20338 - support Hebrew visual order

Status: VERIFIED FIXED

(Core :: Layout, defect, P3)

Description

[Erik van der Poel]

There is a discussion in mozilla-i18n about Hebrew visual order, which is easy
to support:

  news://news.mozilla.org/3842BC95.F5CEB8DB%40technion.ac.il

The current bidi hacks that we have in the Unix version of Mozilla apparently
don't work with these visual order pages. Should be easy to fix. And has quite
an impact in that market, since many pages are still visual. However, this bug
is marked for all platforms and OS's, because we probably want visual order
Hebrew on all platforms.

Comment 1

[Frank Tang]

*** Bug 24869 has been marked as a duplicate of this bug. ***

Comment 2

[Frank Tang]

reassign to bi-di owner

Comment 3

[bobj]

For M16, let's do the minimal to restore Hebrew to pre-Beta1.  This means
deleting some lines from charsetData.properties and changing 1 line in
charsetTitles.properties.  Details follow (cata pls verify):

To reenable what we had prior to beta1 you need to modify
charsetData.properties by removing these lines for Hebrew
   iso-8859-8.notForBrowser = true
   ibm862.notForBrowser = true
   windows-1255.notForBrowser = true

Note, iso-8859-8 defaults to visual order.  But do we want to change this in
the future to use the suffix to make this unambiguous visual order:
    iso-8859-8-e   (See: http://www.ietf.org/rfc/rfc1556.txt)

Are there similar suffices for ibm862 and windows-1255 Hebrew encodings?

If we want to make such changes, let's file another bug and do this later.

We also should change the user visible string to prepend "Visual" => "Visual 
Hebrew".  This needs to be changed in charsetData.properties

(old) iso-8859-8.title = Hebrew (ISO-8859-8)
(new) iso-8859-8.title = Visual Hebrew (ISO-8859-8)

Note this file also contains these entries
  iso-8859-8-i.title = Hebrew (ISO-8859-8-I)
  iso-8859-8-e.title = Hebrew (ISO-8859-8-E)
but let's leave these alone for M16.

We should leave these for Arabic (so they will not show in the menus):
   iso-8859-6.notForBrowser = true
   windows-1256.notForBrowser = true
   ibm864.notForBrowser = true

I have not heard of any demand for Visual Arabic, although there was a mention
that there are visual order Persian pages.

Comment 4

[cata]

Changes look good. Go ahead...

Comment 5

[Mike Kaply [:mkaply]]

Code checked in. I viewed at least one page :)

For some reason www.netvision.net.il traps - looks widget related.

I am leaving this assigned until I figure out the trap.

My visual test is: 

http://www.qsm.co.il/Hebrew/qashetab.htm

Comment 6

[Mike Kaply [:mkaply]]

Fix is checked in. Trap on netvision was unrelated.

Comment 7

[Teruko Kobayashi]

Another Hebrew bug. 

Comment 8

[Laura Yecies]

I verified - the page appears correct.

Comment 9

[bobj]

Does this work on the Mac?  I thought there was some problem that the Mac
will try to apply its own bidi algorithm...

Comment 10

[Laura Yecies]

I verified this on Mac System 9.0

Comment 11

[moshev]

The problem with Mac Bidi will only show on Mac OS Hebrew version. Just to make
sure that this is what you tested it on. Otherwise, bidi is not present on Mac
OS IMHO, and it won't kick in, so you will see no problem.

Comment 12

[bobj]

We should open a separate bug, if there is not one already.

Comment 13

[Frank Tang]

How about English MacOS with Hebrew Language-Lite ? 
I saw the text display correctly. But when you select them, the selection draw 
incorrectly.

Comment 14

[moshev]

I am not that familiar with Mac OS, i made my remark based on general knowledge
of BiDi issues, as discussed on n.p.m.i18n
I also don't think that there should be a separate bug, as long as Visual
don't work correctly on major platforms, i believe this bug should not be marked
as fixed, but i'll leave it to mozilla members to decide...

Comment 15

[Mike Kaply [:mkaply]]

Visual Mode will not work on Bidi Mac OS until full Bidi support is in the 
browser.

On Mac you have to reverse everything and let the OS reverse it back which 
requires a Bidi enabled browser.

We have bidi changes that will drop after nsbeta2 that should make all this 
discussion moot.

Comment 16

[moshev]

Great, my only suggestion is: why not reopen this bug to serve as a QA reminder?