242861 - UnderBar accent doesn't work in MathML

Status: RESOLVED FIXED

(Core :: MathML, defect)

Description

[steve.swanson]

User-Agent:       Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; .NET CLR 1.1.4322)
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8a) Gecko/20040506

I don't see anything. If you use the caret, you see Mozilla thinks something 
is there, but it seems to have zero width.

Reproducible: Always
Steps to Reproduce:
1. Load testcase

Actual Results:  
You see an OverBar and UnderBrace accent, but not an UnderBar.

Expected Results:  
Drawn a bar.

Comment 1

[steve.swanson]

Attachment: testcase

Comment 2

[Austin Lund]

This bug seems to be fixed with the current CVS version (11/11/05).  I see all overbars and underbars and the brace.  

Comment 3

[steve.swanson]

Still fails for me.  What's your OS?  If Windows, which fonts do you have installed?

Comment 4

[Karl Tomlinson (:karlt)]

I can reproduce this with larger font sizes on Linux with Standard Symbols L.

Comment 5

[Karl Tomlinson (:karlt)]

Attachment: fix

The problem is that dy is not set correctly when bm.ascent is -ve for
UnderBar:

http://bonsai.mozilla.org/cvsblame.cgi?file=mozilla/layout/mathml/base/src/nsMathMLChar.cpp&rev=1.139&mark=2382,2402#2382

and so the UnderBar is drawn outside the clip rect.

Comment 6

[Karl Tomlinson (:karlt)]

Attachment: simpler fix

(The baseline has already been calculated in TryParts.)

Comment 7

[Karl Tomlinson (:karlt)]

Comment on attachment 315444 [details] [diff] [review]
simpler fix

Requesting approval1.9.  Although this is an old bug, changes (probably in the fonts used) seem to have made this more common, so it would seem like a regression.

The patch is small so the risk is low, but I have tested manually at http://www.maths-informatique-jeux.com/international/mathml_test/accents.xml

I have a reftest that can be landed with this.

Comment 8

[Mike Beltzner [:beltzner, not reading bugmail]]

Comment on attachment 315444 [details] [diff] [review]
simpler fix

a=beltzner assuming reftest lands with it

Comment 9

[Karl Tomlinson (:karlt)]

http://bonsai.mozilla.org/cvsquery.cgi?module=PhoenixTinderbox&branch=HEAD&cvsroot=%2Fcvsroot&date=explicit&mindate=1208489220&maxdate=1208489284&who=karlt%2B%25karlt.net