Hello Mathias
On 17-Jul-01, you wrote:
>
> Hi!
>
>>>> # is some HTML entity definition for #
>>>
>>> yes but &num without ; isnt
>>> so it shouldnt be converted in a url
>>
>> IE converts it in a URL. Mozilla does as well. I'm damn sure
>> Opera clones this behaviour, as does KHTML and any other
>> non-anal modern HTML rendering application.
>
> No, ? and no. Only Voyager has a problem, and as long as there is no
> semicolon in there, the browser should never ever touch it as far as I
> know.
I gave you a frigging code example of how this is patently not true
in EVERY major browser available.
> What if you wrote "and number 2 finished second", than would become "and
> #ber 2 finished second" unless the browser checks for a semicolon.
No, and no. Entity names start with an ampersand (&), include the entity
name (case-sensitive) and end with a standards-compliant semicolon (;) -
but every browser ON THE PLANET worth using does not require that
semicolon.
.. whether we should make the parser more strict? The question has been
answered for you. No. Not on your life. It was strict and it made a mess
of many sites and a lot of people complained about how it didn't do the
same as IE. Tough shit if it's not 100% strict standards compliant.
Now.. someone pointed out notable exceptions: α and β which
do require the semicolon. Maybe this is because they are not part of
earlier standards definitions - I would assume IE would ignore the
semicolon and/or remove support for older ambiguous entities but
be more strict under certain circumstances.
THIS is worth investigating which entities are employed using these odd
quirks.. I think, however, it's simply that entities lying in the standard
Latin1 character set are parsed non-strict, but others are more
standards compliant.
α and β are technically greek characters, the � character
we use on the Amiga is actually technically the German "eszet" - it
would make sense to be strict for Unicode (in that a certain character
sequence would make a mess if improperly coded) but not for western
codepages.
If this is the case, then just for alpha and beta, we are not going to
arse around making special cases: it can stay how it is. Since the
Amiga uses Latin1 encoding almost exclusively, V will stay on the
non-strict side and not require a semicolon for every entity it
supports.
Unless someone can prove otherwise and give me a concise and quite
undisputable reason to the contrary that doesn't include being a bunch
of anally retentive, standards-hugging whiners, this is how it is.
Thanks
--
Matt <[EMAIL PROTECTED]>
_____________________________________________________________________
Voyager Mailing List - http://v3.vapor.com/
Voyager FAQ....: http://faq.vapor.com/voyager/
Listserver Help: mailto:[EMAIL PROTECTED]?Subject=HELP
Unsubscribe....: mailto:[EMAIL PROTECTED]?Subject=UNSUBSCRIBE
