This has been bugging me for years but because it doesn't effect performance I've never got around to asking. The problem below occurs for me in V8, 9, 10 with OS X 10.8, 10.9 and 10.10. I've tested the same code on Windows over the years and never observed any problems.
Here is a typical example in which a popups display different rendering.
What you will notice is that the rendering of say the first column of popups is consistent with the OS 10.10 however across some of the rows you will notice that the rendering harks back to, from memory, OS 10.6 and below.
As I said above it does not effect performance but it seems like a bug of sorts. Can anyone else 1) reproduce it (by creating say a grid of popups); and 2) know what the cause is and therefore have a possible fix?
Edit
Rendering the code in the comment by @SquareOne with V10.0.2 gives a grid with all popups the same but unfortunately they all render in the old 10.6 format rather than the Yosemite format.
The same with V9 on Yosemite:
Note that in all cases above Mathematica is rendering differently to the defaults for this operating system, just as it did with Mavericks. Here is what the OS popups look like in Yosemite:
Whereas the rendering done by Mathematica harks back to OS X from many years ago.
Edit #2
I've traced this to my default stylesheet. Specifically to the PopupMenu style with an option for FontSize
.
Cell[StyleData["PopupMenu"],
some options,
FontSize->11]
With this setting you get a mix of two different renderings (Yosemite and old OS X), and which popups are rendered differently is random.
Setting a font size for BaseStyle
in the PopupMenuBox
options at the cell level will switch off the problem.
Having said the above the inclusion of a font size in the PopupMenu
style was an error because this was intended to control to menu font size and therefore should be in PopupMenuLabel
style. Nevertheless the problem now seems to be that Mathematica has shown it is capable of rendering the correct Yosemite renderings (even if it is randomly under strange conditions) but by default is not doing that.
Edit 3
It seems no one here had any ideas about the origin of this problem so I am cross posting at Wolfram Community and hopefully someone from Wolfram will respond.
Comments
Post a Comment