Permalink
Commits on Jan 12, 2017
Commits on Dec 30, 2016
Commits on Dec 29, 2016
  1. Editorial: Annotate ToObject calls with `!` (#759)

    These calls are preceded by checks for `null` and `undefined`,
    so they're guaranteed to succeed. But not all such calls were
    annotated with `!`.
    ajklein committed with bterlson Dec 29, 2016
Commits on Dec 24, 2016
Commits on Dec 21, 2016
  1. Editorial: Misc cleanups of new Agents text (#756)

    * Editorial: change tabs to spaces
    * Editorial: change "internal slot" to "field"
    ... because an Agent Record is a record, not an object.
    jmdyck committed with bterlson Dec 21, 2016
  2. Normative: Disregard lastIndex for non-global non-sticky regexps (#627)

    With these changes , non-global non-sticky regexps with nonwritable
    lastIndex property should just work:
    * In RegExpBuiltinExec, remove any access to the lastIndex property for
    non-global, non-sticky regexp.
    * In RegExp.prototype.@@search, avoid to to set the lastIndex property
    when not necessary.
    claudepache committed with bterlson Dec 21, 2016
Commits on Dec 20, 2016
  1. Normative: Remove @@iterator from Module Namespace objects (#747)

    Per consensus at the November 2016 meeting, Module Namespace exotic
    objects no longer have an @@iterator method.
    
    Also changes the [[Configurable]] property of the @@toStringTag
    property to false, as per consensus in the same meeting (and
    because it's impossible to reconfigure due to the exotic
    methods of such objects).
    
    This closes #693 and #710.
    ajklein committed with bterlson Dec 20, 2016
Commits on Dec 8, 2016
Commits on Nov 25, 2016
  1. Editorial: Misc editorial (#739)

    * Editorial: ensure a space at the start of parameter lists
    
    (leftovers from @bterlson's commit
    "Editorial: consistent spacing around parameter lists")
    
    * Editorial: insert missing </emu-grammar> tag
    
    * Editorial: add underscores around metavariables
    
    * Editorial: use of IdentifierReference needs parameters
    
    When IdentifierReference occurs in the RHS of a defining production,
    it must have its Yield and Await parameters set (to either + or ~).
    I've assumed ~ for both; please check that this is correct.
    jmdyck committed with bterlson Nov 25, 2016
Commits on Nov 23, 2016
  1. Editorial: Re-factor ExportSpecifier (#637)

    In prior editions, the ExportSpecifier production was defined using only
    IdentifierName. However, when the ExportSpecifier is part of an "local"
    ExportDeclaration, the IdentifierName was interpreted as an
    IdentifierReference through an Early Error rule. The rule was
    undesirable for a number of reasons:
    
    - it duplicated the definition of IdentifierReference
    - it relied on an abstract operation, ReferencedBindings, which was
      otherwise unused
    - it was not technically necessary (a separate Early Error rule for
      ModuleBody requires that all ExportedBindings are declared within the
      module itself, and it is not possible to create a such a binding)
    
    These details made interpreting ExportSpecifier difficult and tended to
    obscure the motivation for the use of IdentifierName.
    
    Re-factor ExportSpecifier with a production parameter describing whether
    it belongs to a "local" or "indirect" ExportDeclaration. This allows for
    the definition of dedicated grammar productions, obviating the need for
    an explicit early error, and removing branching logic from the static
    semantics rule ExportEntriesForModule. It also more clearly communicates
    the intent behind related definitions of the production.
    jugglinmike committed with bterlson Nov 23, 2016
  2. Editorial: tolength: remove superfluous +Infinity check (#738)

    The `+Infinity` check is superfluous as the `min` function takes care
    of that as well.
    thefourtheye committed with bterlson Nov 23, 2016
Commits on Nov 21, 2016
  1. Add definition of unfolding of AbstractOperation(ReturnIfAbrupt(arg)) (

    …#733)
    
    * Add definition of unfolding of AbstractOperation(ReturnIfAbrupt(arg))
    
    The steps required when a ReturnIfAbrupt is nested in a call to another
    operation are not defined and do not directly follow from the
    ReturnIfAbrupt(AbstractOperation()) case.
    
    * Remove unnecessary "Else" as recommended by ljharb
    edgemaster committed with bterlson Nov 21, 2016
Commits on Nov 15, 2016
  1. Editorial: Misc editorial fixes (#728)

    * Editorial: add '*" around "true" and "false"
    * Editorial: Use capital 'C' for the 'Contains' relation
    * Editorial: put production into <emu-grammar> element
    * Editorial: put period at end of Early Error rules
    * Editorial: "early errors rules" -> "early error rules"
    * Editorial: use present tense in Assert
    * Editorial: Delete extra space in alg step
    jmdyck committed with bterlson Nov 15, 2016
  2. Editorial: Rearrange tables to be consistent with section ordering (#730

    )
    
    * Editorial: Rearrange Table 30 to be consistent with Section 9.5 ordering.
    * Editorial: Rearrange Table 5 to be consistent with Section 9.1 ordering.
    edgemaster committed with bterlson Nov 15, 2016
Commits on Nov 14, 2016
Commits on Nov 11, 2016
  1. Normative: Async Functions (#692)

    bterlson committed on GitHub Nov 11, 2016
Commits on Nov 2, 2016
  1. Normative: Clarify rules around NewTarget inside eval (#686)

    Previously, NewTarget was disallowed in nested eval contexts and in eval
    inside arrow functions. This change allows NewTarget inside eval
    anywhere it would be allowed outside of eval.
    bterlson committed on GitHub Nov 2, 2016
  2. Normative: Apply chain rule to IsFunctionDefinition UnaryExpression: …

    …UpdateExpression (#706)
    
    Since a *FunctionExpression* is a *PrimaryExpression*, I believe
    returning `false` in `IsFunctionDefinition` for every production of
    *UnaryExpression* is incorrect. It's also not consistent with the rules for
    every other expression.
    
    Otherwise I think `IsFunctionDefinition` would return `false` when
    evaluating initializers (e.g. `function foo(bar=function(){}) {}`),
    after "unrolling" all production rules.
    fkling committed with bterlson Nov 2, 2016
  3. Editorial: Minor grammar correction regarding arguments.callee (#707)

    This fixes two pluralization agreement issues introduced by #689.
    bakkot committed with bterlson Nov 2, 2016