User-Agent: Mozilla/5.0 (Macintosh; U; PPC Mac OS X 10.5; en-US; rv:1.9.1b5pre) Gecko/20090504 Shiretoko/3.5b5pre Build Identifier: Mozilla/5.0 (Macintosh; U; PPC Mac OS X 10.5; en-US; rv:1.9.1b5pre) Gecko/20090504 Shiretoko/3.5b5pre Sometime when the "Differentiable Manifold" is displayed at http://eyeasme.com/Joe/MathML/browser_test_bug.html the column align attribute the the mtable is not honored, instead of left aligning the default right alignment is done. (See attached picture 1). Note that when the cursor is moved into the column that it reformats itself and properly left alignes. Reproducible: Sometimes Steps to Reproduce: 1. go to http://eyeasme.com/Joe/MathML/browser_test_bug.html 2. look at "Differentiable Manifold" example 3. sometimes the column alignment is not left (as wanted) but center Actual Results: Sometimes (see attached Picture 1) the generated MathML formula is not properly column aligning to the left (as wanted) but is centering. NOTE: when cursor is moved into the generated MathML column, that the formula then re-aligns itself and becomes left column aligned. Expected Results: The mtable attribute of columnalign should be honored. Sorry for the large example, but the bug I want to show does NOT show itself for small examples (i.e isolating the "Differentiable Manifold" example row). There is some kind of interaction among the examples that triggers the bug. The strange thing is that moving the mouse to hover over the MathML column of the example fixes it. This is bad since the page should be static. I have attached an image of the example displayed incorrectly and then displayed correctly. The incorrectly displayed image should never occur.
Created attachment 375710 [details] bad alignment in the third column example Note bad alignment in the third column example
Created attachment 375711 [details] Correct alignment Correct alignment, this occured after moving mouse into the third column.
When filling out bug 504324, I noticed that the CSS property "text-decoration" did not completely show until I clicked in the table cell. (i.e the MathML equation changed its look (to what it should be) when the table cell it was in, was clicked. It appears that all the CSS code is not being applied when a page is displayed.
The example given in comment #0 is no longer is applicable . The problem is that the columnalign attribute of mtable is not working in Firefox. The work around is to use CSS to set 'text-align' to the columnalign value desired.
(In reply to comment #4) > The example given in comment #0 is no longer is applicable . > > The problem is that the columnalign attribute of mtable is not working > in Firefox. The work around is to use CSS to set 'text-align' to the > columnalign value desired. I'm not sure to understand. The columnalign attribute seems to be implemented (looking to the source code) and works for me. I try to produce a reduced testcase but can not see the bug. Can you give a testcase that shows the incorrect behaviour?
Created attachment 424098 [details] Image of the bug Notice the bug in the third column. Should be column aligned left.
I have updated example (look at third column of "Differentiable Manifold"). (click on the column to see the code (or just look at code with show Source) The problem does not show for a small test page like so https://eyeasme.com/Joe/MathML/browser_test_bug2.html But does show for a bigger test page like so https://eyeasme.com/Joe/MathML/browser_test_bug.html I have included an image of the bug in progress. The bug goes away (I.E. the formula moves to proper configuration) when the mouse enters the formula (same row and column).
Created attachment 427195 [details] Testcase While writing some MathML, I've just met the bug again (this time, the bug disappears when you zoom). Note that the link to a css seems necessary for the bug to happen (but you can use an empty css stylesheet).
It seems zooming in and out can solve this problem. Still annoying, though.
MathJax users have reported issues which looks related to this bug: http://groups.google.com/group/mathjax-users/browse_thread/thread/bbf0c78a19bc565f http://groups.google.com/group/mathjax-users/browse_thread/thread/92a71671b9e2417b I think fixing bug 69409 (or at least to map columnalign to style) is likely to solve this bug.
My example code has bit-rotted and should be ignored. My work around for the bug is to set the CSS style "text-align" in the mtd tag as well as set the "columnalign" in the table. EX: <mtable columnalign="left"> <mtr> <mtd style="text-align: left"> <!--hack for this bug is here -->
(In reply to Frédéric Wang from comment #8) > Created attachment 427195 [details] > Testcase > > While writing some MathML, I've just met the bug again (this time, the bug > disappears when you zoom). Note that the link to a css seems necessary for > the bug to happen (but you can use an empty css stylesheet). This testcase does not seem to work any more.
(In reply to Frédéric Wang from comment #12) > This testcase does not seem to work any more. Actually, the testcase does not seem to exhibit the bug with a nightly build but it does with a release version. Here is another page that demonstrates the issue (see alignments in theorem 9.2 or lemma 9.4): http://www.maths-informatique-jeux.com/international/hsp/web/appendix_B.php#id8.2.
For a more reliable testcase, see attachment 633989 [details] (MathML trees created dynamically by MathJax).
Mass change: setting priority to 2 for bugs that would be nice fix if Gecko's MathML support is enabled by default in MathJax but that are not in my opinion strictly required or for which a workaround could be written in the MathJax code.
Fixed by bug 731667.
@eric: do you remember why you set the "dev-doc-needed" flag? I think the columnalign attribute is already claimed to be supported: https://developer.mozilla.org/en-US/docs/Web/MathML/Attribute It's just that it was (incorrectly) ignored in some random cases.
(In reply to Frédéric Wang (:fredw) from comment #18) > @eric: do you remember why you set the "dev-doc-needed" flag? > > I think the columnalign attribute is already claimed to be supported: > > https://developer.mozilla.org/en-US/docs/Web/MathML/Attribute > > It's just that it was (incorrectly) ignored in some random cases. Yeah, this doesn't need docs. Not sure why I flagged it. Probably misread it during a pass through a list of new bugs once.