209468 - Hebrew Diacritics cause justifed text to be inappropriately shifted

Status: RESOLVED FIXED

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

Description

[Eyal Rozenberg]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.5a) Gecko/20030605
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.5a) Gecko/20030605

In

http://www.earendil.ath.cx/index.shtml.he

observe the second H2 element on the page. One sees not only the manifestation
of bug 60546 (http://bugzilla.mozilla.org/show_bug.cgi?id=60546), but also a
shifting to the right of the header text, which causes some of it to go
off-window. Note also that it is not possible to scroll right and view the text.
Regardless of whether the diacritics appear in their correct position, no such
shift should occur.


Reproducible: Always

Steps to Reproduce:

Comment 1

[Leston Buell]

I can't reproduce this using the Mozilla Linux trunk 2003-06-13 build. Perhaps
you could attach a screenshot. I am assuming that the line in question is one
that says "Mi hu Earandil?" Is that right? For the build i mention, the line is
aligned just like the others.

Comment 2

[Eyal Rozenberg]

Attachment: mis-rendering of the page mentioned in the initial comment

Comment 3

[Simon Montagu :smontagu]

Can you also attach a reduced testcase? The URL gives me a page saying "This
page has been temporarily blocked".

I'm not sure that this is a separate bug from bug 60546, I think it's just a
knock on effect from the incorrect character positioning metrics for the diacritics.

Comment 4

[Leston Buell]

What is odd is that with my Linux installation, the diacritics appear after the
character rather than on the character, as described in 60546, but at the same
time i don't see the misalignment described in this bug. If you look at the
screenshots for some of the bugs duped to 60546 you see the bad diacritics but
no misalignment with respect to the margin, while in others you see the
misalignment, as well.

Comment 5

[Eyal Rozenberg]

Attachment: Reduced Testcase

Note that if the 'align: justify' is removed, the text is rendered properly on
my box, including no manifestation of bug 60546

Comment 6

[Eyal Rozenberg]

Attachment: Screenshot of reduced testcase

Comment 7

[Eyal Rozenberg]

Comment on attachment 125821 [details]
Screenshot of reduced testcase

New screenshot done with build 2003-06-17-04 .

Comment 8

[Leston Buell]

Attachment: testcase which specifies encoding and shows effect of "justify"

I tried to view your testcase using three different encodings, but the
diacritics all came out as question marks. I'm attaching a testcase which both
specifies the encoding (so everyone can view it) and has divs with and without
"text-align: justify" (so that the difference between them can be seen).

Comment 9

[Shoshannah Forbes]

confirming with Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5a)
Gecko/20030721 Mozilla Firebird/0.6

Comment 10

[Shoshannah Forbes]

Also, in the attached testcase, the first character falls of the page for me.
Attaching a screen shot.
(I am running winXP)

Comment 11

[Shoshannah Forbes]

Attachment: text falls up the page

Comment 12

[Eyal Rozenberg]

Changing bug summary to better suit the testcase

Comment 13

[Eyal Rozenberg]

Changing bug summary to better suit the testcase

Comment 14

[Uri Bernstein]

Attachment: screenshot with 2006-11-02 trunk on XP

For the record, this is what the testcase looks like on a recent trunk (Cairo-enabled) WinXP build. Both the justified and unjustified lines are now equally screwed up :-(

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a1) Gecko/20061102 Minefield/3.0a1

Comment 15

[Simon Montagu :smontagu]

(In reply to comment #14)
> For the record, this is what the testcase looks like on a recent trunk
> (Cairo-enabled) WinXP build. Both the justified and unjustified lines are now
> equally screwed up :-(

I suggest opening a new bug in the GFX:Thebes component.

Comment 16

[Simon Montagu :smontagu]

The testcases are all now working in trunk builds, so this was fixed by bug 367177.