I can no longer render LaTeX

  • Thread starter Thread starter renormalize
  • Start date Start date
  • Tags Tags
    Latex
Click For Summary
Users are experiencing issues with LaTeX rendering in posts, particularly after a recent software update. Attempts to refresh or use the preview function often result in LaTeX code being converted to an uneditable ASCII format, complicating recovery efforts. Workarounds include posting, refreshing, and editing, but these are seen as inadequate and frustrating. The editor's performance has also been noted as laggy, further hindering the user experience. Ongoing efforts are being made to resolve these issues, with hopes for improvements in future updates.
  • #31
berkeman said:
Hmm, something seems extra-broken today; testing:

-----------------------------------------------------
Example of in-line LaTeX: ##v(t) = L \frac{di(t)}{dt}##

Example of stand-alone LaTeX:
$$v(t) = L \frac{di(t)}{dt}$$
-----------------------------------------------------

using [ itex ] and [ tex ]:
-----------------------------------------------------
Example of in-line LaTeX: v(t) = L \frac{di(t)}{dt}

Example of stand-alone LaTeX:
v(t) = L \frac{di(t)}{dt}
-----------------------------------------------------
Okay, that's really weird. When I first posted it the double-delimiter things did not work and the [ tex ] tags did. Now it's the opposite when I re-check my post. I'm guessing it's being worked on in the background... :wink:
 
Physics news on Phys.org
  • #32
fresh_42 said:
My current workaround is "post"+"refresh"+"edit" (repeat as long as necessary). I added a comment "Do not answer. This post is still under construction." at the beginning of longer posts and delete this line when I'm done.
It goes without saying that we shouldn't have to do this. This has been two months now.
 
  • #33
PeroK said:
It goes without saying that we shouldn't have to do this. This has been two months now.
I have a couple of freelance Xenforo devs working on this and struggling with the new editor. Xenforo 2.4 (out in a few months) will switch to yet another editor. Maybe that is a good thing. @PeroK I would love your support as a programmer. Please DM me and it would be great for you to have a look. Thanks!!

Also I just got word that MathJax 4 is coming out soon. I'm sure that will change things.
 
Last edited:
  • #34
Another LaTeX issue:
Since last September I have been exchanging PF direct messages with @fluidistic regarding a problem he is interested in. I just went back to review them and although the messages clearly include raw LaTeX code, none of them actually render LaTeX equations anymore, no matter how often I refresh my browser page. That behavior is distinctly different from posts in PF threads which properly display LaTeX after refreshing the post (as long as you never use Preview prior to posting). Frustrating.
 
  • #35
renormalize said:
Another LaTeX issue:
Update: @Greg Bernhardt promptly fixed the issue. All my messages now properly render their LaTeX content.
 
  • Like
  • Informative
Likes Greg Bernhardt and berkeman
  • #36
PeroK said:
It goes without saying that we shouldn't have to do this. This has been two months now.
One other possible workaround is to keep a dummy insight around. Latex preview in Wordpress works just fine. When done, cut and paste to your reply.
 
  • #37
PAllen said:
One other possible workaround is to keep a dummy insight around. Latex preview in Wordpress works just fine. When done, cut and paste to your reply.
Yeah, but it has another big disadvantage: Do not do this with an existing WP text!

I was tempted to repair a broken link today. No big deal within WP. However, whenever you do this (any update) then it automatically destroys the corresponding OP of the PF thread! Hence, I had to decide: keep that broken link, or figure out where the corresponding PF thread is, which by itself is difficult enough, then enter the edit mode, enable the BB mode, copy the entire text to somewhere, then make your update, go back to that thread, enter edit mode again and delete the nonsense that has been created by your update, and insert the saved text again.

Well, I did not repair the broken link.
 
  • #38
fresh_42 said:
Yeah, but it has another big disadvantage: Do not do this with an existing WP text!

