Why does this drop-shadow appear only if overflow-x is available? - safari

This is only observed on Safari Version 15.6 (17613.3.9.1.5); Firefox seems to render as expected.
Observe the following via the Tailwind playground:
https://play.tailwindcss.com/CBCrf9Ujbt
The above has scrollable space on the x-axis. Now if we remove the long comment, observe the drop-shadow disappear:
https://play.tailwindcss.com/74dWNJOG0k
Why is this happening and how can we work around this such that the drop-shadow displays consistently?

Related

Selenium: precision of move_to_element ? It doesn't trigger mouse hover on small elements

I'm having this issue with move_to_element on small elements (dots in a graph) to trigger the mouse over that brings up a tooltip:
If I just move to an element that is ~1.5 pixels round, it fails 20-30% of the time.
I ended up writing a small hack script, spiraling around the element until I hit it and it works 100% of the time, but it's kinda slow.
It consistently seem to find the element about +2,+2 pixels away when not found at first.
To me that just says that the move_to_element function is just not precise enough.
From the doc it says it defaults to the center of the element.
When I look up the element in Chrome dev tools, the tooltip arrow displayed by Chrome to show where the element is located and its size, does not point to the dot itself but several pixels to the right.
So could this be a Chrome issue?
Has anyone experienced this kind of problem? Is there better ways to 'calibrate' the move_to_element functionality?

Bootstrap addons not working properly on different resolutions

It looks like there is a bug in Bootstrap. If you change resolution, for example with ctrl + (+/-) in your browser the addons are not properly adjusted. It looks like this:
Notice the wrong sizes of addons.
Is there a simple way to fix this?
No, there is no simple way (and quite possibly no way period) to fix problems related to the type of zooming that you're referring to (as opposed to the viewport-width-based zooming that you get when using the pinch gesture on touch-based devices).
Per Bootstrap's documentation:
Page zooming inevitably presents rendering artifacts in some components, both in Bootstrap and the rest of the web. [... these issues] often have no direct solution other than hacky workarounds.
Zooming in and out is not changing the resolution at all its zooming and is completely different from changing the view port's aspect ratio and dimensions.
Give the input area a fixed width like input {width:100px}

chrome, cy.js 2.1.0: nodes disappear with mouse motion

macos 10.7.5
chrome Version 32.0.1700.107
Version 32.0.1700.107
and the simple "single.html" helpfully provided by gcpdev:
live code examples: cytoscape.js initialization -- incomplete?
Any movement by the mouse in the canvas makes the nodes disappear. zooming and and out (with middle mouse roller button, for instance) makes them visible.
This problem is only on Chrome. Firefox and Safari work fine.
Any suggestions?
Thanks!
Paul Shannon
I can't reproduce this on Mac OS 10.9 with Chrome 32, which probably indicates this is a problem with Chrome rendering canvas on that version of Mac OS.
I notice that the linked demo is using v2.0.2. Have you tried an updated v2.0.x release or v2.1.0?
Try using pinch-zoom to set the zoom level for the webpage (not for the Cytoscape canvas element, which has its own separate zooming operated by scrolling with two fingers on the trackpad) to 100%. This fixed it for me, based on this question and this github issue, both about an almost identical issue in Firefox.

High resolution display give incorrect viewport when snapping (Surface Pro)

I just got to test IE10 on a Surface Pro with 1920*1080 display resolution where "make text and other items larger or smaller" has been set to Large.
On my website I have added the CSS+JS viewport fix in addition to the viewport meta tag, all asking for width: device-width (plus I added a "min-width: 320px;" to the #-ms-viewport definition to ensure it never gets smaller than that).
I added some javascript to display the value of window.screen.width and $(window).width to see what the browser ended up using for viewport in IE10, and to my surprise the screen size of a 1920*1080 resolution display was reported as 1280x720!
Now, I can live with with that (just like small phone screens report 320px width no matter their actual resolution, since it is a good size to make stuff human readable across devices for the same font size), but when the 'Metro IE10' is snapped to the side of the screen, the problem comes: IE10 tries to make a 320px rendering of the website, but it zooms in so the right side of it is hidden.
I tried Microsofts own test page: http://ie.microsoft.com/testdrive/Graphics/MakeItSnappy/
It does the same thing - on the Surface Pro the right side is hidden, and you need to drag left/right to see it, and you cannot even zoom out to view the full width!
But when trying the same thing on my laptop with a 'normal' 1366x768 display (rendered as 100%), the snapped IE10 display has the perfect size.
Ok, I guess this is a Microsoft Windows scaling bug - but my question is: Has anybody else experimented with changing the "make text and other items larger or smaller" to 125% or 150% and making websites adapt correctly?
Check out this fix from developer Matt Stow
http://mattstow.com/responsive-design-in-ie10-on-windows-phone-8.html

Does Webkit render 8bit PNGs correctly

I've come across a weird problem whilst trying to optimise the following image:
It should have a little shadow down each of the left and right hand sides. However, in webkit (tried mac safari and chrome) the right hand edge loses its little shadow entirely. I've used multiple utilities to create the image and even tried copying a flipped version of the left hand side to the right, all to no avail.
Have I generated the png incorrectly or does webkit have a problem? It looks as it should in Firefox, IE etc.
This may have been a bug with the WebKit rendering engine, but it appears to render correctly now.