The retina and non-retina ipads are running iOS 8.3 (12F69) and 8.4.1 (12H321) respectively.

This is caused by the lazy loaded images on mobile.. T124770: Feature flagged Lazily load images

It seems that the style attribute isn't moved into the child element, but instead remains on the lazy-image-placeholder element, where they basically don't apply (because you can't assign width and height to an inline span element and they are not inherited by child img elements).

<span class="lazy-image-placeholder loaded" style="width: 50.819ex;height: 11.509ex;" data-src="https://wikimedia.org/api/rest_v1/media/math/render/svg/e1ba709c651d23276fe5da94800f5a931101d2b5" data-alt="\begin{array}{rl} \min\limits_{d} &amp; f(x_k) + \nabla f(x_k)^Td + \tfrac{1}{2} d^T \nabla_{xx}^2 \mathcal{L}(x_k,\lambda_k,\sigma_k) d \\
\mathrm{s.t.} &amp; b(x_k) + \nabla b(x_k)^Td \ge 0 \\
  &amp; c(x_k) + \nabla c(x_k)^T d = 0. \end{array}" data-class="mwe-math-fallback-image-inline"><img class="mwe-math-fallback-image-inline" src="https://wikimedia.org/api/rest_v1/media/math/render/svg/e1ba709c651d23276fe5da94800f5a931101d2b5" alt="\begin{array}{rl} \min\limits_{d} &amp; f(x_k) + \nabla f(x_k)^Td + \tfrac{1}{2} d^T \nabla_{xx}^2 \mathcal{L}(x_k,\lambda_k,\sigma_k) d \\
\mathrm{s.t.} &amp; b(x_k) + \nabla b(x_k)^Td \ge 0 \\
  &amp; c(x_k) + \nabla c(x_k)^T d = 0. \end{array}"></span>

Also partially caused by the images using style attributes to set width/height, instead of using width/height attributes in pixels, which might relate to T146298: Certain pages have reflows due to lazy loaded images (possibly due to downscaling via max-width: 100%).

@ovasileva - to add replication steps

@Nicholas.tsg - added steps to reproduce in task description

Screencaps from testing on iPad Air 2 (iOS 10.0.1):

Algebraically_closed_field Firefox

Algebraically_closed_field Safari

Sequential_quadratic_programming Firefox

Sequential_quadratic_programming Safari

While the maths are rendering much larger again than they were in the original screencap in the description, please let me know if I should check more iPad models (and/or browser apps) before confirming this as fixed.

If you're referring to a fix that has not been released to the public then please ignore this message.

I'm still seeing small images (I checked this isn't client-side caching by accessing pages I've never accessed before.) This is on both of the following systems:

iOS 8.3 (12F69) on an Apple iPad Air ME993LL/A
iOS 8.4.1 (12H321) on an older pre-retina ipad MC769LL/A

Algebraically_closed_field Safari

@Nicholas.tsg - Based on the comment above it seems the bug might be showing in different versions of iOS. Could you check some older versions? iOS 8.3 and perhaps on Android as well? Thank you!

I do not have access to an iOS 8 device anymore. Firefox no longer supports iOS 7 which I do have. Screencaps from testing on iPad Mini (iOS 7.1.2):

Algebraically_closed_field Safari

Algebraically_closed_field Chrome

Algebraically_closed_field Opera Mini

This task is probably also related even though it's not a duplicate. T143489

@TheDJ do you have an option to figure out if that's still a problem and do you have an idea to whom this task could be assigned?

@ovasileva can you schedule this task for being addressed? I'm happy to help, with math specific questions, but I'm not familar with the mobile extension.

@Physikerwelt - it's currently in our sprint + 1 column, meaning we will be able to get to it in our next sprint, or the one following that (within the next month - 1.5 months)

@ovasileva oh sorry I did not know that. I just saw that that the sprint 87 label was removed in Dec.

Whenever a full ETC is available, can someone post it here? Just need to be able keep some of our customers informed. (Wikimedia OTRS)

The change has not yet been deployed but will be on English Wikipedia on Thursday.