I was tempted to repair a broken link today. No big deal within WP. However, whenever you do this (any update) then it automatically destroys the corresponding OP of the PF thread! Hence, I had to decide: keep that broken link, or figure out where the corresponding PF thread is, which by itself is difficult enough, then enter the edit mode, enable the BB mode, copy the entire text to somewhere, then make your update, go back to that thread, enter edit mode again and delete the nonsense that has been created by your update, and insert the saved text again.

Well, I did not repair the broken link.
Can you clarify? I see you would never want to use a posted insight; that's why I suggested a dummy (draft) insight. Once the latex issue is solved, get rid of the dummy draft. Do you see any problems with that?
 
  • #39
PAllen said:
Can you clarify? I see you would never want to use a posted insight; that's why I suggested a dummy (draft) insight. Once the latex issue is solved, get rid of the dummy draft. Do you see any problems with that?
No. I just wanted to point out that the update function on WP has a very negative impact on the corresponding thread on PF. A dummy thread is ok. You just shouldn't use an existing one.
 
  • #40
A latex preview update is shipping soon. Need to walk the dogs first :)
 
  • #41
Dang this is weird. It's working great on CivicsWatch, but here, it auto renders in the editor outside of preview.
 
  • #42
Greg Bernhardt said:
Dang this is weird. It's working great on CivicsWatch, but here, it auto renders in the editor outside of preview.
Is it already live, since as of now, it does not work on CW for me?
 
  • #43
fresh_42 said:
Is it already live, since as of now, it does not work on CW for me?
Yeah it's live on CW and works for me. Clear your cache.
 
  • #44
Greg Bernhardt said:
Yeah it's live on CW and works for me. Clear your cache.
Well, yes and no.

Code:
##
\begin{pmatrix}a&b\end{pmatrix}
##

[itex]
\begin{pmatrix}c&d\end{pmatrix}
[/itex]

$$
\begin{pmatrix}e&f\end{pmatrix}
$$

[tex]
\begin{pmatrix}g&h\end{pmatrix}
[/tex]
worked. But using the hashtags or double dollars in the same line did not work, and an error made once remains in the cache without showing the wrong code ever again in the editor. If I type

Code:
##  \begin{pmatrix}a&b\end{pmatrix} ...
(ab)
in one line, then it immediately turns it into a preview without a request and the possibility to correct it.

Same on both platforms.

Correction: PF still has the <br> repair stuff.
 
  • #45
Try now, I've been tinkering this entire time and maybe I got it. Clear cache and reload.
 
  • #46
##\begin{pmatrix}a&amp;b\end{pmatrix}## is a vector.

\begin{pmatrix}c&amp;amp;d\end{pmatrix} is a vector.

$$\begin{pmatrix}e&amp;f\end{pmatrix}$$ is a vector.

\begin{pmatrix}g&amp;amp;h\end{pmatrix} is a vector.

CW: No preview anymore at all.
PF: Inserting "amp;" stuff at the ampersand.

Hint @all: The {align*} statement has now to be put into double dollars. The previous version recognized them automatically as math.
 
Last edited:
  • #47
fresh_42 said:
CW: No preview anymore at all.
I removed that as all dev is on here now
fresh_42 said:
PF: Inserting "amp;" stuff at the ampersand.
can you elaborate on this in simple terms
fresh_42 said:
Hint @all: The {align*} statement has now to be put into double dollars. The previous version recognized them automatically as math.
are you suggesting some config has changed?
 
  • #48
Greg Bernhardt said:
can you elaborate on this in simple terms
Works now.

Test:
Code:
##\begin{pmatrix}1&0\\0&1\end{pmatrix}##
[itex]\begin{pmatrix}1&0\\0&1\end{pmatrix}[/itex]
$$\begin{pmatrix}1&0\\0&1\end{pmatrix}$$
[tex]\begin{pmatrix}1&0\\0&1\end{pmatrix}[/tex]

##\begin{pmatrix}1&0\\0&1\end{pmatrix}##
\begin{pmatrix}1&amp;0\\0&amp;1\end{pmatrix}
$$\begin{pmatrix}1&0\\0&1\end{pmatrix}$$
\begin{pmatrix}1&amp;0\\0&amp;1\end{pmatrix}
Only copying this code directly from post #44 on the screen by Ctrl+C inserted some <br> and amp; and the preview ignores the carriage returns within the following code tags.

