Two-way become exists in VA Smalltalk already

Just as a curiosity, simulating a two-way become with a one-way become
is a lot more efficient...
    aPrime := A shallowCopy.
    bPrime := B shallowCopy.
    A oneWayBecome: bPrime.
    B oneWayBecome: aPrime.
This one's cute because both objects are replaced with copies, and nobody
knows they're not the originals (like InvasionOfBodySnatchers).  One of
the copies can be avoided, but it's not like that's going to be a major
bottleneck (unless you're fiddling with huge collections).

[Instantiations, please be sure to preserve this functionality if you don't implement an official two-way become.]

In porting GBS functionality from VisualWorks to VA Smalltalk, I needed two-way become support. Mark van Gulik posted an algorithm by which one can simulate two-way become using a one-way become (see below), but it required two copy operations and two separate become operations.

It turns out the #multiBecome: will do a two-way become if you convert a,b to b,a.

Example:

| date time dateHolder timeHolder |
date := Date today.
time := Time now.
dateHolder := Array with: date.
timeHolder := Array with: time.
(Array with: date with: time) multiBecome: (Array with: time with: date).
^Array with: dateHolder "now time" with: timeHolder "now date"

Mark's explanation of how to do it using brute force:

Just as a curiosity, simulating a two-way become with a one-way become
is a lot more efficient...
    aPrime := A shallowCopy.
    bPrime := B shallowCopy.
    A oneWayBecome: bPrime.
    B oneWayBecome: aPrime.
This one's cute because both objects are replaced with copies, and nobody
knows they're not the originals (like InvasionOfBodySnatchers).  One of
the copies can be avoided, but it's not like that's going to be a major
bottleneck (unless you're fiddling with huge collections).

Hi Richard,

Yes, we plan to preserve the functionality we have.

For next release we will review basicHash swapping.  I would like to understand the rationale behind the decision that was made originally not to do it...and what happens (if anything) if we do.

I found this old thread (<a href="http://forum.world.st/identityHash-and-become-question-td1404098.html" target="_blank" onmousedown="this.href='http://www.google.com/url?q\75http%3A%2F%2Fforum.world.st%2FidentityHash-and-become-question-td1404098.html\46sa\75D\46sntz\0751\46usg\75AFQjCNHtjzq0jpKWt2ZvMCMxo8vB5jN_AA';return true;" onclick="this.href='http://www.google.com/url?q\75http%3A%2F%2Fforum.world.st%2FidentityHash-and-become-question-td1404098.html\46sa\75D\46sntz\0751\46usg\75AFQjCNHtjzq0jpKWt2ZvMCMxo8vB5jN_AA';return true;">http://forum.world.st/identityHash-and-become-question-td1404098.html) as a start for research.  I don't have much time at the moment to dig deeper, but as I said, it will be revisited.
If you have additional thoughts on the matter, I will be sure to include that in the discussion.  Thanks

On Wednesday, October 1, 2014 12:25:27 PM UTC-7, Seth Berman wrote:

Hi Richard,

Yes, we plan to preserve the functionality we have.

For next release we will review basicHash swapping.  I would like to understand the rationale behind the decision that was made originally not to do it...and what happens (if anything) if we do.

I found this old thread (<a href="http://forum.world.st/identityHash-and-become-question-td1404098.html" target="_blank" onmousedown="this.href='http://www.google.com/url?q\75http%3A%2F%2Fforum.world.st%2FidentityHash-and-become-question-td1404098.html\46sa\75D\46sntz\0751\46usg\75AFQjCNHtjzq0jpKWt2ZvMCMxo8vB5jN_AA';return true;" onclick="this.href='http://www.google.com/url?q\75http%3A%2F%2Fforum.world.st%2FidentityHash-and-become-question-td1404098.html\46sa\75D\46sntz\0751\46usg\75AFQjCNHtjzq0jpKWt2ZvMCMxo8vB5jN_AA';return true;">http://forum.world.st/identityHash-and-become-question-td1404098.html) as a start for research.  I don't have much time at the moment to dig deeper, but as I said, it will be revisited.
If you have additional thoughts on the matter, I will be sure to include that in the discussion.  Thanks

Thanks, Seth. That was a very interesting and - I think - enlightening reference. It may well be that my changing their identity hashes when I swap them is actually creating problems for me. I'll have to check this out carefully. [I don't know about anyone else, but for me, "reasoning" about #become: is very difficult! :-)]

On Wednesday, October 1, 2014 12:25:27 PM UTC-7, Seth Berman wrote:

For next release we will review basicHash swapping.  I would like to understand the rationale behind the decision that was made originally not to do it...and what happens (if anything) if we do.

I found this old thread (<a href="http://forum.world.st/identityHash-and-become-question-td1404098.html" target="_blank" onmousedown="this.href='http://www.google.com/url?q\75http%3A%2F%2Fforum.world.st%2FidentityHash-and-become-question-td1404098.html\46sa\75D\46sntz\0751\46usg\75AFQjCNHtjzq0jpKWt2ZvMCMxo8vB5jN_AA';return true;" onclick="this.href='http://www.google.com/url?q\75http%3A%2F%2Fforum.world.st%2FidentityHash-and-become-question-td1404098.html\46sa\75D\46sntz\0751\46usg\75AFQjCNHtjzq0jpKWt2ZvMCMxo8vB5jN_AA';return true;">http://forum.world.st/identityHash-and-become-question-td1404098.html) as a start for research.  I don't have much time at the moment to dig deeper, but as I said, it will be revisited.

... It may well be that my changing their identity hashes when I swap them is actually creating problems for me. I'll have to check this out carefully. ...