wrote: of course it is there -- Mariano http://marianopeck.wordpress.com |
In reply to this post by Stéphane Ducasse
On 21 April 2011 18:42, Stéphane Ducasse <[hidden email]> wrote:
> of course it is there > > newArray at: 34 put: Point new. > > now could not we have a well initialized point? > What is the purpose of this one? > probably there was a primitives which doing fast Point allocaiton, and for this it takes an existing point instance as a prototype. I don't know if this slot are used by something .. > Stef > > > On Apr 21, 2011, at 6:39 PM, Stéphane Ducasse wrote: > >> with manuel oriol we started to play with the idea of a randomTester in Pharo. It poses a lot of questions. >> >> Now doing the first experiments, I found an instance of point (nil@nil) (besides reciprocal broken on Point (0@0) and >> >> (Point allInstances select: [:each | each x isNil]) first pointersTo >> >> points me to recreateSpecialObjectsArray >> >> I looked into the literals and others and I could not find. >> >> Each time I execute >> Smalltalk recreateSpecialObjectsArray >> >> I get a new nil@nil objects.... but forcing a GC garbage them. >> But I still have this nil@nil object around. >> >> Does anybody have an idea from where nil@nil could come from? >> >> Stef >> >> >> >> > > > -- Best regards, Igor Stasenko AKA sig. |
On Apr 21, 2011, at 19:32 , Igor Stasenko wrote: > On 21 April 2011 18:42, Stéphane Ducasse <[hidden email]> wrote: >> of course it is there >> >> newArray at: 34 put: Point new. >> >> now could not we have a well initialized point? >> What is the purpose of this one? >> > > probably there was a primitives which doing fast Point allocaiton, and > for this it takes an existing point instance as a prototype. > I don't know if this slot are used by something .. It doesn't seem so. I (quickly) checked the VM sources and couldn't find an access to this array slot... Adrian |
In reply to this post by Igor Stasenko
I would like at least to have a well initialized point because it can break tools using the system to generate
examples. Stef On Apr 21, 2011, at 7:32 PM, Igor Stasenko wrote: > probably there was a primitives which doing fast Point allocaiton, and > for this it takes an existing point instance as a prototype. > I don't know if this slot are used by something .. |
Free forum by Nabble | Edit this page |