On 3/22/19, patrick.rein@hpi.uni-potsdam.de patrick.rein@hpi.uni-potsdam.de wrote:
Sounds good to me!
How about the following: I would deprecate #asMutator/#isMutator in favor of #asSimpleSetter and #isSimpleSetter. At the same time I would move the #asSimpleSetter up from Symbol to String where #asSetterSelector already resides. Finally, I would move both pairs of selectors to the Tools (or System or Compiler?) package (I would rather not keep them scattered across MorphicExtras and Etoys as common selectors would then also be used by System).
+1
Bests, Patrick
On Thu, Mar 21, 2019 at 1:03 PM Chris Muller asqueaker@gmail.com wrote:
My main objection is about growing the "mutator" nomenclature further instead of trimming it back.
+1
I'm working on a GraphQL framework. GraphQL has a first-class notion of "mutations".
https://graphql.org/learn/queries/
Currently, we have almost zero mention of the word "mutator" in the image. I would appreciate if, instead of germinating the overloading the word, we uproot it and keep the nomenclature for this Smalltalk tooling confined to "getter" and "setter" that we have now.
+1
So, instead of #asMutator and #isMutator, how about #asSimpleSetter and #isSimpleSetter?
+1
We can end up removing a method instead of adding one. We already have many existing methods which are dealing with this conversion, all of which use the traditional Smalltalk / developer linguisitic of "setter". Browsing through all the "setter" nomenclature we have in a message names browser, including
Utilities class>>#getterSelectorFor: Utilities class>>#setterSelectorFor: Utilities class>>#simpleSetterFor:
which instruct Etoys users to use existing methods on String, #asSetterSelector or #asSimpleSetter.
Of special note is SyntaxMorph>>#isStandardSetterKeyword: which, by its use of a utility-method implementation, acknowledges that this behavior is _application-specific_.
I, myself, would just write:
someSelector asSimpleSetter = someSelector "same as isMutator / isSimpleSetter"
instead of #isSimpleSetter, but can appreciate both styles. They both better avoid the awkwardness of "mutator" as a general, reliable behavior when:
#at: isMutator "true"
while
#at:put: isMutator "false"
+1000. Chris is right. isMutator is a mistake.
Best, Chris
On Thu, Mar 21, 2019 at 1:16 AM patrick.rein@hpi.uni-potsdam.de wrote:
Which one are you refering to? I only see Symbol>>#asMutator in Collections in the method category converting. It is also used in a few other places beyond generating instance variable accessors.
Bests, Patrick
#asMutator: is for "generate instVar accessors" code-generation, and that's why it's in the Tools package.
IMHO, if #isMutator belongs anywhere, it isn't in Collections.
Regards, Chris
On Wed, Mar 20, 2019 at 3:08 PM Chris Muller asqueaker@gmail.com wrote:
Assuming a single-keyword is a mutator? Object>>#at: is not a mutator. #indexOf: is not. #first: is not. Obviously many more..
If API balance is the goal of this, perhaps #asMutator should be removed instead. This feels application-specific...
On Wed, Mar 20, 2019 at 2:05 PM commits@source.squeak.org wrote:
Patrick Rein uploaded a new version of Collections to project The Trunk: http://source.squeak.org/trunk/Collections-pre.822.mcz
==================== Summary ====================
Name: Collections-pre.822 Author: pre Time: 20 March 2019, 8:05:20.383677 pm UUID: 483c4461-cee4-4a4f-82d3-fbc03e7201cc Ancestors: Collections-dtl.821
Adds #isMutator to Symbol which is analogous to asMutator.
=============== Diff against Collections-dtl.821 ===============
Item was added:
- ----- Method: Symbol>>isMutator (in category 'testing') -----
- isMutator
- ^ self isKeyword and: [self numArgs = 1]!
-- _,,,^..^,,,_ best, Eliot --000000000000004a690584a05ffa--