• Discussion Feed

Vanilla 1.1.9 is a product of Lussumo. More Information: Documentation, Community Support.

1.

@noah, almost certainly, there's some extra step to load new fonts that isn't being performed when you preview. Perhaps try editing a question that already has \mathrm, and seeing if preview works in that case.

2.

I can't reproduce Noah's bug, but it's consistent with weird jsMath behavior people have reported lately. Perhaps there's some sort of a race condition that prevents certain packages from loading sometimes. jsMath does seem to be sensitive to the order in which I load various things (things that I wouldn't expect to matter).

3.
Apologies---I'm sure this has been flagged before, but I see 152 comments and I've not been keeping up with this thread and I can't face reading all of them. The issue is that I just wrote a mathbf Q p (p-adic numbers) and a mathbf Z p (p-adic integers) with an underscore to do the subscripts, and the p-adic integers came out great (umm what am I using---looks like firefox on my girlfriend's mac) and the p-adic numbers it looks like the p fell off. See my answer at http://mathoverflow.net/questions/23477/z-p-flatness-and-irreducible-components but this has happened to me several times before.
4.

When doing mathematics with underscores or asterisks in it, you need to escape the formula with backquotes (outside the dollar signs).

5.
@harald: I'm going to call your bluff. At the time you wrote your comment there were no backquotes to escape the formulae. But this isn't for the reason you think :-) When I wrote the answer I backquoted to escape all the formulae with underscores in. It was only after that that I noticed the droopy p's. I removed all the backquotes in an attempt to fix the problem. The problem did not go away. I couldn't be bothered to put them back and I just posted. So I believe this is not the issue. As a result of your comment I just went and put them all back. The issue is still there on....umm...this time it's firefox under debian linux.
6.
PS @Harald: can you confirm that on your browser the p's under the Q's are much much lower than the p's under the Z's?
7.

Looks fine to me (firefox 3.6 on linux). The p under the Q is ever-so-slightly lower than that on the Z, but not unpleasingly so. I could send you a screenshot if you like.

8.
Here's a screenshot of what it looks like for me:

9.
PS I cannot reconcile that monstrosity with your "ever-so-slightly lower" comment so I am assuming you can see something else. I appear to be using Iceweasel 3.0.6 .
10.

This is what I see:

Comparing the two, I think that another difference is in the rendering of the Z and Q themselves.

11.

@buzzard: There is no bluff. I misunderstood your point. When you wrote “it looks like the p fell off” I thought it was a cute way of saying the p was missing or something. Usually, when things go bump in the night the first thing you think of is the backquotes (or lack thereof), hence my post. If you had been more specific, I would of course have answered differently. Anyway, what I see is again quite different than what either you or Andrew sees, but not outrageously wrong:

12.
@Harald: :-) But to both Harald and Andrew---what is going on? I am currently using Chrome 5.0.342.9 beta on a mac and I still have this p mayhem: p _really_ low on the Q. I've now seen this on a mac under firefox and chrome and on a linux box under iceweasel. Are you telling me that somehow all three set-ups are somehow not optimal? Both of you have non-droopy p's, which is what I yearn for.
13.

On Opera on Linux, I see:

which is clearly gibberish!

I don't have iceweasel installed, it's a rebranding of firefox so I don't see why it should be so different, plus you see it on firefox as well.

Looking at your screenshot, the subscript 'p's don't look any smaller than the non-subscript ones. It's a little hard to see, though, so I may be wrong.

14.

My screenshot was firefox on mac. Maybe the available fonts influence the result?

15.

I think I found a hack. I replaced "\mathbf{Q}_p" by "\mathbf{Q}{}_p". That fixed the dropping p for me.

16.

I just installed the TeX-Illustrator fonts (for the obvious purpose of using Computer Modern inside Illustrator), but now jsmath is messed up. Here's an example from http://mathoverflow.net/questions/23437/are-the-q-catalan-numbers-q-holonomic/23930#23930:

Any ideas? I hadn't wanted to installed the jsmath fonts locally for the sake of keeping my experience in sync with "new users" on mathoverflow. I'm using OS X 10.6.

17.

jsMath wraps math in a span of class "typeset" and then has lots of internal spans with classes like "cmsy10" (or other font families). There's a css rule that says that such a thing should use the font family "jsMath-cmsy10". Perhaps the cmsy10 TeX-Illustrator font is somehow taking the place of jsMath-cmsy10, but isn't exactly the same. The ≺ character appears two steps before the - character in jsMath-cmsy10, so you would expect this kind of problem if you used an almost-but-not-quite-the-same font.

If you weren't using jsMath fonts before and don't want to use them, you can always click the jsMath box in the lower right hand corner of your browser, click "options", then click the "use image fonts" radio button. That will save a cookie which tells jsMath to always use image fonts.

18.

Thanks Anton, I'd forgotten I could override the font mode.

19.

The math preview doesn't seem to be working in Chrome. Is anyone else having this problem?

20.

