Initial Record/Tuple standard library spec #131

rickbutton · 2020-06-09T22:03:15Z

NOTE: This is an early and incomplete draft. Comments welcome.

package.json

spec/immutable-data-structures.html

ljharb

It really seems like we'll have to rename the existing concept in the spec of "Record" before landing this proposal. What are the thoughts/plans around that?

spec/immutable-data-structures.html

rricard · 2020-06-12T18:29:57Z

It really seems like we'll have to rename the existing concept in the spec of "Record" before landing this proposal. What are the thoughts/plans around that?

I agree, we're tracking that in #96 but maybe we need to work on that earlier because the html output is very misleading right now

littledan

This all looks generally reasonable, though I haven't reviewed every word extremely closely. I guess you're positing two operations here:

CreateTupleFromList
This throws an exception if List has a non-primitive in it.
I think we might want to move responsibility for checking for non-primitives to the caller. And in general, do this more eagerly, e.g., for map, check each time something is output.
In the context of Record and Tuple object model spec #132 , we might write this as, "A new Tuple value whose [[Sequence]] is list." (The JS spec generally tries to avoid unnecessary indirection for things like construction.)
CreateListFromTuple
This needs returns a fresh copy of the list, as sometimes the caller modifies it.
In the context of Record and Tuple object model spec #132, we may instead write this as "tuple's [[Sequence]] value". In this case, we'd want to make a fresh copy of the list before mutating it.

spec/immutable-data-structures.html

littledan · 2020-06-16T21:37:14Z

spec/immutable-data-structures.html

+          1. Let _items_ be a List whose first element is _T_ and whose subsequent element are, in left to right order, the arguments that were passed to this function invocation.
+          1. Repeat, while _items_ is not empty,
+            1. Remove the first element from _items_ and let _E_ be the value of the element.
+            1. Let _spreadable_ be ? IsConcatSpreadable(_E_).


I guess we should make sure that Tuples are concat spreadable (either by updating that algorithm or giving them the symbol; I prefer the former).

It's a protocol; it's fine to have default behavior here, but I think the symbol needs to exist and have its value respected.

I'm not sure what you're trying to say here. Nothing in the JS spec has that symbol defined. I'm not saying that we would stop querying the symbol, just that it would have this default behavior built into the IsConcatSpreadable algorithm, like Arrays do.

rickbutton · 2020-06-16T21:52:04Z

@littledan on CreateListFromTuple / CreateTupleFromList are you suggesting that I replace the usag e of this abstract op with the verbiage you mentioned, or define them using that verbiage? In either case, seems fine to me.

littledan · 2020-06-16T21:54:13Z

@rickbutton I'm suggesting you replace the usages with the verbiage I mentioned.

spec/immutable-data-structures.html

ljharb · 2020-06-23T03:15:05Z

spec/immutable-data-structures.html

+            1. Let _name_ be ? GetV(_obj_, `"0"`).
+            1. Let _value_ be ? GetV(_obj_, `"1"`).


Suggested change

1. Let _name_ be ? GetV(_obj_, `"0"`).

1. Let _value_ be ? GetV(_obj_, `"1"`).

1. Let _name_ be ! GetV(_prop_, `"0"`).

1. Let _value_ be ! GetV(_prop_, `"1"`).

(the ! is since we know these can't fail, because of EnumerableOwnPropertyNames's algorithm)

Not sure why this was resolved

apologies, part of my broken push, I've pushed a commit that contains all of @ljharb 's suggested changes. Should be good now.

spec/immutable-data-structures.html

ljharb

For the Tuple methods that make a new Tuple (map/filter/concat/spliced/etc), it seems like rather than duplicating the array method steps and adjusting them, there should be new abstract operations that both the Array methods and the Tuple methods use.

