[Vm-dev] [spur] literalsManager class shouldn't be an ancilliary class?

Eliot Miranda eliot.miranda at gmail.com
Wed Jun 10 14:37:48 UTC 2015

Hi Esteban,  

    things are going to be broken fir a few more commits yet.  The out-of-line literals changes are bigger than I'd anticipated (easy to add to a one pass JIT, not so easy to a multi pass one like Cog).  I expect to get the code working today and fix the VM generation/simulation issues tomorrow.  Fingers xed.

Eliot (phone)

On Jun 10, 2015, at 3:56 AM, Esteban Lorenzano <estebanlm at gmail.com> wrote:

> Hi, 
> (btw… since last week I have an hourly update from VMMaker.oscog, building the VMs for Pharo, you can check it here: https://ci.inria.fr/pharo/view/5.0-VM-Spur/ … I think since we are not merging anymore, and because of that we are not a de facto fork anymore, this CI can be useful for regular development too…)
> Now, the problem :)
> latest commit introduced literalsManager but since then spur vm is not compiled, because (I think), literalsManagerClass should be ancilliary class. Something like this: 
> ancilliaryClasses: options
> 	ProcessorClass ifNil:
> 		[Cogit initializeMiscConstants].
> 	^{	CogMethodZone.
> 		CogAbstractInstruction.
> 		self activeCompilerClass.
> 		self activeCompilerClass literalsManagerClass.
> 		CogBlockStart.
> 		CogBytecodeDescriptor.
> 		CogBytecodeFixup.
> 		CogInstructionAnnotation.
> 		CogPrimitiveDescriptor.
> 		CogBlockMethod.
> 		CogMethod },
> 	((options at: #NewspeakVM ifAbsent: [false])
> 		ifTrue: [{NewspeakCogMethod. NSSendCache}]
> 		ifFalse: [#()])
> but well, that causes some other problems, like I need to rename a couple of variables and include literalsManager into #isNonArgumentImplicitReceiverVariableName: 
> then VM builds, but does not seems to work fine either, so I suppose I’m doing something fundamentally wrong :)
> Any idea?
> thanks,
> Esteban
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.squeakfoundation.org/pipermail/vm-dev/attachments/20150610/7f57b2e5/attachment-0001.htm

More information about the Vm-dev mailing list