MathJax version 2.0 now available

[Davide P. Cervone]

MathJax v2.0 is now available on the CDN, and for download from GitHub or the download page at

http://www.mathjax.org/download/

Version 2.0 is available on the CDN at

http://cdn.mathjax.org/mathjax/2.0-latest/MathJax.js

and the files at the

http://cdn.mathjax.org/mathjax/latest/MathJax.js

address have been switched over the v2.0 as well, but it will take some time for them to propagate out to the distributed cloud servers (I am not sure exactly how long that takes).  

During the time that the files are making their way out to the servers, there may be a mixture of files in your users' caches, and so they may need to clear their cache and restart their browser in order to get a consistent version of the files.  If this turns out to be problematic for you, you can change (temporarily) to the mathjax/2.0-latest URL instead of mathjax/latest since that is a new address that will not have any cached older versions to worry about.  Change back after a day or two when the new version has migrated to all the locations in the cloud.

See

http://www.mathjax.org/docs/2.0/whats-new-2.0.html

for details about the changes in v2.0, and some caveats about the effect of these changes on existing sites.

We anticipate a smooth upgrade from v1.1 to v2.0, but as always, let us know if you experience problems with this new version of MathJax.  Thank you for your continued support.

Davide Cervone

MathJax developer

[d^3p]

This morning I received quite a few emails from my students about math
error in the math pages. When I checked out the page, I knew MathJax
was some how not available to process. Interestingly, math was being
rendered at MathJax.org. So, I came here to see if there is any news
about this.

I think there should be a way to avoid prolong unavailability of
mathjax processing due to upgrade like this. I will change the url for
the time being as I have midterm for my classes next week and the
students can't wait to view sample test problems.

d^3p

[d^3p]

with the new version, the rendering seems little faster. :-)

d^3p

[Davide P. Cervone]

The switch didn't occur until 11:30, so it has only been the last hour
that anything was changed on the CDN. If you were getting messages
earlier than that, it is not due to the upgrade.

Sorry that you have been having problems. We are looking into two
different ways to improve the changeover in the future.

Davide

[Davide P. Cervone]

> with the new version, the rendering seems little faster. :-)

I'm not sure if this is a complaint that it is not very much faster
(it is only a LITTLE faster), or a compliment that it IS faster. Most
of the improvement will be for IE users. If your pages have less than
20 expressions each, you probably won't see much of a difference in
other browsers.

Davide

[d^3p]

it is faster and it is noticiable in some pages. it was already quite
fast before.

it was not complaint but compliment. please keep up the good work.

d^3p

[d^3p]

just checked the emails i had received from some of my students. they
were sent after 12:13pm eastern time. so i think must be due to the
change. I am still seeing the error in my web site where i have not
made the changes for the cdn url.


d^3p

[Davide P. Cervone]

OK, thanks for the kind words!

Davide

[Davide P. Cervone]

I'm working on an approach for the future that will (I hope) eliminate
the problem in the future, but it wasn't ready in time for this
release. I expect that the roll-over should be accomplished in a few
hours (and certainly within a day) if I understand the cloudfront
documentation properly.

Davide

[Davide P. Cervone]

Well, it looks like the files in mathjax/latest tree on Cloudfront aren't expiring as quickly as I hoped, and people are getting a mixture of the two versions, which causes math processing errors (since the two versions aren't interchangeable).  The files were supposed expire after an hour, but it looks like the expiration dates didn't propagate, and so they still have the 1 day expirations.  That means it may take until tomorrow for this to clear.  I'm working on trying to force the expiration to occur, but haven't accomplished that.

If you can, you should switch to 

http://cdn.mathjax.org/mathjax/2.0-latest/MathJax.js

if you are currently using mathjax/latest/MathJax.js, as that will work properly.

I regret the trouble this is causing you and your users.  We are working on the problem and hope to have it cleared up soon.  Your patience is appreciated.

