Misc editorial #1053

jmdyck · 2017-12-21T03:44:14Z

In the first commit, add a [lookahead <! HexDigit] restriction to these
two right-hand sides of NotEscapeSequence:

    u { NotCodePoint
    u { CodePoint

to force them to consume all the HexDigits there.
Otherwise, the NotEscapeSequence could end 'early',
causing 'TemplateCharacters' to be ambiguous.

(I think the TRV is the same regardless which parse you pick, but presumably we want to avoid ambiguity anyway.)

Fix some typos in a recent commit.

Resolve a technical (though not substantive) ambiguity raised in issue #1059.

Attn @mathiasbynens re recent commits (starting at "consistify grammar params in defining prodns").

mathiasbynens · 2018-01-26T22:56:45Z

This needs a rebase, but the changes look great — thank you @jmdyck!

jmdyck · 2018-01-26T22:57:46Z

Working on the rebase now.

jmdyck · 2018-01-26T23:25:45Z

rebase done

ljharb · 2018-01-26T23:28:26Z

spec.html

@@ -30395,7 +30395,7 @@ <h1>Runtime Semantics: UnicodeMatchProperty ( _p_ )</h1>
          <p>Implementations must only recognize the property aliases listed in <emu-xref href="#table-nonbinary-unicode-properties"></emu-xref> and <emu-xref href="#table-binary-unicode-properties"></emu-xref>.</p>
          <p>Implementations must only recognize the property value aliases and canonical property value names listed in <emu-xref href="#table-unicode-general-category-values"></emu-xref> and <emu-xref href="#table-unicode-script-values"></emu-xref>.</p>
          <emu-note>
-            <p>For example, `Script_Extensions` (property name) and `scx` (property alias) are valid, but `script_extensions` or `Scx` aren’t.</p>
+            <p>For example, `Script_Extensions` (property name) and `scx` (property alias) are valid, but `script_extensions` or `Scx` aren't.</p>


typographically, ‘ is correct here (and ' isn't); why these changes?

consistency with the rest of the spec. (scan for /[a-z]'[a-z]/)

Gotcha, that's unfortunate.

ljharb · 2018-01-26T23:28:40Z

spec.html

@@ -30415,7 +30415,7 @@ <h1>Runtime Semantics: UnicodeMatchPropertyValue ( _p_, _v_ )</h1>
          </emu-alg>
          <p>Only the canonical property values and property value aliases listed in <emu-xref href="#table-unicode-general-category-values"></emu-xref> and <emu-xref href="#table-unicode-script-values"></emu-xref> must be recognized.</p>
          <emu-note>
-            <p>For example, `Xpeo` and `Old_Persian` are valid `Script_Extension` values, but `xpeo` and `Old Persian` aren’t.</p>
+            <p>For example, `Xpeo` and `Old_Persian` are valid `Script_Extension` values, but `xpeo` and `Old Persian` aren't.</p>


(also here)

ljharb · 2018-01-26T23:29:30Z

spec.html

@@ -30549,7 +30549,7 @@ <h1>CharacterClassEscape</h1>
            1. Return the CharSet containing all Unicode code points whose character database definition includes the property &ldquo;General_Category&rdquo; with value _LoneUnicodePropertyNameOrValue_.
          1. Let _p_ be ! UnicodeMatchProperty(_LoneUnicodePropertyNameOrValue_).
          1. Assert: _p_ is a binary Unicode property or binary property alias listed in the &ldquo;Property name and aliases&rdquo; column of <emu-xref href="#table-binary-unicode-properties"></emu-xref>.
-          1. Return the CharSet containing all Unicode code points whose character database definition includes the property _p_ with value |True|.
+          1. Return the CharSet containing all Unicode code points whose character database definition includes the property _p_ with value &ldquo;True&rdquo;.


Should the value true be lowercased?

The current casing is correct. The binary property values according to the Unicode Standard are N/No/F/False and Y/Yes/T/True.

gotcha, I read it as "JS value" :-) thanks for clarifying.

ljharb

LGTM, but def needs more eyes

jmdyck · 2018-01-27T19:25:25Z

added two more commits re lookbehind assertions, attn @mathiasbynens

Specifically, add a [lookahead <! HexDigit] restriction to the two right-hand sides: u { NotCodePoint u { CodePoint to force them to consume all the HexDigits there. Otherwise, the NotEscapeSequence could end 'early', causing 'TemplateCharacters' to be ambiguous.

... in an algorithm step

... in algorithm steps.

(Could go either way. I chose to resolve it in the direction of smaller diff.)

... because that RHS made it ambiguous. (Resolves issue tc39#1059.)

