mozilla :: #mathml

12 Jun 2014
12:08fredwphysikerwelt: hi
12:09fredwI'm trying to think about the modules/ext.math.css stylesheet
12:09fredwAs I see, we don&#39;t use the PNG fallback for IE < 9
12:10fredwphysikerwelt: I was wondering why &quot;display: none;&quot; is done in getFallbackImage
12:11fredwor actually getHtmlOutput
12:22physikerweltI think we won&#39;t support PNG fallback in the long run.
12:23physikerweltCurrently PNG fallback requires texvc to be installed. That does not make sense at all
12:23physikerweltFor me PNG is an easy way to compare the texvc output with the Mathoid output
12:25physikerweltI&#39;ll focus on the bug with the missing MATHML tags in the HTML output
12:27physikerweltIt would be highly interstign to see if we find a way to execute JavaScript code before the page is fully loaded. This would allow for much better Javascript based decission of the output w.r.t. MathML SVG or PNG
12:29fredwphysikerwelt: I&#39;m trying to clean up a bit the setup and do the setup in the stylesheet
12:31fredwI wonder if the work in correctSvgStyle could just be done in the stylesheet
12:35physikerweltI&#39;m not sure if a style sheet is able to include information from a linked file
12:37physikerweltthe size information is copied from the SVG file
12:37physikerweltotherwise the svg images are very very large
12:40fredwyes, I overlooked that
12:40fredwAnyway, I&#39;ll just try to remove the &quot;display: none&quot; and do the selection in the CSS
12:42fredwphysikerwelt: should the SVG have vertical-align: middle; ?
12:43physikerweltGabriel proposed the display none as fallback for the case that no css stylesheets are used ... For some reason, I don&#39;t remember at the moment but which sounded reasonable when he made this point
12:44fredwphysikerwelt: so you are only doing display=none for the PNG
12:45physikerweltI&#39;m unsure there are different opinions on baseline vs middle
12:49ranierefredw: fscholz: jkitch: KhaledHosny: qheaden: teoli: WG9s: Our meeting will start in ~10 minutes
12:50qheadenOkay, great.
13:01ranierefredw: fscholz: jkitch: KhaledHosny: qheaden: teoli: WG9s: Can we start our meeting?
13:02jkitchfine with me
13:02ranierethe topics for this meeting are available at
13:04raniereStarting with the &quot;Important topics&quot;
13:05raniereI and fredw had our paper accept for MathUI and we will attend the conference
13:05WG9sraniere: shure go ahead.
13:06raniereand we also propose a demo section
13:07raniereSince the conference will be on July 10th I want to propose do next meeting on July 17th
13:08ranierethis is suitable?
13:08WG9sthat date is ok with me.
13:09fredwok for me too
13:10raniereMozilla Science Lab will run a Multi-Site Sprint on July 22-23
13:11raniereany small project needing a kickstart will be welcome and also any long time project that want to participate
13:12ranierethey will be rooms at most of all Mozilla Offices (full list at
13:13ranierefredw: fscholz: teoli: WG9s: any suggestion of project for this sprint?
13:16fredwraniere: I hadn&#39;t really thought about that, but perhaps we should see with some LaTeXML people.
13:17fredwSome of them had some stuff with Wiki and git to help writing & publishing papers.
13:17ranierefredw: I will send a email
13:19raniereMoving to &quot;Update of last meeting&quot;
13:19ranierefredw and physikerwelt talk a little about MathML at MediaWiki before the meeting
13:20fredwraniere: and
13:20raniereqheaden: you can view the log at
13:21fredwI&#39;ve been reviewing physikerwelt&#39;s patches for MediaWiki.
13:21fredwAt this point, the new MathML+SVG/PNG fallback modes is now merged in the development branch
13:22fredwHowever, there was a problem when testing on the MediaWiki machines. physikerwelt is investigating
13:23ranierefredw: thanks for the links and the update
13:24raniereMoving to &quot;Web developments&quot;
13:24fredw(for now the PNG mode is still the default. registered users can try MathML+SVG/PNG fallback)
13:25ranierefredw did some update in some addons for Firefox and create new ones.
13:25ranierefredw: Could you talk about it?
13:26fredwIn Gecko 31 (currently in beta) we will have the new support for OpenType MATH fonts
13:26qheadenraniere: Thanks for the log link.
