334203 - line underneath inline mathml isn't shifted down resulting in text overlap

Status: RESOLVED WORKSFORME

(Core :: MathML, defect)

Description

[Chris Chiasson]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.0.2) Gecko/20060308 Firefox/1.5.0.2
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.0.2) Gecko/20060308 Firefox/1.5.0.2

Find the paragraph that begins:
The objective function is minimized at

After that text is a MathML equation. Back when I made this page a month ago, that MathML rendered correctly. Now, there is a problem with the line below it not being shifted down, so the text overlaps. This problem is not present on other pages on my site, like this one:

http://chris.chiasson.name/Engineering_Optimization/ch01.xhtml

Anyway, has something changed about the MathML rendering in recent FF updates? Am I just loony? :-/

Reproducible: Always

Steps to Reproduce:
1. Load http://chris.chiasson.name/Engineering_Optimization/ch04.xhtml
2. Find "The objective function is minimized at" and look at the equation after it.
3. The equation overlaps the text below.

Actual Results:  
The equation overlaps the text below it.

Expected Results:  
The equation should cause a shift in the line height of the line below it, adjusting the text so it doesn't overlap.

http://chris.chiasson.name/Engineering_Optimization/ch01.xhtml

^^ doesn't have this problem but I don't know why.

Comment 1

[rbs]

*** This bug has been marked as a duplicate of 324857 ***

Comment 2

[Phil Ringnalda (:philor)]

Can't be Cairo, since both the reporter and I see it with Firefox 1.5.0.2 (this is probably where I get yelled at for choosing Trunk while moving out of Firefox, but then, either choice gets me yelled at).

Comment 3

[Phil Ringnalda (:philor)]

Ah, good, trunk: it works in a 20051231 trunk build, and fails in a 20060128 build.

Comment 4

[Jungshik Shin]

(In reply to comment #2)
> Can't be Cairo, since both the reporter and I see it with Firefox 1.5.0.2 

I can't see it with 1.5.0.2

Comment 5

[Phil Ringnalda (:philor)]

I'd blame fonts, but I see the same thing on a Mac where I just installed the Mathematica 4.1 fonts and on Windows where I've had a full load of working math fonts for years.

Comment 6

[Chris Chiasson]

In reply to comment #4, make sure you loaded the correct page - I gave an example of a page where the bug did not appear and an example of a page where the bug did appear.

Comment 7

[Chris Chiasson]

Phil, are you using Mathematica or another program that generates <semantics> elements? I just noticed a formatting problem, where a table border was overlapping MathML, that went away when I used some XSLT to remove the <semantics> elements.

Comment 8

[Frédéric Wang (:fredw)]

The link is currently broken. Can someone gives a testcase that demonstrates the bug?

Comment 9

[Chris Chiasson]

I think the correct URL is now http://test.chris.chiasson.name/Engineering_Optimization/Linear_Programming_HW3.xhtml. However, I may have changed the page in the intervening time, as the text I pointed out doesn't appear verbatim. The part of the page I think I was referring to now occurrs in the paragraph following 
Table 4.5. P 4-4 Canonical Form of Initial Tableau. It has a vector solution to an optimization problem in the middle of the paragraph, so Firefox has to adjust the line heights. I think I may have worked around the problem back then by removing the semantics tag, since I don't see it anymore. BTW, perhaps the status of this bug should have been "confirmed" based on others comments above.

Comment 10

[Frédéric Wang (:fredw)]

Thanks.

I'm not able to reproduce this bug. According to comment 2, this bug happens before the switch to Cairo, so it may have been fixed. How did you generate the <semantics/> tag?

Comment 11

[Frédéric Wang (:fredw)]

Attachment: Testcase

Comment 12

[Chris Chiasson]

If you have Mathematica, you can create markup similar to what I originally had using the following command:

ExportString[{{Subscript[X, 1]}, {Subscript[X, 2]}} == {{2/3}, {2/3}},
   "MathML", {"Semantics" -> True, 
   "NamespacePrefixes" -> {"http://www.w3.org/1998/Math/MathML" -> 
      "ns303"}, "ElementFormatting" -> None}] /. 
 x_String :> StringReplace[x, "" -> ""]

It produces the following output, which still doesn't break the current Firefox.

<ns303:math xmlns:ns303='http://www.w3.org/1998/Math/MathML' \
xmlns='http://www.w3.org/1998/Math/MathML'><ns303:semantics><ns303:\
mrow><ns303:mrow><ns303:mo>(</ns303:mo><ns303:mtable rowspacing='1ex' \
columnspacing='1em' rowalign='baseline' \
columnalign='center'><ns303:mtr><ns303:mtd><ns303:msub><ns303:mi>X</\
ns303:mi><ns303:mn>1</ns303:mn></ns303:msub></ns303:mtd></ns303:mtr><\
ns303:mtr><ns303:mtd><ns303:msub><ns303:mi>X</ns303:mi><ns303:mn>2</\
ns303:mn></ns303:msub></ns303:mtd></ns303:mtr></ns303:mtable><ns303:\
mo>)</ns303:mo></ns303:mrow><ns303:mo></ns303:mo><ns303:mrow><ns303:\
mo>(</ns303:mo><ns303:mtable rowspacing='1ex' columnspacing='1em' \
rowalign='baseline' \
columnalign='center'><ns303:mtr><ns303:mtd><ns303:mfrac><ns303:mn>2</\
ns303:mn><ns303:mn>3</ns303:mn></ns303:mfrac></ns303:mtd></ns303:mtr><\
ns303:mtr><ns303:mtd><ns303:mfrac><ns303:mn>2</ns303:mn><ns303:mn>3</\
ns303:mn></ns303:mfrac></ns303:mtd></ns303:mtr></ns303:mtable><ns303:\
mo>)</ns303:mo></ns303:mrow></ns303:mrow><ns303:annotation-xml \
encoding='MathML-Content'><ns303:apply><ns303:eq \
/><ns303:list><ns303:list><ns303:apply><ns303:ci>Subscript</ns303:ci><\
ns303:ci>X</ns303:ci><ns303:cn \
type='integer'>1</ns303:cn></ns303:apply></ns303:list><ns303:list><\
ns303:apply><ns303:ci>Subscript</ns303:ci><ns303:ci>X</ns303:ci><\
ns303:cn type='integer'>2</ns303:cn></ns303:apply></ns303:list></\
ns303:list><ns303:list><ns303:list><ns303:cn \
type='rational'>2<ns303:sep \
/>3</ns303:cn></ns303:list><ns303:list><ns303:cn \
type='rational'>2<ns303:sep \
/>3</ns303:cn></ns303:list></ns303:list></ns303:apply></ns303:\
annotation-xml></ns303:semantics></ns303:math>

Maybe I am misremembering the example, but it just appears like Firefox is handling this correctly now.

Comment 13

[Chris Chiasson]

BTW, that stringreplace rule is supposed to say "" -> "=". It is replacing a character that most computers won't be able to display with a regular equal sign.

Comment 14

[Frédéric Wang (:fredw)]

Thanks for these informations. I haven't got Mathematica. I copy your code inside an XHTML paragraph, but I still can't see the bug. I guess the bug has been fixed since your initial report.