This is possibly related to T377657, but Blockquote is a more common case. The same bug occurs if either {{blockquote}} or <blockquote> is used.

Original example

https://en.wikipedia.org/w/index.php?title=Dijkstra%27s_algorithm&oldid=1263378061#Dynamic%20programming%20perspective

Now changed to remove <math> tags.

The parser is inserting <p> inside the <blockquote> and the <math> element isn't ending up inside those. It works as expected if you insert the <p> yourself.

Besides <blockquote> elements, floating divs (e.g. images) in the same paragraph also trigger this problem: <math>A</math> implies <math>B</math> ocassionally. <div style="display:float">

Another example is mentioned in T389021 affecting the page

https://en.wikipedia.org/w/index.php?title=Eadie%E2%80%93Hofstee_diagram&oldid=1277717249

Even if the <math> tag is outside the <blockquote> we still get the same behaviour. The example

For the group <math>G=C_m</math> given <blockquote>m=5</blockquote>

Renders as

This example is simplified from https://en.wikipedia.org/w/index.php?title=Group_cohomology&oldid=1293195916#Basic_examples
where an editor has used block quotes to format display equations.

This is quite a big problem, initial check indicates it might affect 474 main space articles on en-wiki.

I don't have an idea why this happens, or what could be done from within the Math extension to improve the situation.

I though as much, alas it is enough to block T271001 as rendering on many pages is substandard. Worse example I've seen is from https://en.wikipedia.org/wiki/Optical_metric

Here a parenthesised expression is broken over multiple lines.

@SalixAlba I don't see anything on Optical metric, which version of Chrome are you using? It seems it is a browser bug. If so, shouldn't it be reported upstream?

Its a bit hacky but if we see

</p>
<math class="mwe-math-element mwe-math-element-inline">
...
</math>
<p>

Then it might be safe to conclude that the </p> and <p> are spurious. You need the mwe-math-element-inline class to know its an inline formula.

Hello,

I added also T375238 as a parent task. I think this is also the problem seen in the picture

The code is:

Stammfunktionen und der Hauptsatz der Differential- und Integralrechnung

{{Hauptartikel|Fundamentalsatz der Analysis}}Die Integration ist eine nicht-eindeutige Umkehrung der Differentiation. Um dies zu präzisieren, wird der Begriff der [[Stammfunktion]] benötigt: Ist <math>f</math> eine Funktion, so heißt eine Funktion <math>F</math> eine ''Stammfunktion'' von <math>f</math>, wenn die [[Differentialrechnung|Ableitung]] von <math>F</math> gleich <math>f</math> ist:
: <math>F' = f</math>

In T382267#11002727, @SalixAlba wrote:

I'm pretty sure its not my browser set up. You need to get down to the Electrodynamics in media residing in curved space-times section.

The wikitext

The fundamental quantity of electrodynamics is the potential 1-form <math>\boldsymbol{A},</math> from which the field strength tensor is the 2-form <math display="inline">\boldsymbol{F} = d \boldsymbol{A}.</math> From the nilpotency of the exterior derivative one immediately has the homogeneous Maxwell equations <blockquote><math>d \boldsymbol{F} = 0,</math></blockquote>

Gets converted to

<p>
The fundamental quantity of electrodynamics is the potential 1-form </p>
<math class="mwe-math-element mwe-math-element-inline" xmlns="http://www.w3.org/1998/Math/MathML"><mrow data-mjx-texclass="ORD"><mstyle displaystyle="true" scriptlevel="0"><mrow data-mjx-texclass="ORD"><mi>𝑨</mi></mrow><mo>,</mo></mstyle></mrow></math>
<p> from which the field strength tensor is the 2-form </p>

Note how closing </p> elements have been added. If you remove the <blockquote> element. The wikitext is converted to

<p>The fundamental quantity of electrodynamics is the potential 1-form <math class="mwe-math-element mwe-math-element-inline" xmlns="http://www.w3.org/1998/Math/MathML"><mrow data-mjx-texclass="ORD"><mstyle displaystyle="true" scriptlevel="0"><mrow data-mjx-texclass="ORD"><mi>𝑨</mi></mrow><mo>,</mo></mstyle></mrow></math> from which the field strength tensor is the 2-form <math class="mwe-math-element mwe-math-element-inline" xmlns="http://www.w3.org/1998/Math/MathML"><mrow data-mjx-texclass="ORD"><mstyle displaystyle="false" scriptlevel="0"><mrow data-mjx-texclass="ORD"><mi>𝑭</mi></mrow><mo stretchy="false">=</mo><mi>d</mi><mrow data-mjx-texclass="ORD"><mi>𝑨</mi></mrow><mo stretchy="false">.</mo></mstyle></mrow></math> From the nilpotency of the exterior derivative one immediately has the homogeneous Maxwell equations 
</p>

With no closing </p> before the <math> tag.

I think this happens before it get to the maths parser which is why its hard to fix.

@Christian1985 This explains the situation pretty well I think. From within the Math extension, I don't think anything can be done about it. I was thinking maybe the problem vanishes once we switch to parsoid parsing and created https://de.wikipedia.beta.wmcloud.org/wiki/T382267 where I can not reproduce the problem. See https://www.mediawiki.org/wiki/Parsoid/Parser_Unification maybe @cscott has more insights

This does not seem to be related to MathJax

In T382267#11636680, @Christian1985 wrote:

Hello,

I added also T375238 as a parent task. I think this is also the problem seen in the picture

The picuture is gereated from https://de.wikipedia.org/w/index.php?title=Integralrechnung&oldid=259297415 And I used MathJax for rendering.

The code is:

Stammfunktionen und der Hauptsatz der Differential- und Integralrechnung

{{Hauptartikel|Fundamentalsatz der Analysis}}Die Integration ist eine nicht-eindeutige Umkehrung der Differentiation. Um dies zu präzisieren, wird der Begriff der [[Stammfunktion]] benötigt: Ist <math>f</math> eine Funktion, so heißt eine Funktion <math>F</math> eine ''Stammfunktion'' von <math>f</math>, wenn die [[Differentialrechnung|Ableitung]] von <math>F</math> gleich <math>f</math> ist:
: <math>F' = f</math>

@Physikerwelt , what do you mean by "This does not seem to be related to MathJax"? The rendering in the picture in the quoted message is MathJax and the output is ugly as hell. With standard svg the page looks much better.

I think it works by accident for the images generated by the SVG rendering mode. https://www.mediawiki.org/wiki/Help:Formatting/de links to the https://phabricator.wikimedia.org/T309778 . I think that's the problem to be solved here.

Change #1262984 had a related patch set uploaded (by Physikerwelt; author: Physikerwelt):

[mediawiki/extensions/Math@master] Hide native MathML from RemexHTML

https://gerrit.wikimedia.org/r/1262984

Change #1262984 merged by jenkins-bot:

[mediawiki/extensions/Math@master] Hide native MathML from RemexHTML

https://gerrit.wikimedia.org/r/1262984

Hello @Physikerwelt ,

should this bug be fixed for now? Using client side MathJax. The layout of the page https://de.wikipedia.org/w/index.php?title=K%C3%A4hler-Einstein-Mannigfaltigkeit&oldid=269053772 is still broken.

Hello @Physikerwelt ,

what is Pardoid? What do I have to do to deactivate it? Do I have to open a new ticket?

My settings

In T382267#12228497, @Christian1985 wrote:

what is Parsoid? What do I have to do to deactivate it?

Hi, see https://www.mediawiki.org/wiki/Help:Extension:ParserMigration for options - either via your user preferences/settings, or by passing the parameter useparsoid=0 in the URL/web address.