ECMA/ISO meeting

Document explaining the differences between ECMAScript and JScript/JavaScript from ISO/ECMA/NSCP/Microsoft. This will be available from NSCP/MSFT/ECMA/ISO sites for download.

The group voted unanimously in favor of the standard to be ratified.

Standard is now DIS-16262.

Technical Meeting

July 13th 9:00 PDT Telephone meeting.

September 15th/16th @ NSCP from 10:00 'til 5:00 for next face to face.

Andrew Clinick will send round details for newsgroup and investigate getting a mail reflector for the newsgroup.

Reply to Martin Duerst

Drafted by Clayton. The references to URL's have been removed from the V1 standard.

There is some question over RFC 1738 but this will be resolved via email.

Rok will talk to Norris to ensure that MSFT's and NSCP's browser implementations conform and to see if there are any further anomalies.

The group will ask if he wants to be on the reflector.

Clayton will rephrase the reply from the committee and ask if wants to be on the reflector and inform him of the next meeting date in September.

Thanks to Karl for putting together the disposition of comments and Bill for editing the document.

Propose that all Version 1 changes get merged into the current Version 2 document.

The corrections for ISO will be referred to as the second edition of ECMA 262 not the second version.

SPICE/XSL

Andrew Clinick will invite XSL and SPICE representatives to a technical meeting on September 16th to discuss their requirements/needs of the language in the future. Copy in Martin Duerst.

Exception handling

Everyone seems to be in agreement on the current proposal but the existing document does not conform to the new thinking. Waldemar will be rewriting the document and provide a draft prior to the September 15th meeting.

I18N Unicode Issues

Tom McFarlands comments on I18N will discussed in the afternoon session.

Section 16 Errors

The existing section is fine for Version 1 but exception handling will require some major changes to this.

15.13.3.6

15.12.2.11 Max and 15.12.2.12 Min

Content agreed but what about when no argument is passed in then it would be positive or negative infinity.

15.7.x Array

Changes in this section should be backed out to reflect V1 of the spec.

15.7.4.3 Array.prototype.toSource

Get the length once and return undefined if it changes. We will have a general discussion on persistence and then revisit toSource when we have made a decision of what to do.

15.7.4.4 Array.prototype.concat

Differences in implementation (NSCP use real arrays MSFT use sparse arrays) resulted in the rules in step 7 allowing for implementation dependent mechanism for dealing with arrays.

The 'this' argument is treated differently than the array arguments themselves. It is proposed that it should be handled the same as arguments. Mike (NSCP) will provide a write up on this and the key issues with array.

Array.prototype.pop

Returns undefined. Functionally correct.

Array.prototype.push

Waldemar will look into this and write up his proposals to the group.

Array.prototype.shift

Naming came from PERL implementation and is shipping with Navigator. Rok Yu suggested an argument to the implementation and will provide a specification for it.

Array.prototype.unshift

Naming came from PERL implementation and is shipping with Navigator.

Sparse Arrays

The group decided that further discussion was required on this subject since it affects so many methods in the language. A document discussing extensions to array is required. It will be delivered before the meeting where this will be discussed.

Array.prototype.slice

Slice needs review, NSCP to provide this.

Array.prototype.splice

Not ready. Needs considerable review from the group. Bill Gibbons will own this.

 

15.8.4.x String

String.prototype.concat

Group agreed in principle to the proposal. Step 6 has a typo.

String.prototype.slice

String.prototype.substring and String.prototype.substr

NSCP to provide a proposal about which ones to keep in the specification.

 

InstanceOf operator

Keywords and Reserved words

Should be in ISO approved version 1.

8.6.2 Internal Properties/Methods

Requires separate discussion other than HasInstance.

11.8 Relational operators

Can go in the draft document as functional complete.

11.8.6 instanceOf

The differences between left hand side and the right hand side. The right hand side is very strict yet the left-hand side is much more forgiving. This is inconsistent.

Step 8 has to change from return result 6 to return result 7.

11.8.7 in operator

The proposal is functionally correct. There were some concerns raised over whether in should be an operator or a function.

Everything 15.4.2.5

Agreed.

15.2.4.5 Obect.prototype.hasProperty

Discussion over naming hasImmediateProperty was proposed. Step 3 required some more work. There are no real precedents from prototype based languages. Rok Yu will look into the name.

15.2.4.6 hasDelegate

Renamed to isPrototypeOf.

15.2.4.7 propertyIsEnumerable

15.3.5.5 hasInstance

Agreed to functionally accept. Language will be cleaned up by NSCP.

7.7 Why did XXXX get moved

Bill Gibbons explained that it had to be moved for compliance.

8.9 Completion Type

Covered by exception handling.

9.8 ToString

Functionally accepted.

11.1 Primary Expressions

Leave this open for sharp variable discussion.

11.1.4 Array initializer

Leave this open for sharp variable discussion.

13 Function definition

Leave for Stack frame discussion.

14 Program

Change to content agreed dependent on the exception handling proposals.

15.1.2.1 Eval (x)

Changed for V1 already so just needs to revert to V1.

15.1.4.2 Really Regular Expressions

Leave for RegExp discussion.

15.2.4 Properties of the Global Object

Leave for toSource discussion.

15.3.4 Properties of the Function Prototype Object

Converted to a C.

15.3.5 Properties of Function Instances

Arguments deleted in V1.

15.3.5.4 prototype

ReadOnly doesn't make sense so ReadOnly should be deleted.

15.4 Call Objects

Defer to Stack frames discussion.

7.3 Comments

Version 1 (ISO revision) has the correct changes so when this is merged into the V2 doc then it will be correct.

7.4 Tokens

Other than the sharp variables reference this is OK.

Undefined

Wanted a keyword but perhaps it could be added as an undeletable property to the global object. Herman will write up a proposal for this.

Exception Handling

MSFT will write up the current JScript implementation for Error object.

 

Agenda for Next meeting

July Call

Sparse Arrays [WH]

Array like [WH]

Array & Push[RY]

Array concat[MM]

String [NB]

September Meeting

15th

Try Catch [WH]

"Stack Frames" [CL]

16th

W3C (Spice/XSL) [AC] a.m.

I18N [MK] p.m.

Regular Expression [AC]