User-Agent: Mozilla/5.0 (Macintosh; U; PPC Mac OS X 10.5; en-US; rv:1.9b3pre) Gecko/2008020201 SeaMonkey/2.0a1pre Build Identifier: Mozilla/5.0 (Macintosh; U; PPC Mac OS X 10.5; en-US; rv:1.9b3pre) Gecko/2008020201 SeaMonkey/2.0a1pre These math accents are supposed to stretch. In current trunk, they don't. Reproducible: Always Steps to Reproduce: 1. 2. 3.

Hmmm, in the latest nightlies (Mozilla/5.0 (Macintosh; U; PPC Mac OS X 10.5; en-US; rv:1.9b5pre) Gecko/2008030701 SeaMonkey/2.0a1pre), the \overline and \hat accents in the testcase are screwed up as well.

(In reply to comment #4) That sounds like bug 421955.

It sure looks like it is. But will be good to check back when that bug is fixed.

In the latest nightlies, the \hat (and \widehat) accents have entirely disappeared.

(In reply to comment #7) > In the latest nightlies, the \hat (and \widehat) accents have entirely > disappeared. I don't know why that's changed. ^ represents U+005E and which seems to render (even if it doesn't stretch), but this test case uses U+0302, which doesn't render. Would you mind filing a separate bug on that being missing, please? Do you know how often U+0302 is used for accents? It doesn't seem to be in the sample operator dictionary: http://www.w3.org/TR/2007/WD-MathML3-20071214/appendixb.html#oper-dict.entries

>^ represents U+005E and which seems to render (even if it doesn't stretch), but this test case uses U+0302, which doesn't render. Hmmm. This was changed in itex2MML 1.1.2, to work around a then-current (5/2006) bug in Gecko. I could change it back, but there would still be content out there, created in the past two years, using U+0302. >Would you mind filing a separate bug on that being missing, please? I will do so tomorrow.

This bug still seems to be present in the latest Gecko Trunk

(In reply to comment #9) > >^ represents U+005E and which seems to render (even if it doesn't stretch), > but this test case uses U+0302, which doesn't render. > > Hmmm. This was changed in itex2MML 1.1.2, to work around a then-current > (5/2006) bug in Gecko. > > I could change it back, but there would still be content out there, created in > the past two years, using U+0302. > U+0302 is defined as a combining circumflex. It's purpose is to display a circumflex over the following character thus combining this pair of Unicode characters to form a single displayed symbol. As such, IMHO it should never stretch.

This is the same as the original testcase with the exception of using U+005E for the hat rather than the incorrect U+0302. This testcase renders correctly in builds that contain the patches attached to bu 519126, bug 524275 and bug 414277.

(In reply to comment #11) > U+0302 is defined as a combining circumflex. It's purpose is to display a > circumflex over the following character thus combining this pair of Unicode > characters to form a single displayed symbol. > > As such, IMHO it should never stretch. I realize it is incorrect to say it should never stretch. I suppose if the following character stretches it could be stretched to the same width. But the point is that by definition, it should always be only over the following character, so should not stretch to cover the width of the 3 characters in the given testcase. Sorry for the bugspam.

> This is the same as the original testcase with the exception of using U+005E > for the hat rather than the incorrect U+0302. I think your right. MathML Operator Dictionary (and also our version) contains U+005E but not U+0302.