V supports every applicable and usable entity defined for the ISO-8859-1
(Latin1) character set.
V supports entities in this character set exactly as Internet Explorer,
Netscape Navigator, Mozilla, Opera and KHTML does.
V does not support other, non-Latin1, truly unicode glyphs or entities
and as such does not have the scope to detect when one is specified
(such as α or β - β is supported as a fluke in that we
have a similar character in Latin1 - • is exactly the same in that
I am supporting it via the same codepoint that • is allocated to)
and apply more strict rules in parsing.
As such, V's entity parser is PERFECT, for all intents and purposes, for
all behaviours of IE, for entities that represent characters inside the
Latin1 codepage. If you really want that to be ABSOLUTELY true, I will
drop support for • and β today. Please tell me if anally
retentive standards-hugging is more important than the functionality
this might provide.
This is not an issue that requires more in-depth investigation: it was
investigated back in February, and V's entity parser was updated to
match that of Internet Explorer.
3.3.45 is the revision that added support for the lack of a semicolon,
such that every browser does:
- Layouter: fixed the "broken entity" bug (Olli)
We have a precedent set - converting entities in URLs - with the
, entity, which is fixed as of 3.3.89
- Layouter: removed , entity, to avoid freak interactions
with the IE precedent of converting entites in URLs (grrr),
and IE not knowing , (grrr) (Olli)
And for the next beta we have this bug fixed, exactly the same bug,
but a different entity:
- Parser: removed the # entity so that dumbass webboards didn't
nuke when they used it as a variable name [IE5 doesn't support the
# entity, BAH!] (Neko)
Now, until V supports font character codepoints above 255, V has no
way of distinguishing between β and ß such that stricter
parsing would be required without a slight rework - a rework that
would be useless if real font character codepoints became available,
and would need to be undone.
The # bug is fixed. The implementation is precisely the same as
IE does for the entities V provides, minus two rare, untrappable
exceptions, which gives V decent and unbroken support for the
webpages out in the real world.
Your mission - Jools, Ian, Mathias - should you choose to accept it, is to
read this mail, understand that V works just fine, and quit telling me
that it isn't working fine, and that I require looking at it more than I
already have to understand something I've known to be the case and
I have prompted, testcased, overseen, and verified the fixing of already
this year.
Of course if any of your team members should be annoyed by this mail,
all reported bugs will be disavowed by the secretary :)
Regards
--
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