Code:
##<br>\begin{pmatrix}a&amp;b\end{pmatrix}<br>##<br><br>[itex]<br>\begin{pmatrix}c&amp;d\end{pmatrix}<br>[/itex]<br><br>$$<br>\begin{pmatrix}e&amp;f\end{pmatrix}<br>$$<br><br>[tex]<br>\begin{pmatrix}g&amp;h\end{pmatrix}<br>[/tex]

Greg Bernhardt said:
are you suggesting some config has changed?
Not anymore. I had tested
Code:
\begin{align*}
1&=-1\\[16pt]
0&\neq 1
\end{align*}

\begin{align*}
1&=-1\\[16pt]
0&\neq 1
\end{align*}
and it wasn't recognized automatically as math code, but now that I tested it again, it does (as in the previous version).
 
  • #49
Well done!
 
  • Love
Likes Greg Bernhardt
  • #50
fresh_42 said:
Works now.

Test:
Code:
##\begin{pmatrix}1&0\\0&1\end{pmatrix}##
[itex]\begin{pmatrix}1&0\\0&1\end{pmatrix}[/itex]
$$\begin{pmatrix}1&0\\0&1\end{pmatrix}$$
[tex]\begin{pmatrix}1&0\\0&1\end{pmatrix}[/tex]

##\begin{pmatrix}1&0\\0&1\end{pmatrix}##
\begin{pmatrix}1&amp;0\\0&amp;1\end{pmatrix}
$$\begin{pmatrix}1&0\\0&1\end{pmatrix}$$
\begin{pmatrix}1&amp;0\\0&amp;1\end{pmatrix}
This was directly after I posted it correct, but now it looks broken. Strange.

Second try:

##\begin{pmatrix}1&0\\0&1\end{pmatrix}##
\begin{pmatrix}1&amp;0\\0&amp;1\end{pmatrix}
$$\begin{pmatrix}1&0\\0&1\end{pmatrix}$$
\begin{pmatrix}1&amp;0\\0&amp;1\end{pmatrix}

Looks fine in the preview, but my screen shows

1754760198876.webp
 
  • #51
Testing inline ##\Omega=ab\mu^{\nu}\text{ test}## code. Testing display$$\Omega=ab\mu^{\nu}\text{ test}$$code.
Both inline and display tags now work reliably for me in preview. Eureka!
 
  • Love
  • Like
Likes FactChecker and Greg Bernhardt
  • #53
Let's see.

Code:
\begin{align*}
\begin{pmatrix}1&0\\0&1\end{pmatrix}&=\mathbb{I}_2
\end{align*}
\begin{align*}
\begin{pmatrix}1&0\\0&1\end{pmatrix}&=\mathbb{I}_2
\end{align*}

results in ...

a) no preview at all
and then corrected to
Code:
$$
\begin{align*}
\begin{pmatrix}1&0\\0&1\end{pmatrix}&=\mathbb{I}_2
\end{align*}
$$
$$
\begin{align*}
\begin{pmatrix}1&0\\0&1\end{pmatrix}&=\mathbb{I}_2
\end{align*}
$$
resulted in ...

b) a correct preview, and (sic!) it also previewed the first version without math tags correctly this time. But it did not on the first attempt. (I have no idea why.)

So now we will see what happens after hitting the post reply button.

Edit 1:
c) Page refresh is necessary

Edit 2:
d) Still, both versions are correct.

See what will happen after a while and/or leave the page.

Edit 3:
e) No additional amp; notes this time. I don't know what happened the first time.

I try to do it with the same code again:
Code:
##\begin{pmatrix}a&b\end{pmatrix}##
[itex]\begin{pmatrix}c&d\end{pmatrix}[/itex]
$$\begin{pmatrix}e&f\end{pmatrix}$$
[tex]\begin{pmatrix}g&h\end{pmatrix}[/tex]
##\begin{pmatrix}a&b\end{pmatrix}##
\begin{pmatrix}c&amp;d\end{pmatrix}
$$\begin{pmatrix}e&f\end{pmatrix}$$
\begin{pmatrix}g&amp;h\end{pmatrix}

f) Ctrl+C still copies the additional commands <br> and amp; into the buffer.
g) The preview is correct.

