Version 28.2.x problems with MathJax

[Gammatester]

The Palemoon 28.2.x series has problems to display MathJax-Rendering in comments at https://math.stackexchange.com/, see e.g. the post https://math.stackexchange.com/question ... y-frach-n2 . With 28.2.x you see that you see no math formula in the comments.

SInce this is an essential feature I am still using version 28.1.0, which displays the page correctly. I tried all 28.2.x versions, but immediately revert to 28.1.0.

[satrow]

I'm no mathematician but we may need screenshots and more details.

[Moonchild]

No problems with default settings either.
MathJax does seem to have settings of its own, though, so maybe you accidentally changed something there that won't work in Pale Moon?

[Gammatester]

Your screen shot is not from the comments sections. How can a post imges? Meanwhile look at the comments in the answer:

Code: Select all

Isn't this summing as opposed to ? – Badam Baplan

Thank You, that is something yet. Any ideas about the second part? – Isak

@BadamBaplan: it is summing the right thing, since . – Jack D'Aurizio

nice, thanks for explaining. this is a nice approach. – Badam Baplan

@Isak: the difficult part is essentially which appears to be manageable by summation by parts. – Jack D'Aurizio

With 28.2.2 you can see no formula, the source code for the last comment from above is

Code: Select all

<span class="comment-copy">@Isak: the difficult part is essentially $\sum_{n\geq 1}\frac{H_n^{(2)}}{(2n+1)^2}$ which appears to be manageable by summation by parts.</span>
                    &ndash;&nbsp;<a href="/users/44121/jack-daurizio"
                       title="283,999 reputation"
                       class="comment-user">Jack D'Aurizio</a>

The MathJax part $\sum_{n\geq 1}\frac{H_n^{(2)}}{(2n+1)^2}$ is not displayed.

I have added two images as attachments the first from 28.1.0 the second from 28.2.2

[Moonchild]

Hmm.. weird. It indeed doesn't display, until I open the devtools, THEN it displays... and remains displayed after closing devtools again.
With there being no issue with devtools open, it becomes really difficult to troubleshoot, though.

[JustOff]

Looks like the Firefox 52 compatibility mode is required here:

Code: Select all

general.useragent.override.math.stackexchange.com -> Mozilla/5.0 (Windows NT 6.1; WOW64; rv:52.9) Gecko/20100101 Goanna/4.1 Firefox/52.9 PaleMoon/28.2.2

[Moonchild]

OK, so there's a workaround. Add a useragent override for math.stackexchange.com:

Code: Select all

Mozilla/5.0 (Windows NT 6.1; rv:52.9) Gecko/20100101 Firefox/52.9

Since the mathjax library does UA sniffing and changes code based on what it finds, it'll be difficult to find out what change exactly in Firefox code is responsible for it not displaying with the later compatibility mode version.

[Moonchild]

Raised it as an issue on https://github.com/mathjax/MathJax/issues/2096

[Gammatester]

Thank you.

After changing general.useragent.compatMode.version to 52.9 version 28.2.2 worked as expected.

[JustOff]

it'll be difficult to find out what change exactly in Firefox code is responsible for it not displaying with the later compatibility mode version.

Perhaps this can help:

Last bad: http://ftp.mozilla.org/pub/firefox/tind ... .win32.zip
First good: http://ftp.mozilla.org/pub/firefox/tind ... .win32.zip
Pushlog: https://hg.mozilla.org/mozilla-central/ ... e1febee19e

[Moonchild]

Thanks. Looks like it's bug #1338678 since its testcase displays the same devtools oddity-behavior.

[JustOff]

Looks like it's bug #1338678 since its testcase displays the same devtools oddity-behavior.

Nice catch, I can confirm this one helps! But you better port it yourself, because lately I'm trying not to alter the code, which I don't understand quite clearly. And perhaps bug #979782 also needs attention, as a follow-up.