spec/immutable-data-structures.html

Co-authored-by: Jordan Harband <[email protected]>

…onstruction

…ertyNames

ljharb · 2020-06-24T22:31:37Z

spec/immutable-data-structures.html

+          1. Let _adder_ be a new Abstract Closure with parameters (_key_, _value_) that captures _fields_ and performs the following steps when called:
+            1. If Type(_value_) is Object, throw a *TypeError* exception.
+            1. Let _field_ be { [[Key]]: _key_, [[Value]]: _value_ }.
+            1. Append _field_ to the end of list _fields_.
+          1. Perform ! AddEntriesFromIterable(`undefined`, _iterable_, _adder_).


this looks much better, thanks!

spec/immutable-data-structures.html

Co-authored-by: Jordan Harband <[email protected]>

rickbutton · 2020-06-25T19:00:03Z

Oh, I realize what happened @ljharb , I messed up a force-push. Whoops! I will reapply your suggestions asap.

littledan

Overall, this seems good to land.

spec/immutable-data-structures.html

littledan · 2020-06-25T18:52:57Z

spec/immutable-data-structures.html

+            1. Let _name_ be ? GetV(_obj_, `"0"`).
+            1. Let _value_ be ? GetV(_obj_, `"1"`).


Not sure why this was resolved

spec/immutable-data-structures.html

littledan · 2020-06-25T19:08:49Z

spec/immutable-data-structures.html

+          1. Let _T_ be ? thisTupleValue(*this* value).
+          1. Let _list_ be a new List containing the elements of _T_'s [[Sequence]] value.
+          1. Let _len_ be the number of elements in _list_.
+        </emu-alg>


It would be good if you could avoid duplicating the below section with Array.prototype.sort, since it's just the same, right?

Most of it is the same, except with the handling of sparse arrays removed. We could share more here. What would be the best way to share more of this? Not sure how to handle it, since it isn't quite an abstract-op, very prose-y.

spec/immutable-data-structures.html

rickbutton · 2020-06-26T02:31:16Z

Merging this in so that we can continue to iterate on this in smaller chunks and merge it into the other and upcoming PRs.

rricard · 2020-06-26T08:19:36Z

I'll rebase on this as well since you improved the ecmarkup pipeline

Initial Record/Tuple standard library spec

rricard reviewed Jun 10, 2020

View reviewed changes

package.json Outdated Show resolved Hide resolved

littledan reviewed Jun 11, 2020

View reviewed changes

spec/immutable-data-structures.html Outdated Show resolved Hide resolved

ljharb reviewed Jun 12, 2020

View reviewed changes

spec/immutable-data-structures.html Outdated Show resolved Hide resolved

papb mentioned this pull request Jun 16, 2020

Names of Tuple.prototype.* methods #121

Closed

littledan reviewed Jun 16, 2020

View reviewed changes

littledan reviewed Jun 17, 2020

View reviewed changes

spec/immutable-data-structures.html Outdated Show resolved Hide resolved

littledan reviewed Jun 17, 2020

View reviewed changes

spec/immutable-data-structures.html Outdated Show resolved Hide resolved

ljharb reviewed Jun 19, 2020

View reviewed changes

spec/immutable-data-structures.html Outdated Show resolved Hide resolved

spec/immutable-data-structures.html Outdated Show resolved Hide resolved

ljharb requested changes Jun 23, 2020

View reviewed changes

ljharb reviewed Jun 23, 2020

View reviewed changes

spec/immutable-data-structures.html Outdated Show resolved Hide resolved

ljharb reviewed Jun 23, 2020

View reviewed changes

spec/immutable-data-structures.html Outdated Show resolved Hide resolved

ljharb reviewed Jun 23, 2020

View reviewed changes

spec/immutable-data-structures.html Outdated Show resolved Hide resolved

ljharb reviewed Jun 23, 2020

View reviewed changes

spec/immutable-data-structures.html Outdated Show resolved Hide resolved

ljharb reviewed Jun 23, 2020

View reviewed changes

littledan reviewed Jun 23, 2020

View reviewed changes

spec/immutable-data-structures.html Outdated Show resolved Hide resolved

littledan reviewed Jun 23, 2020

View reviewed changes

spec/immutable-data-structures.html Show resolved Hide resolved

littledan mentioned this pull request Jun 24, 2020

Record and Tuple object model spec #132

Merged

rickbutton and others added 10 commits June 24, 2020 17:14

initial shell for record/constructor stdlib

d6bd914

implemented more record/tuple constructor/prototype properties

5d1d336

TupleCreate -> CreateTupleFromList:

fda0d1e

more constructor/prototype, Tuple.prototype.concat

13e2b40

pushed/popped

c122a48

shifted/unshifted/reversed

5b73fa4

reference some existing array methods, for now?