The math preview doesn't seem to be working in Chrome. Is anyone else having this problem?

Yes. I don't know what's going on. The live preview doesn't work in Chrome (on linux), but does work in Firefox. It works in both on another machine running Windows XP. The weird thing is that the one-shot preview works fine, so it's clear that jsMath is working properly. Hmmm ... maybe the modified wmd.js isn't getting loaded for some reason?

Edit: indeed, it looks like the custom version of wmd.js (which incorporates jsMath processing) isn't loading properly. I don't understand why. More Edit: it sometimes works and sometimes doesn't.

21.

Some lines of math are also jutting out of the margins in Chrome for some reason. (A little, uh, math overflow, if you will.) I'm using Vista.

22.

Some lines of math are also jutting out of the margins in Chrome for some reason. (A little, uh, math overflow, if you will.) I'm using Vista.

That's some weird interaction between jsMath and webkit browsers. I can't do anything about that. It doesn't seem to happen with MathJax.

23.
OK so here's something that is probably related to the jsMath (re)processing. On a mac, with chrome, *not logged in*, in this question: http://mathoverflow.net/questions/27229/polynomials-with-rational-coefficients as I type there is a line "$\boldsymbol a=(a_1,\dots,a_n)$" in the source , which, when not logged in as I say, fails to render at all and gives me "Unknown control sequence '\boldsymbol'" in bright yellow. No amount of hitting "(Re)process jsMath" fixes this. In fact, when not logged in and in chrome, this question is unreadable to me because there are six TeX errors all involving failure to understand boldsymbol.

OK so the moment I log in on chrome and reprocess jsMath, this becomes fixed.

Now here's the behaviour under firefox (still on a mac). The first time I looked at the page, the boldsymbols didn't render but didn't give me an error either. I see a partly rendered page, with blackboard-bold R's but with the word "boldsymbol" here and there. When I hit "reprocess jsMath" everything is fixed and I can finally read the question.

So what is going on? Not only is jsMath not rendering, it's somehow not rendering in two different ways. Or is this just two different browers' reactions to the same problem?
24.

Odd. It looks fine to me in firefox on the mac, but looking at the source, I notice that formulas containing underscores are not surrounded by backticks. Possibly that is part of the problem.

•
CommentAuthorJon Awbrey
• CommentTimeJun 11th 2010

I've noticed that when I come to the main site by way of a search engine the jsmath is blocked somehow. This is all in Firefox.

25.
Up until recently I was able to see the Latex, but now I can't. I am using explorer and I am using 3 different computers and in 2 different locations. Today I have checked whether I have the right chracters, I didn't. So I uploaded the characters to 2 of these computers, but it didn't seem to help.

Any suggestions? You are welcome to contact me via email if you prefer.

Thanks!
26.

Hi Yiftach, sorry you haven't received a response yet. Perhaps you could try a more modern browser, e.g. Firefox, Safari, Opera or Chrome? You might also try clearing the browser cache in Explorer in case it's has somehow mis-cached the javascript that renders the LaTeX.

27.
Hi Scott, I have tried Firefox and it works. Thanks!
28.
I've encountered a strange bug when trying to append subscripts to things of the form "\operatorname{Foo}." Specifically, I want to be able to write "\operatorname{Aut}_{\textbf{Tw}}(k[G])." I can do this, and it will work. Once! But if I put another one in, both of them turn to gobbledygook, both in the preview and in the actual post. The problem seems to occur whenever there are two instances of "\operatorname{something}_{somethingelse}"; the curly braces in the subscript seem to be relevant. Looking through the comments here, it appears that \operatorname has been problematic in the past; were all of the kinks (supposedly) worked out?
29.

@Evan: the problem is that _phrases between underscores_ get converted to italics, like this, before the math processor gets a chance to look at them. The accepted workaround is to tell markdown to leave the underscores alone by enclosing your math in backticks, like this:

$\operatorname{Foo}_{bar}$

Another alternative, assuming you don't have any markdown syntax in the paragraph, is to enclose the paragraph in P tags, like this:

<p>
I really like using $\operatorname{Foo}_{bar}$ as an operator.
</p>

30.
@Anton: Thanks! I'm sure I must've encountered (and fixed!) this problem before, but I completely forgot about it.
31.

Charles Rezk mentioned a \varprojlim problem on April 16 in this thread, but I'd like to point out that when I looked at this question, $\varinjlim$ rendered correctly in preview, but threw up an error when being displayed after submission. However, when I look at the edit history now, the old version seems to render fine.

• CommentAuthorjbl
• CommentTimeJul 27th 2010

When I look at the old version in the edit history, I have four broken LaTeX formulas; the warning message asserts "Unknown control sequence '\underrightarrow'"
32.

I had replaced the \varinjlim commands in the old version with \underset{i}{\lim_{\longrightarrow}}, since I had seen the \underrightarrow problem before. Unfortunately, the bug (whatever it is) does not seem to manifest consistently. Now I get jbl's error message when I look at the edit history (with the same browser as before).