General observation from other threads: the quoted texts do not resolve TeX commands.

I close the edit session now.

Edit 4:
h) The post has been displayed correctly (after one refresh).
i) After the second refresh, I have again the amp; tags on the screen.
j) This occurs only in the lines that used [ tex ] and [ itex ] commands, not on the lines with hastags or dollars.
 
Last edited:
  • #54
The motional emf across an oriented segment ##d\mathbf l## is $$d(\text{emf})=\mathbf E\cdot d\mathbf l= (\mathbf v\times \mathbf B)\cdot d\mathbf l.$$Now $$\begin{align}& \mathbf E=(\mathbf v\times \mathbf B)=(\boldsymbol{\omega}\times \mathbf{r})\times\mathbf B \nonumber \\& = \mathbf B\times(\mathbf {r}\times \boldsymbol{\omega}) \nonumber \\& =\cancelto{0}{\mathbf r(\mathbf B\cdot \boldsymbol{\omega})}-\boldsymbol{\omega}(\mathbf B\cdot \mathbf r) =-\omega B r\cos\!\theta ~\mathbf {\hat y}.\end{align}$$
Code:
The motional emf across an oriented segment ##d\mathbf l## is $$
d(\text{emf})=\mathbf E\cdot d\mathbf l= (\mathbf v\times \mathbf B)\cdot d\mathbf l.$$ Now $$
\begin{align}& \mathbf E=(\mathbf v\times \mathbf B)
=(\boldsymbol{\omega}\times \mathbf{r})\times\mathbf B \nonumber \\
 = \mathbf B\times(\mathbf {r}\times \boldsymbol{\omega}) \nonumber \\
& =\cancelto{0}{\mathbf r(\mathbf B\cdot \boldsymbol{\omega})}-\boldsymbol{\omega}
(\mathbf B\cdot \mathbf r)=-\omega B r\cos\!\theta ~\mathbf {\hat y}.\end{align}$$

I had to change all &amp; to just &.
Looking good! Thanks Greg.
 
  • #55
Greg Bernhardt said:
Try now, I've been tinkering this entire time and maybe I got it. Clear cache and reload.
Thanks, Greg

Edit:
And it also seems to render much faster than before.
 
Last edited:
  • #56
The preview seems to function fine (thanks again), but LaTeX inside quotations is now presented as text even after submission of the reply. For example, the quote in this post doesn't show the equations as LaTeX on my browser.
 
  • #57
JimWhoKnew said:
The preview seems to function fine (thanks again), but LaTeX inside quotations is now presented as text even after submission of the reply. For example, the quote in this post doesn't show the equations as LaTeX on my browser.
I see now that I have a similar problem with the "Homework Statement", "Relevant Equations" and "TL;DR Summary" sections.
 
  • #58
I assume that it doesn't render in any tags.

Test 1:

##\mathbb{Z}##
##\mathbb{Z}##

##\mathbb{Z}##


Edit: Strange, it had worked in the preview.
 
  • #59
latex should be fixed in message fields aka in homework help, and in quote blocks now
 
  • Like
Likes renormalize and JimWhoKnew
  • #60
Greg Bernhardt said:
hmmm this is a tough one. I'm at a bit of a loss. Nothing I'm trying is close to working.
Can you revert to a previous version that is known not to have this problem?
 

Similar threads

  • · Replies 11 ·
Replies
11
Views
460
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 16 ·
Replies
16
Views
3K
  • · Replies 6 ·
Replies
6
Views
4K
  • · Replies 4 ·
Replies
4
Views
4K
Replies
15
Views
9K
  • · Replies 12 ·
Replies
12
Views
2K
  • · Replies 32 ·
2
Replies
32
Views
8K
  • · Replies 147 ·
5
Replies
147
Views
18K
  • · Replies 3 ·
Replies
3
Views
2K