Davide Cervone

MathJax developer

[Davide P. Cervone]

The MathJax CDN distribution at mathjax/latest should (finally) be working properly again, serving MathJax version 2.0.  I apologize for the disruption in service, and regret the inconvenience it caused you and your users.  Our upgrade procedures are being modified so that the problem will not occur again in the future.  

If you changed from mathjax/latest to mathjax/2.0-latest it should be safe to switch back again.  At this point, clearing the browser cache and reloading the page should restore functionality for anyone who is still experiencing problems.

Please let us know if there are any additional issues.  Again, my apologies.  Your patience was very much appreciated.

Davide Cervone

MathJax developer

[Phillip Lord]

"Davide P. Cervone" <dp...@union.edu> writes:

> MathJax v2.0 is now available on the CDN, and for download from GitHub or the
> download page at
>
> http://www.mathjax.org/download/
>
> Version 2.0 is available on the CDN at
>
> http://cdn.mathjax.org/mathjax/2.0-latest/MathJax.js
>
> and the files at the
>
> http://cdn.mathjax.org/mathjax/latest/MathJax.js

Congrats on release.

Just to let you know, my initial testing on the beta, and now this
release with my mathjax wordpress plugin

http://wordpress.org/extend/plugins/mathjax-latex/

suggests that the change over has been transparent.

Thanks!

Phil

[dethfire]

running pretty flawless on my end, no reports of bugs or annoyances.
what is on the development track now?

[Davide P. Cervone]

> running pretty flawless on my end, no reports of bugs or annoyances.

Great!

> what is on the development track now?

First, a little bit of a break. Then probably a bug fix for things
that will (inevitably) crop up. We have not yet prioritized the tasks
that are to be performed. It will partly depend on what the needs
are, and I am hoping that the sponsors, supporters, partners, and
friends will help by making their most pressing needs known. (That
was how we got linebreaking, equation numbering, SVG output, and a
number of the other new features.)

Davide

[Davide P. Cervone]

Great! I'm glad you didn't experience any problems. There was a
period where the mathjax/latest URL was out of synchonization, but
that was cleared up this morning, so I think everything should go well
from her on out. Thanks for letting us know that the WordPress plugin
is functioning well with v2.0.

Davide

[dethfire]

Hey Davide, just got a report that it doesn't work in IE9. I
confirmed. Here is an example
http://www.physicsforums.com/showthread.php?t=570158

thanks!

[Davide P. Cervone]

It works for me. What version of Windows are you using? What are the
symptoms of "not working"? (E.g., do you get math processing error
messages, or is the TeX not processed at all? Does it turn grey or
stay black?

Davide

[Larry Evans]

We had problems with both Firefox and Safari using the "latest"
version. We've forced back to 1.1. Our users reported "overlapping
layers" on Safari. Unfortunately, none of them knew how to take screen
shots. On Firefox, I had problems with the MML code not properly
adhering the to the text-indent CSS rules I had. I have something as
follows:

.message { /* First line flushed left, 2nd, 3rd, etc. indented a bit
*/
padding-left: 1em;
text-indent: -1em;
}

Then when I have inline latex, the rendered math is moved left what
looks like several ems, overlapping the surrounding text. Forcing HTML
fixes the issue.

[dethfire]

Win7.

Here is an example
http://www.physicsforums.com/misc/mjprob.gif

[Davide P. Cervone]

Can you provide a URL of a page that is affected?

I'm not sure what "overlapping layers" means. Can you get a more
precise description? (Probably not since you have got back to v1.1.)

Firefox should default to HTML-CSS already in v2.0 (but your own
personal setting can override that, so if you had set MathML using the
contextual menu earlier, it would stay with that).

Can you be more precise about what "not properly adhering" to the CSS
means? Or perhaps a screen shot demonstrating the problem? Or an
example equation that is causing problems, together with the
surrounding HTML (it is not clear how the CSS you have provided has
been applied to the equation). It is hard to diagnose the problem
from what you have given.

