321438 - MathML entities φ and ϕ are reversed

Status: RESOLVED FIXED

(Core :: MathML, defect)

Description

[David Harvey]

User-Agent:       Mozilla/5.0 (Macintosh; U; PPC Mac OS X; en) AppleWebKit/312.5.2 (KHTML, like Gecko) Safari/312.3.3
Build Identifier: Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.8) Gecko/20051111 Firefox/1.5

The MathML entities φ and ϕ are the wrong way around. The entity φ should correspond to unicode 0x3D5 ("straight phi") and ϕ ("curly phi") to 0x3C6, but in Firefox 1.5 they are reversed.


Reproducible: Always

Steps to Reproduce:
1. View a MathML document containing φ and ϕ
2.
3.

Actual Results:  
φ gets mapped to unicode 0x3c6 and ϕ gets mapped to unicode 0x3d5

Expected Results:  
φ should get mapped to 0x3d5 and ϕ should get mapped to 0x3c6

It appears the problem can be traced back to a mistake in the MathML 1.x DTD spec. In that earlier version (for example see http://www.w3.org/1999/07/REC-MathML-19990707/chap6/byalpha.html), there are the lines:

phi     isogrk3       966 03C6 /straightphi - small phi, Greek
phiv   isogrk3       981 03D5 /varphi - curly or open phi 

Clearly something is wrong here , since 3C6 is NOT "straight phi", and 3D5 is NOT "curly or open phi". (See the unicode chart at http://www.unicode.org/charts/PDF/U0370.pdf).

In MathML 2.0 the error appears to have been corrected. See for example http://www.w3.org/TR/2003/REC-MathML2-20031021/bycodes.html.

It could be argued that this is not a Firefox bug because it follows a particular version of the MathML DTD faithfully. Nevertheless, as that version of the spec was clearly in error, I think it's reasonable to update Firefox's behaviour to match the corrected (MathML 2.0) DTD.

Comment 1

[AndrewM]

Does this bug still exist on the Trunk?

Comment 2

[Karl Tomlinson (:karlt)]

http://www.unicode.org/reports/tr25/#_Toc23
discusses the representative glyphs being swapped with Unicode 3.0.

"it is important to make sure that codes or entities, such as phi1, that require the straight form of the phi symbol be mapped to U+03D5 and not to U+03C6."

Comment 3

[Karl Tomlinson (:karlt)]

(In reply to comment #1)
> Does this bug still exist on the Trunk?

layout/mathml/content/src/mathml.dtd seems to still have this bug:

<!ENTITY phi '&#x03C6;'>  <!ENTITY straightphi '&#x03C6;'>
<!ENTITY phiv '&#x03D5;'>  <!ENTITY varphi '&#x03D5;'>

Comment 4

[Karl Tomlinson (:karlt)]

I think this'll be fixed at the same time as bug 289938.

Comment 5

[Karl Tomlinson (:karlt)]

Fixed on trunk in bug 289938