... because it's already derived by UnicodeIDContinue RHS. (See issue tc39#1059.)

... so that it doesn't look like PromiseResolve is a property of the Promise Constructor.

... in prodns for AtomEscape and ClassEscape. (Please check that [?U] is correct, and not [+U] or [~U].)

... from non-defining CharacterClassEscape prodns.

... because backtick is for ECMAScript code.

... because pipe is for nonterminals.

... from UnicodeMatchProperty to UnicodeMatchPropertyValue.

... for UnicodeMatchProperty and UnicodeMatchPropertyValue. (And make it parallel between the two.)

... for UnicodeMatchProperty and UnicodeMatchPropertyValue (as in typeof, DateString, GetSubstitution)

(The spec tends to put a nonterminal's definition after its right-hand-side uses.)

_UnicodePropertyName_, _UnicodePropertyValue_, and _LoneUnicodePropertyNameOrValue_ are invalid because UnicodePropertyName etc are nonterminals, not metavariables. And simply changing them to |UnicodePropertyName| etc wouldn't be valid, because that would be passing a Parse Node to UnicodeMatchProperty/UnicodeMatchPropertyValue, which isn't what they're expecting. So instead, use the SourceText operation to 'extract' the List of code points for each.

... in UnicodeMatchProperty.

... to resolve BackreferenceMatcher's reference to _direction_ at step 1.f

The note's content was marked up as if the code sample would be rendered inline, but it's a <pre>, so it'll be a block.

... so that the <td> element doesn't have both inline and block content.

jmdyck · 2018-02-08T17:50:04Z

Added 3 markup commits re #890, attn @littledan.

(I should have caught these in the branch review I did 5 days ago, but I guess I only looked at the "reparse" -> "covering" changes.)

littledan · 2018-02-08T22:11:54Z

Markup commits for #890 LGTM

bterlson · 2018-02-12T23:51:18Z

spec.html

@@ -30298,12 +30298,12 @@ <h1>Atom</h1>
              1. Let _xe_ be _x_'s _endIndex_.
              1. Let _ye_ be _y_'s _endIndex_.
              1. If _direction_ is equal to +1, then
-                1. Assert: _xe_ &lte; _ye_.
-                1. Let _s_ be a fresh List whose characters are the characters of _Input_ at indices _xe_ (inclusive) through _ye_ (exclusive).
+                1. Assert: _xe_ &le; _ye_.


I think the intention here was less than or equal to. While gte/lte are not entities, ≤= would preserve the intent. Is < a better assert? /cc @littledan

never mind I don't know HTML entities.

jmdyck · 2018-02-13T00:23:16Z

Yay! Thanks.

jmdyck force-pushed the editorial branch from 58c4083 to 4083a4b Compare January 4, 2018 18:30

jmdyck changed the title ~~Editorial: add more lookahead-restrictions to NotEscapeSequence~~ Editorial: add more lookahead-restrictions to NotEscapeSequence (plus tweaks to recent commits) Jan 4, 2018

jmdyck changed the title ~~Editorial: add more lookahead-restrictions to NotEscapeSequence (plus tweaks to recent commits)~~ Misc editorial Jan 5, 2018

anba mentioned this pull request Jan 17, 2018

LocalTimeZoneAdjustment/LocalTZA follow-ups necessary #1070

Closed

jmdyck force-pushed the editorial branch 3 times, most recently from 7663d7f to 77bfec2 Compare January 26, 2018 00:09

jmdyck mentioned this pull request Jan 26, 2018

Missing [N] in RegExp grammar? #1081

Open

mathiasbynens approved these changes Jan 26, 2018

View reviewed changes

jmdyck force-pushed the editorial branch from 77bfec2 to 2180ded Compare January 26, 2018 23:24

ljharb reviewed Jan 26, 2018

View reviewed changes

ljharb approved these changes Jan 27, 2018

View reviewed changes

mathiasbynens approved these changes Jan 27, 2018

View reviewed changes

jmdyck force-pushed the editorial branch 2 times, most recently from a2bcffb to e6f8436 Compare February 2, 2018 02:02

jmdyck added 11 commits February 8, 2018 11:46

Editorial: insert missing </emu-eqn>

c564154

Editorial: delete extraneous dot

dbf7751

Editorial: delete extra spaces and add a dot

cf8c857

... in an algorithm step

Editorial: add '*' around 'true' and 'false'

d7f1c6c

... in algorithm steps.

Editorial: fix typo "algorthm"

12dc570

Editorial: resolve "LocalTimeZoneAdjustment" vs "LocalTZA" inconsistency

0f1d570

(Could go either way. I chose to resolve it in the direction of smaller diff.)