Davide

[Davide P. Cervone]

Also, are the Safari users on Windows or Mac OS? If Mac OS, is it Mac
OS 10.7 (Lion)? Does it continue to happen if they reload the page?
Have they cleared their cache and restarted Safari?

Davide

On Feb 28, 2012, at 12:32 PM, Larry Evans wrote:

[Davide P. Cervone]

IE9 on Windows 7 works fine for me. Have you tried clearing your
cache and restarting the browser? It may be that you have some cached
copies of v1.1 files that are causing problems when used with v2.0
files.

Try entering

javascript:alert(MathJax.Message.Log())

in the URL typein area and send me what you see.

Also, try opening the Developer Tools (in the gear menu at the upper
right), and select the Console tab. Then reload the page. See if
there are any error messages.

Davide

[dethfire]

In the console in IE9 I get

SCRIPT16389: Unspecified error.
TeX-AMS-MML_HTMLorMML.js, line 52 character 575

[Davide P. Cervone]

IE messages are so helpful, aren't they? "Unspecified error", sheesh!

Ok, I'll look into it. Do you administer the site, or are you just a
user? If you administer it, and are willing to do a test, could you
use mathjax/latest/unpacked/MathJax.js instead of mathjax/latest/
MathJax.js? This should not affect the output on your site (though it
will be a bit slower), but will give more useful line numbers and
positions. Thanks.

Davide

[Davide P. Cervone]

OK, I've been able to reproduce the issue. It seems to occur only in
IE9 standards mode, so a temporary fix would be to add

<meta http-equiv="X-UA-Compatible" content="IE=EmulateIE7" />

at the top of the <head> section of the page (before any other tags).

I'm working on a better resolution, but that should work for now.

Davide

On Feb 28, 2012, at 5:37 PM, dethfire wrote:

[dethfire]

yup that works Davide. looking forward to a permanent solution,
thanks!

Greg

[Davide P. Cervone]

I have identified the source of the issue, and am looking into what
the best work-around should be. It is actually the MathML input
preprocessor that is triggering the issue, and even though you are not
using TeX notation, the fact that you are using TeX-AMS-MML_HTMLorMML
does mean that you are loading the MathML code, and the preprocessor
does run looking for MathML. Switching to TeX-AMS_HTML is another
choice. The only down side is that IE users with MathPlayer will get
HTML-CSS output rather than NativeMML output, but they can still
change by hand using the contextual menu if they want. (This will be
important for people with screen readers, but it should be OK for
everyone else.)

Switching configurations would allow you to go back to IE=edge rather
than IE=EmulateIE7 and get better performance from IE8 and IE9.

Davide

[Davide P. Cervone]

OK, I've put a fix for the IE9 problem in place on the CDN, so you
should be able to go back to IE=edge in your pages again. People may
need to clear their caches to get the new version.

Davide

[nageh]

Awesome! Thanks for the new release! I have just updated my user
script on the English Wikipedia, and so far everything seems to be
smooth.

On a side note, both in v1.1 and in the current version the subscript
in the following line is positioned too close to its parent.
\begin{pmatrix} q_1 & 1 \\ 1 & 0 \end{pmatrix}

This does not happen when q_1 is outside of the matrix environment.

-nageh

On 26 Feb., 18:15, "Davide P. Cervone" <d...@union.edu> wrote:
> MathJax v2.0 is now available on the CDN, and for download from GitHub
> or the download page at
>
>        http://www.mathjax.org/download/
>
> Version 2.0 is available on the CDN at
>
>        http://cdn.mathjax.org/mathjax/2.0-latest/MathJax.js
>
> and the files at the
>
>        http://cdn.mathjax.org/mathjax/latest/MathJax.js
>

[Davide P. Cervone]

> Awesome! Thanks for the new release! I have just updated my user
> script on the English Wikipedia, and so far everything seems to be
> smooth.

Great to hear. Thanks for the support.