13:26fredw allows to select the math fonts and size for the rendering of equations
13:27fredwAlso, allows to zoom-in equations, like MathJax does
13:27fredw is an experimental add-on to convert images of formulas into MathML
13:27fredwWhen the LaTeX source is provided in the alt text
13:28fredwit uses the LaTeXML web service to do the conversion
13:28fredwFor details:!topic/
13:30ranieresome of fredw&#39;s addons need help with localization:
13:31raniereany help will be grateful
13:31ranierefredw: I will send pull request for portuguese localization.
13:32raniererelated with Firefox OS, I receive my device three weeks ago
13:33raniereyou will find the link to some blog post about it at lines 36 and 37 from the PAD
13:34fredw(also, we skip that but TeXZilla 0.9.7 is released:
13:34fredwWG9s: did you actually apply to the program?
13:35WG9si tried but it was already closed.
13:35raniereyesterday I finally buit my own image of Firefox OS. There still some problems with my build but after I solve it I plan to make it available if anyone want to use it
13:35fredwOK, I think they said they will launch another campaign again at some point.
13:35fredw(I didn&#39;t receive my tablet yet either)
13:35ranierefredw: sorry to skip TeXZilla =(
13:36fredwraniere: which version of Gecko are you using?
13:36ranierefredw: now 33
13:36fredwDid you also see on the tablet?
13:37raniereit is almost the nightly build (Mozilla has a nightly build but only available to employers)
13:37fredwI&#39;m not aware of that
13:37fredwI thought Mozilla source code was 100% open :-)
13:38ranierefredw: gecko is 100% open source (at least without the DRM stuff)
13:38fredwSo what is this nightly build for employers?
13:38ranierethe problem with Firefox OS is the kernel and some libraries/drives that aren&#39;t
13:39ranierethis nightly build for employers has the last version of proprietary libraries/drivers
13:40ranierethe community has to pull this libraries from the device and copy it back when flashing it
13:40fredwok, I see
13:41fredwFor the demo at the MathUI, I think it would be good to try to fix MathML bugs on mobile
13:41ranierebecause of this proprietary libraries/drivers no one directly conect with Mozilla can make public available a image/rom for the tablet =(
13:42fredwSo normally you can not legally give use a copy of your image either
13:42ranierefredw: yes, people are still talking about it at #TCP
13:43fredwgive *us*
13:43raniereand send links peer-to-peer
13:44fredwIn, we see that the math fonts are missing from the tablet.
13:44fredwI wonder if you should try to apply the patch for
13:44ranierefredw: I will try to build Firefox OS with your patches
13:44fredw(or to bundle the WOFF fonts in our demos)
13:45ranierefredw: this screenshots was before I flash my own build
13:46fredwYes, but I guess the fonts are still missing
13:46raniereI think that it was running gecko 29 when I took the screenshots
13:46fredwAlso, I think we have the same issue for some keys on the virtual keyboard
13:47ranierefredw: yes.
13:48ranieresince we are already talking about gecko&#39;s bugs, lets move to next topic
13:48fredwraniere: I think should be enough. But I don&#39;t know if you see the font-inflation bug with the tablet too.
13:49WG9sSpeaking of bugs shold I add all the bugs mentioned in the etherpad that have patches to my builds?
13:50fredwWG9s: I don&#39;t think so
13:50ranierefredw: could you comment the bugs from the PAD?
13:50fredwraniere just added bugs that have changed status
13:50WG9syes looking for an opionion on which might be rady to include in my builds.
13:51fredwWG9S: I think the main one missing is columnspacing/rowspacing
13:51fredwroc finally agreed to add support for these attributes, but I don&#39;t know if jkitch&#39;s patches are really up to date
13:52jkitchthey need some rebasing, but other than that only minor changes are needed.
13:52jkitchI&#39;ll upload a new set in the next day or so.
13:53fredwjkitch: ok thanks
13:53fredwI think this bug as well as bug 687807 are currently the main issues blocking the &quot;enable native MathML in MathJax&quot;
13:54fredw(and of course the &quot;missing font&quot; bugs I&#39;m working on too)
13:54fredwjkitch: have you tried to do what karl suggested for
13:55jkitchI have been distracted by other bugs, but revising the patches will take some time.
13:56raniereany other bug that we need to mention?
13:57fredwok. I didn&#39;t get a chance to look at the bug either. and why it is conflicting with the tooltip patch according to WG9s
13:57jkitchI&#39;m not even going to start until I have the revised patch mostly working.
13:58fredwyes, I think the tooltip patch is of lower priority anyway
13:58WG9sI ws trying to find time today to comment out the 687807 patch in peices to at least figure out what part of the patch causes the issue. I suspect sometihng about it is preventing the mose event from proagating so that it does not interfere with the tooltip patch just the mouseover event is not happening.
