75034 - Bidi: selection highlight on visual page is in the opposite position

Status: RESOLVED FIXED

(Core :: Layout: Text and Fonts, defect)

Description

[Maha Abou El-Rous]

From Bugzilla Helper:
User-Agent: Mozilla/4.0 (compatible; MSIE 5.01; Windows 98)
BuildID:    00000000 - Bidi build dated 20010404

Bidi: When double clicking on a word to select it, the opposite position is the 
one highlighted

Reproducible: Always
Steps to Reproduce:
1.Open file cp864.htm
2.Double click any word to select it


Actual Results:  Notice that the highlight denoting the selection is in the 
opposite position

Expected Results:  the word you double click on is the one that is highlighted

When you copy the selection to the clipboard and paste into notepad, the word 
you double clicked on is the one pasted (correct action).

Comment 1

[Maha Abou El-Rous]

Attachment: Arabic visual file

Comment 2

[Frank Tang]

looks like bidi selection highlight drawing issue. move this to moz0.9.1. Focus 
on landing IBMBIDI without breaking latin and CJK on moz0.9

Comment 3

[Frank Tang]

erik resign. reassign all his bug to ftang for now.

Comment 4

[Frank Tang]

mark it as assign

Comment 5

[Frank Tang]

decommit other bidi issue until we land IBMBIDI default on.

Comment 6

[Frank Tang]

reassign to [e-mail]

Comment 7

[Andreas Becker]

Changing QA contact to [e-mail].

Comment 8

[Maha Abou El-Rous]

this bug depends on the fix for bug # 81664

Comment 9

[Frank Tang]

move to "Arabic/Hewbrew" component

Comment 10

[Simon Montagu :smontagu]

*** Bug 94856 has been marked as a duplicate of this bug. ***

Comment 11

[Zach Lipton [:zach]]

Mass-move all BiDi Hebrew and Arabic qa to me, [e-mail]. 
Thank you Gilad for your service to this component, and best of luck to you 
in the future.

Sholom. 

Comment 12

[Katsuhiko Momoi]

Examples are Arabic though the issue does not seem to be
Arabic only. For now, QA contact back to maha.

Comment 13

[Zach Lipton [:zach]]

*** Bug 99785 has been marked as a duplicate of this bug. ***

Comment 14

[Christopher Hoess (gone)]

*** Bug 101339 has been marked as a duplicate of this bug. ***

Comment 15

[Simon Montagu :smontagu]

The checkin to bug 81664 has fixed the Arabic case, but this bug still exists in
Visual Hebrew and text with {unicode-bidi: bidi-override; direction: rtl} or
<BDO dir="rtl">

I have a patch which needs some polishing, will attach it soon.

Comment 16

[Maha Abou El-Rous]

fixed

Comment 17

[Simon Montagu :smontagu]

Reopening: as I said before, Visual Arabic is fixed but other cases from the 
bugs duped to this aren't

Comment 18

[Simon Montagu :smontagu]

Attachment: Patch for Visual Hebrew and bidi-override

Comment 19

[Marc Attinasi]

Comment on attachment 54070 [details] [diff] [review]
Patch for Visual Hebrew and bidi-override

sr=attinasi

Comment 20

[kinmoz]

Are these 2 'if' expressions really equivalent?


+   *  if ( (!isBidiSystem && isOddLevel) ||
+   *       (isBidiSystem &&
+   *        ((isRTLChars && !isOddLevel) ||
+   *         (!isRTLChars && isOddLevel))))
+   */
+  if (isRTLChars ^ (isOddLevel && isBidiSystem)) {


It seems like the first expression in the comment (!isBidiSystem && isOddLevel) 
would cause the 'if' clause to be entered, where as it is not possible in your 
new 'if' statement without factoring in the value of isRTLChars? Or is there an 
assumption here that isRTLChars guaranteed to be PR_TRUE if !isBidiSystem and 
isOddLevel?

Comment 21

[Simon Montagu :smontagu]

kin: good catch. That was a typo for

  if (isOddLevel ^ (isRTLChars && isBidiSystem))

Comment 22

[Simon Montagu :smontagu]

Attachment: New patch addressing kin's comments

Comment 23

[kinmoz]

Comment on attachment 54695 [details] [diff] [review]
New patch addressing kin's comments

r=[e-mail]

Comment 24

[Simon Montagu :smontagu]

Fix checked in

Comment 25

[Jerry]

http://bugzilla.mozilla.org/show_bug.cgi?id=101339

the bug on mozilla 097 is still existed,
but When I checked the source code, the fix is there.

Is there something wrong with the fix?

Comment 26

[Simon Montagu :smontagu]

Jerry: can you specify the platform you are testing on and the test page you are
using? I don't see a problem in either Win2k (with Hebrew and Arabic support) or
Linux, on any of the pages referenced in this bug or the bugs duped to it.

If you are using big-endian Solaris, there are known bugs in 0.9.7 that should
be fixed in an up-to-date nightly.

Comment 27

[Jerry]

The testcase is 
http://mozilla.org/quality/browser/standards/xhtml/transitional/bdo.xml

I running mozilla 097 on win2k. it is wrong.

but when I run mozilla 094 on win2k, it is OK.

An more serious problem is as below:
when I use mozilla 094 source code , build it on solaris, 
and access the testcase, it is wrong.

When I debug the sourcecode, it didnot go though AdjustSelectionPointsForBidi 
function, in fact, in the code of 
GetBidiProperty(aPresContext, nsLayoutAtoms::embeddingLevel,
                         (void**) &level,sizeof(level));
it gives two different Level result on win2k platform and solaris platform.

Comment 28

[Simon Montagu :smontagu]

You are seeing bug 115921 (for the bdo testcase) and bug 116976 (the levels not
being set correctly on solaris). Both are fixed in current nightlies, so you
should have no problem in 0.9.8.