1d5f049

entries/keys/values/with

735c5f1

Update spec/immutable-data-structures.html

dce438d

Co-authored-by: Jordan Harband <[email protected]>

@@iterator

acb4543

rickbutton added 10 commits June 24, 2020 17:14

replace CreateListFromTuple

52b8bb6

re-use Array iterator instead of creating Tuple iterator

1942649

Record.prototype is null

cee72f2

replace CreateTupleFromList with explicit object checks and a tuple c…

ca8267f

…onstruction

don't create a new list when you don't have to

da54093

remove extraneous asserts

88e3c43

record constructor, Record.fromEntries

6833c41

tweaked Record.fromEntries to fetch prop values via EnumerableOwnProp…

18cba33

…ertyNames

use AddEntriesFromIterable for Record.fromEntries

2d8686c

Tuple.prototype.sorted !!!

8b61abe

rickbutton force-pushed the rb/stdlib branch from 766b746 to 8b61abe Compare June 24, 2020 21:14

rickbutton added 4 commits June 24, 2020 17:17

Record [[Prototype]] is null

55b8b4f

concrete Record.isRecord and Tuple.isTuple

14af093

use abstract closure for Record.fromEntries

0218a00

use AddEntriesfromIterable for Tuple.from

949ba86

ljharb reviewed Jun 25, 2020

View reviewed changes

rickbutton and others added 2 commits June 25, 2020 14:46

Apply suggestions from code review

dd1df80

Co-authored-by: Jordan Harband <[email protected]>

fixing some missing wording changes

16c53f3

littledan approved these changes Jun 25, 2020

View reviewed changes

rickbutton added 4 commits June 25, 2020 15:16

apply suggestions from @ljharb

b24d79f

typo in Record()

752087d

casing

f6ca881

Added Tuple(number) option for Tuple function

29f8e6c

littledan approved these changes Jun 25, 2020

View reviewed changes

rickbutton marked this pull request as ready for review June 26, 2020 02:26

rickbutton merged commit be0d278 into master Jun 26, 2020

rickbutton added a commit that referenced this pull request Jan 4, 2021

Merge pull request #131 from tc39/rb/stdlib

4dc445d

Initial Record/Tuple standard library spec

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Initial Record/Tuple standard library spec #131

Initial Record/Tuple standard library spec #131

rickbutton commented Jun 9, 2020

ljharb left a comment

rricard commented Jun 12, 2020

littledan left a comment

littledan Jun 16, 2020

ljharb Jun 23, 2020

littledan Jun 23, 2020

rickbutton commented Jun 16, 2020

littledan commented Jun 16, 2020

ljharb Jun 23, 2020

littledan Jun 25, 2020

rickbutton Jun 25, 2020

ljharb left a comment

ljharb Jun 24, 2020

rickbutton commented Jun 25, 2020

littledan left a comment

littledan Jun 25, 2020

littledan Jun 25, 2020

rickbutton Jun 25, 2020

rickbutton commented Jun 26, 2020

rricard commented Jun 26, 2020

		1. Let _name_ be ? GetV(_obj_, `"0"`).
		1. Let _value_ be ? GetV(_obj_, `"1"`).

Initial Record/Tuple standard library spec #131

Initial Record/Tuple standard library spec #131

Conversation

rickbutton commented Jun 9, 2020

ljharb left a comment

Choose a reason for hiding this comment

rricard commented Jun 12, 2020

littledan left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rickbutton commented Jun 16, 2020

littledan commented Jun 16, 2020

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ljharb left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rickbutton commented Jun 25, 2020

littledan left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rickbutton commented Jun 26, 2020

rricard commented Jun 26, 2020