W3C home > Mailing lists > Public > www-voice@w3.org > July to September 2010

Re: CCXML: comment on ISSUE-737

From: Chris Davis <davisc@iivip.com>
Date: Tue, 03 Aug 2010 09:26:02 -0500
Message-ID: <4C5826FA.7010702@iivip.com>
To: Petr Kuba <kuba@optimsys.cz>
CC: www-voice@w3.org
Petr,

Did you quote the correct email? I don't see any discussion of strict vs 
not-strict
there. I'm confused about how STRICT is related to the discussion of 
line-by-line processing
in that email.

Regards,
Chris

Petr Kuba wrote:
> Hello Chris,
>
> Have you considered the following email from RJ?
>
> http://lists.w3.org/Archives/Public/www-voice/2010JulSep/0003.html
>
> I believe it gives pretty good guide for this issue.
>
> If I understand the new text from RJ the basic behavior is STRICT but 
> no-STRICT optimization is allowed. However, then you should note the 
> last sentence:
>
> "Application developers SHOULD NOT depend on this behavior and SHOULD 
> instead assume code is executed line by line for maximum portability 
> between implementations."
>
> Therefore I reported the problems because I expect that the 
> Implementation Report SHOULD also assume code is executed line by line.
> Then the tests will pass in both STRICT and no-STRICT modes.
>
> Regards,
> Petr
>
> On 2.8.2010 18:15, Chris Davis wrote:
>> Hello www-voice,
>>
>> We suspect that Optimsys' repeated issues of "undeclared vars" keeps 
>> occurring
>> because they are running their javascript engine in "STRICT" mode.
>> Previous issues raised by Optimsys on the same subject are 727, 715 
>> and 709.
>>
>> We never saw in the spec where "STRICT" is required so we don't run 
>> our engine
>> configured that way and as a result we pass these tests.
>>
>> What the spec *does* say is only "Attempting to assign to an 
>> undeclared variable causes an
>>    |error.semantic", and even then that is listed just for 
>> the<assign>  tag. We don't see how
>> that demands STRICT, because we assume this is just for the<assign>  
>> tag.
>> We check that with a pre-pass and thus pass tests that check such 
>> behavior (like #729).
>>
>> We request that a final ruling be made: STRICT or no-STRICT? The 
>> decision should go in
>> the recommendation. If STRICT then you could also strike defining 
>> some behaviors of STRICT
>> such as the text under the<assign>  tag. If the decision is no-STRICT 
>> then issues 737,727,715, and 709
>> should all be rejected.
>>
>> We lobby for a no-STRICT decision, as this would allow the most 3rd 
>> party javascript to run inside CCXML.
>> It has been our observation that many web-browsers(like firefox) run 
>> in no-STRICT mode and as a result
>> there is a huge amount of no-STRICT code out there.
>>
>> Regards,
>> Chris
>> |
>>
>> -- 
>> Chris Davis
>> Interact Incorporated R&D
>> 512-502-9969x117
>>
>
>


-- 
Chris Davis
Interact Incorporated R&D
512-502-9969x117
Received on Tuesday, 3 August 2010 14:26:40 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 August 2010 14:26:47 GMT