14:00fredwWG9s: ok, that makes sense. Does it break also maction features?
14:00WG9snot sure i can test that i guess
14:01fredwperhaps you can try to do some addEventListeners
14:01fredwBut anyway, I&#39;d suggest to drop my patch and keep jkitch&#39;s one in your build, if the conflicts is a problem
14:02fredwraniere: the other important bug was &quot;Use Open Type MATH constants for fractions, stacks, radicals and scripts&quot;
14:03fredwthe progress has been a bit erratic because of various bugs on Windows. but based on jkitch&#39;s feedback, I think I&#39;ll be able to ask a review again soon.
14:05fredwI guess that&#39;s essentially all for Gecko
14:06raniereAnd the last topic is &quot;Documentation&quot;
14:06ranierefredw has update MDN based on the new font support
14:08fredwYes, these will be the new instructions for Gecko 31
14:08ranierefredw: fscholz: jkitch: physikerwelt: qheaden: teoli: WG9s: Any last comment?
14:09fredwPlease read the MDN page and check if the new instructions are clear. And feel free to improve the pages by fixing the wording or adding missing info.
14:10WG9sfredw: i was thinking keep the patch but remove the entry on my builds page that there i a patch that adds maction@tooltip that way I can continue to track merge conflicts but not confuse people running my builds.
14:11fredwWG9s: yes, I guess that could work too.
14:11fredwIf you don&#39;t mind hiding to people some patches included in your build
14:12WG9sI guess i caould just add a note expalining that it is not currently working
14:12WG9sor a note explaining that I guess I can&#39;t type! ;-)
14:13fredwAbout, I&#39;m still lobbying to have MATH fonts installed by default on systems and to ask font designers to fix the known bugs.
14:14fredwAt the moment Cambria Math is on Windows and I hope a MATH font will be installed on Mac too (this will be required for WebKit MathML).
14:14ranierefredw: how can I help lobbying?
14:15fredwOn Android, I created a git repository and asked developers about how to integerate my changes, but no reply so far.
14:15fredwThe Debian maintainer updated STIX to have the MATH version.
14:15fredwAnd on FirefoxOS, we are working on that :-)
14:16fredwAnd the GUST and STIX groups are working on fixing the bugs, I think
14:17fredwraniere: I don&#39;t really know. But perhaps making people aware of the new support when Firefox 31 is released will help.
14:17ranierethanks for the tip
14:18raniereI believe this is all.
14:19ranierefredw: fscholz: jkitch: physikerwelt: qheaden: teoli: WG9s: Thanks very much for your presence at the meeting
14:19fredwFYI, partial OpenType MATH support also landed in WebKit Nightly but works only on the Linux ports and more or less on the Windows one. Not mac yet
14:20fredwAnd Khaled is working on integrating that to (a fork of) gtkmathview
14:20fredwTo use that in OpenOffice, I guess.
14:22ranierefredw: fscholz: jkitch: physikerwelt: qheaden: teoli: WG9s: Hope to meet you all again next month.
14:22raniereanujagarwal464: welcome!
14:22qheadenSame here.
14:22physikerweltsorry I was absend most of the time...
14:22anujagarwal464I missed meeting
14:22raniereanujagarwal464: sorry about what?
14:23physikerwelthave a nice day
14:23ranierephysikerwelt: you too
14:23fredwanujagarwal464: physikerwelt for the logs, see
14:46WG9si was only absent for the very end. got a my real job emergency call.
15:38fredwphysikerwelt: so I&#39;ve opened
15:39fredwI guess the main question is how do we intend to make the new mode compatible with IE < 9
15:39physikerweltwe need to find a way to get feedback on that question
15:40fredwWe can probably do Javascript instead, but I don&#39;t really understand your problem about the fact that it is not executed before page load
15:40physikerweltI&#39;ll provide some statiscs later
15:40fredwnote: I haven&#39;t actually verified the patch on IE
15:41fredwphysikerwelt: your concern is that it will lead to page reflow (thus performance issue) if e.g. we show PNG at the beginning and then switch to SVG, right?
15:45fredwphysikerwelt: an alternative that I suggested was to keep SVG visible and PNG/MathML hidden by default, but to make the stylesheet change that for IE < 9 and Gecko. But I&#39;m not sure there is a CSS selector for IE < 9 only. Except perhaps if we use IE&#39;s special comment syntax
16:02physikerweltI&#39;ll have a look at that later... I have to prepare my lecture for tomorroy
13 Jun 2014
No messages
Last message: 46 days and 21 hours ago