Re[2]: [cgmo-webcgm] ISSUE: delimited string inconsistency?

[Lofton Henderson <lofton@rockynet.com>]

At 11:24 AM 5/4/2005 -0400, Benoit Bezaire wrote:
> [...]
> I have a few questions that need to be answered before I can 
> provide
> my opinion.
>
> 1) Are we talking about the DOM string value? the XCF string value?
> or
> both?

I was talking DOM string value.  I guess we need to answer the
question also for XCF, and I think the answer should be the same. 


> 2) I wonder how much importance
> should be put into parsing the string
> values in JS? Who wants to do that?

Good question.  It is why I asked about the options -- I don't
recommend that we make life simple for JS parsing, if it means
significant tradeoff elsewhere.

>  I can see a lot of get/set
> calls,
> but I expect they will be done without the actual parsing ex:
>
> obj1.setAppStructureAttr( "viewcontext",
> obj2.getAppStructureAttr(
> "viewcontext" ) );
>
> 3) Would the following string "0  0 100 100" (contains two
> spaces
> between 0  0), give your 'split' method a hard time? Would you end
> up
> with an empty string for one of the strings?

Yes, 
"  0 0   100      
100  ".split(" ") 
results in lots of empty strings.


> My opinion is that I would like for
> the XCF string value to be the
> same as DOM string value (or else I have to duplicate code).

Agreed.

> And in
> XML, a wsp+ is (#x20 | #x9 | #xD | #xA)+. Thus my preference for
> option 2.

Okay.  Using regular expressions, this JS works (tested on
"  0\t\t 0\n100     100\n"):

var wsp = /[\x20\x09\0D\x0A]+/;
var nbrs =
obj2.getAppStructureAttr("viewcontext").split(wsp);

So JS parsing convenience isn't the issue.

I don't care strongly. 

-Lofton.