> On a side note, both in v1.1 and in the current version the subscript
> in the following line is positioned too close to its parent.
> \begin{pmatrix} q_1 & 1 \\ 1 & 0 \end{pmatrix}
>
> This does not happen when q_1 is outside of the matrix environment.

I'm not sure I'm seeing it. Here is a snapshot of your equation and
the plain q_1:

[mal.m...@gmail.com]

Sorry for the delay in responding.

Here is what I am seeing. I'm using the HTML/CSS renderer with TeX fonts on Firefox 11.0/Windows Vista. Note that the subscripts are not positioned in the same way for the two instances of q_1; this may be an issue due to font hinting and/or rounding (it is definitely not related to one q_1 being in the matrix and the other one outside of it). One can also clearly see the issue for maths expressions covered under root signs: often, the cap line does not connect to the initial hook of the root symbol.

-nageh

I'm not sure I'm seeing it.  Here is a snapshot of your equation and  
the plain q_1:

Both seem positioned the same, and while it is a little tight, it  
seems acceptable.  If this isn't what you are seeing, can you say what  
browser, OS, and MathJax renderer you are using, and what fonts are  
being used (from the About MathJax dialog box)?  Thanks.

Davide

[mal.m...@gmail.com]

PS: Here is how the same matrix looks in Wikipedia's PNG rendering mode.

[Larry Evans]

Sorry for not getting back to this sooner, I just now found the time
to make a demo page.

Please see http://aops.com/MathJax/index.html with FireFox (the only
browser for which I could get MML working at all) for a demonstration
on what broke from 1.1 to 2.0. The bottom right box should show the
formula indented to the right more.

[Davide P. Cervone]

The issue turns out to be the text-indent: -1em on the message style.
The reason this worked for you in v1.1 and not v2.0 is that in v1.1
the MathML was in class MathJax and that class has resets for a number
of CSS properties that could affect the HTML-CSS output, but in v2.0,
the MathML is in MathJax_MathML, which doesn't have the resets. I'll
have to add some of them back, but for now, you can use

<style>
.MathJax_MathML { text-indent: 0 }
</style>

or

<script type="text/x-mathjax-config">
MathJax.Hub.Config({
NativeMML: {
styles: {
".MathJax_MathML": { "text-indent": 0 }
}
}
});
</script>

to override it.

Davide

[cipr...@gmail.com]

Is anyone able to get the "Show source" back in the right click menu? I can't! How do I copy the original tex math code?

[Davide P. Cervone]

If you are using Chrome 18, there is a known issue (lloks like a bug
in Chrome) that is preventing the "show source" window from working
properly. See

https://github.com/mathjax/MathJax/issues/221

and

http://meta.math.stackexchange.com/questions/3901/mathjax-bugs-show-math-as-tex-commands-and-bold-math-in-comments

It appears that it may be fixed in Chrome 19. It looks to be this bug:

http://code.google.com/p/chromium/issues/detail?id=115585

If you are not using Chrome, then can you explain your situation in
more detail?

Davide

[ciprianmp]

Ah, thanks, yes I was rendering with latest Chrome 18.

I will switch now to FF/IE and see what I get.

Will a fix for Chrome be implemented any soon?

[Davide P. Cervone]

I do not yet have a workaround for the issue, but have some ideas.  I can't guarantee that they will work out, however.  When browser bugs are involved, you aren't always able to work around them.

Davide

[lumo]

If you are holding CTRL while clicking at Show math as / TeX commands, the source will appear in a new tab of your Chrome 18. If you're at least a little bit faithful Chrome user, I would avoid the exaggeration of this tiny glitch as a justification to switch to another browser.

Yours

LM

Dne sobota, 21. dubna 2012 17:09:23 UTC+2 Davide Cervone napsal(a):

[lumo]

One more funny thing. If you hadn't held CTRL and just opened the new window with the source, you can't see it. But if you focus on this window and press F11, Full Screen, the source will show up correctly in the full screen mode. ;-)

Dne sobota, 21. dubna 2012 20:07:42 UTC+2 lumo napsal(a):