Editorial: Remove '_' RHS for IdentifierPart

ec1cd8c

... because that RHS made it ambiguous. (Resolves issue tc39#1059.)

Editorial: entity-encode < and >

b4081da

Editorial: tweak some indentation

0e04451

Editorial: consistify grammar params in defining prodns

342fdb0

jmdyck added 23 commits February 8, 2018 11:46

Editorial: "Let" -> "Set" for already-bound metavariable

261bc18

Editorial: remove _ RHS from RegExpIdentifierPart

a1f915b

... because it's already derived by UnicodeIDContinue RHS. (See issue tc39#1059.)

Editorial: tweak indentation

f09e801

Editorial: move PromiseResolve clause into Promise.resolve clause

35ab8d5

... so that it doesn't look like PromiseResolve is a property of the Promise Constructor.

Editorial: add "then" to "If" step

feaa751

Markup: entity-encode “ and ”

cd12b08

Editorial: right-single-quotation-mark -> apostrophe twice

57b25af

Markup: <emu-nt>Foo</emu-nt> -> |Foo|

f2ed860

Editorial: Add [?U] to RHS occurrences of CharacterClassEscape

0fc5f71

... in prodns for AtomEscape and ClassEscape. (Please check that [?U] is correct, and not [+U] or [~U].)

Editorial: delete extraneous backslashes

2ad9c5f

... from non-defining CharacterClassEscape prodns.

Editorial: General_Category -> “General_Category”

2a7bb57

... because backtick is for ECMAScript code.

Editorial: |True| -> “True”

6a5870c

... because pipe is for nonterminals.

Editorial: move a sentence about property values

392b8d9

... from UnicodeMatchProperty to UnicodeMatchPropertyValue.

Editorial: tighten up the "must support/must not support" wording

607e7d1

... for UnicodeMatchProperty and UnicodeMatchPropertyValue. (And make it parallel between the two.)

Editorial: Put algorithm before supporting tables

9a4fa7a

... for UnicodeMatchProperty and UnicodeMatchPropertyValue (as in typeof, DateString, GetSubstitution)

Editorial: reorder UnicodePropertyValueExpression-related prodns

e3d9eff

(The spec tends to put a nonterminal's definition after its right-hand-side uses.)

Editorial: avoid reusing/overwriting metavariable _p_

b756cef

... in UnicodeMatchProperty.

Editorial: misc tweaks re lookbehind assertions

d26d24a

Editorial: add _direction_ to BackreferenceMatcher's param list

6e04f06

... to resolve BackreferenceMatcher's reference to _direction_ at step 1.f

Markup: fix well-formedness error

03bd376

Markup: tweak note re parsing the same String multiple times.

17da467

The note's content was marked up as if the code sample would be rendered inline, but it's a <pre>, so it'll be a block.

Markup: insert <p>...</p> tags

528eb61

... so that the <td> element doesn't have both inline and block content.

jmdyck force-pushed the editorial branch from e6f8436 to 528eb61 Compare February 8, 2018 17:45

bterlson reviewed Feb 12, 2018

View reviewed changes

bterlson merged commit 528eb61 into tc39:master Feb 13, 2018

ljharb mentioned this pull request Apr 1, 2018

IdentifierPart is ambiguous re '_' ? #1059

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Misc editorial #1053

Misc editorial #1053

jmdyck commented Dec 21, 2017 •

edited

Loading

mathiasbynens commented Jan 26, 2018

jmdyck commented Jan 26, 2018

jmdyck commented Jan 26, 2018

ljharb Jan 26, 2018

jmdyck Jan 26, 2018

ljharb Jan 27, 2018

ljharb Jan 26, 2018

ljharb Jan 26, 2018

mathiasbynens Jan 26, 2018

ljharb Jan 26, 2018

ljharb left a comment

jmdyck commented Jan 27, 2018

jmdyck commented Feb 8, 2018

littledan commented Feb 8, 2018

bterlson Feb 12, 2018

bterlson Feb 13, 2018

jmdyck commented Feb 13, 2018

Misc editorial #1053

Misc editorial #1053

Conversation

jmdyck commented Dec 21, 2017 • edited Loading

mathiasbynens commented Jan 26, 2018

jmdyck commented Jan 26, 2018

jmdyck commented Jan 26, 2018

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

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

jmdyck commented Jan 27, 2018

jmdyck commented Feb 8, 2018

littledan commented Feb 8, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jmdyck commented Feb 13, 2018

jmdyck commented Dec 21, 2017 •

edited

Loading