ECMA meeting 9/15/98

Present

Mike Cowlishaw (IBM)

Herman Venter (MSFT)

Jeff Dyer (Nombas)

Bill Gibbons (NSCP)

Karl Matzke (SUN)

Clayton Lewis (NSCP)

Norris Boyd (NSCP)

Mike McCabe (NSCP)

Waldemar Horwat (NSCP)

Andrew Clinick (MSFT)

Rok Yu (MSFT)

Next meeting:

10/16/98 12:00 noon to 6:00 p.m. at Microsoft

11/19/98 & 11/20/98 at Netscape including ECMA GA report with Gary Robinson

12/14/98 Telephone meeting

Karl Matzke will be taking a sabbatical for a couple of years so wonít be able to attend the meetings. Sun will provide a replacement.

ECMA 262 Edition 2

The version of the ECMA spec on the web does not include all the "disposition of comments" changes. The existing mechanism is not working well at the moment we need to ensure that we all sign off on the document.

AC will check with the Word team about change tracking and sub documents.

Resolutions:

 

Version 2.0 Ė ECMAScript Edition

Bill Gibbons will merge content agreed work into the new base document (the Edition 2).

CL will check with ECMA to see if a mailing list can be set up for technical discussion. If not then he will set up one from NSCP.

NB noted a change to 12.6.1 and will send details to the technical reflector

11.9.6 boundary case for undefined and Null. WH will send email to the reflector

Array.prototype.concat/pop/push/shift

MM will write up proposed text and submit to the group. There was some discussion whether it should be generic.

String.prototype.slice

Rok will draft a recommendation to the mailing list. Get rid of 15.8.4.12 and merge with 15.8.4.13

Try Catch

WH will write up the definition of behavior for side effects in the catch guard. Section 12.14. WH will circulate this definition by 9/28

Computational

Bill Gibbons will provide a document that will describe how ECMAScript Edition 3 should be updated for the November meeting.

7.4 Tokens

Mentions sharp variables so will be looked at after discussion. Tokens will be categorized as a D.

7.7 Literals

This can now be moved to complete since it is already updated in the ECMA 262 Edition 2 specification. Moved to Complete.

7.7.4 String Literals

WH will write up recommendation.

11.8 Relational Operators

instanceOf and in have been submitted as proposals. MM would like to make some changes to the proposals.

12.1 Block

Changed to C

12.2 Variable Statement

Part of the instanceOf proposal. Moves to C

12.4 Expression Statement

Moves to C

12.7 The continue statement and 12.8 The break statement

Moved to C

12.10 With Statement

More discussion required

12.11 Switch

Moved to complete

12.14 Try Statement

WH will write up proposal

15.3.5.1 Properties of Function Instances

Should be removed when the Edition 2 document is merged with the Edition 3 proposals

15.3.5.2 Arity

15.5.5 HasInstance

15.12.2.11 min and max

WH will prepare a draft

Regular Expressions

AC will document differences and provide a proposal of what we might standardize in ECMA Edition 3

ECMA 262 9/16/1998

W3C Related Issues

Martin Duerst

Xerox were involved in this decision

Schematic

Characters

UCS

UTF-8

%HH

RFC 2396 covers UTF-8 and %HH

2 new functions uriEncode and uriDecode that will deal with encoding into the %HH

Martin Duerst will provide a URL to his talk from the last Unicode meeting

W3C international site

Spice/XSL

Dave Raggett to mail the docs to AC

AC will organize a meeting with Spice, NSCP and MSFT

Docs by the beginning of October

Meeting in the UK in the first week of November

Internationalization

IBM suggested that ECMAScript should have a minimal set and the host application should provide more features.  This was discussed further and the group felt that the locale features should be in an external function/class library.
Single local model does have numerous problems.
7.7.5 Regular Expression Literals
www.unicode.org/techreports number 15 contains more information

Deals with computer readable text not natural language at the moment. The group will consider having language sensitive regexp syntax input is requested for this in the future.

9.3.1 toNumber Applied to the String Type

The StrWhiteSpaceChar should be extended to include all Unicode white space characters. The spec should refer to the Unicode character class. The function should stay the same as it is today and the new spec should include a solution to provide locale specific methods.

9.8.1 toString

See 9.3.1

11.8.5 Comparison operators and 11.9.3 Equality operators

Define separate methods to deal with this problem e.g. mystring.compare(otherstring)

12.11 The Switch statement

This is covered assuming the host normalizes the data before the script is run. It was suggested that the string object have a normalize method to provide normalized data at runtime.

15.7.4.8 Array.prototype.reverse()

Since an array is not a string in ECMAScript this does not apply

15.7.4.9 Array.prototype.shift()

Since an array is not a string in ECMAScript this does not apply

15.8.* String Object

Some functionality for identifying characters in the string

ToLower, toUpper, and toTitleCase

Should have a locale argument to allow for locale

15.9.1 \w & \W

Doesnít include $ which is a valid ECMAScript control character. The regExp spec should be reviewed and in general use Unicode categories