Hi Eliot --

> But lookup: *always* interns:

No, it does not. You should use #lookup: instead. It has the very same semantics as #findInterned:.

In any case, I will re-load 60Deprecated again into Trunk so that you can keep your code unchanged. We can then discuss the benefit of having two identical methods for the same thing.

#lookup: has the better wording and hides implementation details.
#findInterned: leaks the implementation detail of "internalization"

#asSymbol always interns. #lookup: does not.

Best,
Marcel
Am 15.06.2022 00:27:54 schrieb Eliot Miranda <eliot.miranda at gmail.com>:
Hi All,

  I've just run into what I think is a bad mistake.  My understanding of Symbol class>>findInterned: is that it answers a Symbol if its (string) argument already exists as a Symbol, and answers nil otherwise.  Indeed looking at Collections-mt.935, 13 April 2021, this is the case:

Symbol class>>findInterned:aString

    self hasInterned:aString ifTrue:[:symbol| ^symbol].
    ^nil.


In Collections-nice.936,13 April 2021 it was deprecated, but the new method in 60Deprecated-nice.95, 13 April 2021 reads:

Symbol class>>findInterned:aString
   self flag: #deprecated. "use lookup: instead, they are synonym"
   ^self lookup: aString

But lookup: *always* interns:

Symbol intern: 'union {
sqIntptr_t valword;
struct { int low, high; } valleint64;
struct { int high, low; } valbeint64;
double valflt64;
struct { void *addr; sqIntptr_t size; } valstruct;
}' #'union {
sqIntptr_t valword;
struct { int low, high; } valleint64;
struct { int high, low; } valbeint64;
double valflt64;
struct { void *addr; sqIntptr_t size; } valstruct;
}' 


So I think it is *urgent* that we undeprecate findInterne3d: and restore its old meaning.  I apologixze for not finding this earlier.  i just ran VMMaker today after an update and that triggered the bug for the first time.  I'm guessing the update flushed symbols.

I'm going to make this change now.

_,,,^..^,,,_

best, Eliot
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20220615/95be990e/attachment.html>