This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
Right-to-left scripts (Hebrew and Arabic for the purposes of this specification) may be encoded either logically (right to left) or ‘visually’ (left to right). For instance, Opera, Safari, IE and Firefox all decode the bytes E0 E1 E2 E3 as Aleph Beth Gimel Daleth (right to left) when given the label iso-8859-8-i, but as Daleth Gimel Beth Aleph (still from right to left) when given the label iso-8859-8. This does not seem to be taken into account at the moment.
What is the best way to address this? And which labels map to which variant?
Actually what happens is that they decode identically (which makes sense as I tested that already), but there is something going on layout-wise based on the label.
Gecko appears to be the only browser (have not tested IE) to treat iso-8859-8-e as a different encoding name from iso-8859-8. So they have three variants, the others two. The other label for -i is csiso88598i The above also applies to iso-8859-6. So I guess the fix here is to split out the -i and cs...i variants and treat them as separate encodings and then somewhere we have to define the effect of those encoding names on layout. Does that sound about right?
https://github.com/whatwg/encoding/commit/d2a4f321ff9d73d9db5380b954ab9dd26a18b43e http://encoding.spec.whatwg.org/ This was not needed for iso-8859-6 after all because no browser implements logic for that encoding as far as I could tell. I still need to file a bug of sorts on CSS I suppose.
I filed bug 19505 on CSS.