hi bill
i'm not really too sure one way or the other - i'd need to examine the
html spec and have a think about it. so it's quite possible that somebody
will be able to jump in and give a definite answer straight away.
but what would be really useful is a patch submitted against cvs (cvs diff
-u) so that we can all take a good look at exactly what you're proposing.
- robert
On Thursday, August 2, 2001, at 11:33 PM, Bill Hackmann wrote:
> The HTML 4.01 specification provides that the value of a keyword/value
> pair must
> be delimited in quotes (single or double) if it contains any characters
> other
> than (upper and lower case letters, numerals, and period, semi colon,
> hyphen and
> underscore).
>
> I modified ElementAttributes.createStartTag as follows:
>
> 1) The special value "ECS_NO_ATTRIBUTE_VALUE" was being placed in the
> html occasionally
> 2) The "value" was scanned and the value enclosed as follows:
> �
> 2.1) If "attribute_quote" is false, and value is valid to be unquoted, it
> is left unquoted
> 2.2) If the "value" contains EITHER a single or a double quote, but not
> both, and the value is to be quoted, it is enclosed in the delimiter
> which does not appear in the string
> 2.3) If the "value" contains both a single and a double quote, a single
> quote is used, and embedded single quotes are converted to %#39;� It
> might be better to count and use the one which make the final string the
> shortest.
>
> examples:
>
> <input name=thisIsAnInputField_Name ...>
> need not be quoted unless all values must be due to ecs defaults.
>
> onClick=work('securities',';insurance',"true");
> Single quoted value:
> onClick='work('securities','insurance',"true");'
> which is properly executed by both IE 5.5 and NN 4.75.
>
> This implementation may introduce an incompatibility with prior ECS code,
> and so� maybe it should be controlled by an optional ECS Attribute, such
> as attribute_quote_either=true
>
> This relieves the developer from needing to examine the contents of every
> string used as a value in an attribute, and adjusting the contents to
> match the expected value in the ecsdefaults file.
>
> Is there a reason this is a bad idea?
>
> Also, how closely does the Character.isLetter(char c) function match
> "letters" as defined in the 4.01 HTML spec?� I implemented using the
> Latin alphabet, but isLetter() uses a definition which is much more broad.
>
> Comments, please.
>
> Hackmann
>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
