The Inbox: Kernel-eem.1191.mcz

Eliot Miranda uploaded a new version of Kernel to project The Inbox:
http://source.squeak.org/inbox/Kernel-eem.1191.mcz

==================== Summary ====================

Name: Kernel-eem.1191
Author: eem
Time: 11 October 2018, 10:52:37.059231 pm
UUID: 465e029d-66ff-492d-a465-79f2867d2d8b
Ancestors: Kernel-cmm.1190

Remove adoptInstance:'s fallback on primitiveChangeClassTo:, which obscures potential failures on Spur.  Committing to Inbox; this isn't urgent but I want it in soon enough.

=============== Diff against Kernel-cmm.1190 ===============

Item was changed:
  ----- Method: Behavior>>adoptInstance: (in category 'instance creation') -----
  adoptInstance: anInstance
+ "Change the class of anInstance to the receiver.
+ Primitive. Change the class of the argument anInstance into the receiver, provided
+ that the format of the receiver matches the format of the argument's class.
+ Fail if the argument is an immediate, or when the pointerness of the receiver is different
+ from the pointerness of the argument, or when the receiver is a fixed pointer class and
+ anInstance's size differs from the size that an instance of the receiver should have,
+ or when anInstance's size is not an integer multiple of the receiver's unit size."
- "Change the class of anInstance to me.
- Primitive (found in Cog and new VMs)  follows the same rules as primitiveChangeClassTo:, but returns the class rather than the modified instance"
 
  <primitive: 160 error: ec>
+ ^self primitiveFailed!
- anInstance primitiveChangeClassTo: self basicNew.
- ^self!

Hi Eliot,

On Fri, Oct 12, 2018 at 05:52:51AM +0000, [hidden email] wrote:
By any chance do you have a reproducible code snippet that shows the
original problem? I realize this may not be easy if it involves Form
and endianness and so forth.

If so I will volunteer to turn it into a unit test. That is a good
way to guard against regression, and also easily identify the problem
when a user is running an older VM that may not have the fix.

Dave