[Vm-dev] VM Maker: VMMaker.oscog-eem.1026.mcz

Eliot Miranda eliot.miranda at gmail.com
Sat Jan 24 02:45:10 UTC 2015


Hi David,

On Thu, Jan 22, 2015 at 5:48 PM, David T. Lewis <lewis at mail.msen.com> wrote:

>
> On Fri, Jan 23, 2015 at 12:42:03AM +0000, commits at source.squeak.org wrote:
> >
> > Eliot Miranda uploaded a new version of VMMaker to project VM Maker:
> > http://source.squeak.org/VMMaker/VMMaker.oscog-eem.1026.mcz
> >
> > ==================== Summary ====================
> >
> > Name: VMMaker.oscog-eem.1026
> > Author: eem
> > Time: 22 January 2015, 4:41:41.71 pm
> > UUID: 58161d66-0b23-4356-89f1-8e951fd1a1d9
> > Ancestors: VMMaker.oscog-eem.1025
> >
> > Fix type regression in primitiveBitShift due to 64-bits.
> >
>
> Is there a new issue here? As far as I know, this primitive has been
> working properly for all combinations of 32/64 bit host and image for
> quite a while. Is this something new related to the extended range of
> small integers on Spur? Just curious.
>
> Yes, this is a new issue.  I rewrote the primitive so that it worked with
Spur 61-bit SmallIntegers, hence I replaced


integerReceiver := self popPos32BitInteger.

with
objectMemory wordSize = 4
ifTrue:
[integerReceiver := self positive32BitValueOf: integerReceiver]
ifFalse:
[integerReceiver := self signed64BitValueOf: integerReceiver].

and in the Cog VMMaker that changed the type computed for integerReceiver.
I just added a declaration to override the type inferencer.  So it's a new
issue, and Cog branch only.
-- 
best,
Eliot
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.squeakfoundation.org/pipermail/vm-dev/attachments/20150123/d170e915/attachment.htm


More information about the Vm-dev mailing list