I'm writing a game engine and using opengl/qt. I'm trying to use webkit as a 2D UI overlay for menus etc. I found some examples but none of them work for some reason. For example adding two widgets QGLWidget and a QWebView doesn't seem to work.
How should I do this without losing much performance?
Related
My current UI doesn't use QT widgets paint and it's mostly done with QSS markup. The current styles are defined in pixels and I could not find anything in the QT documentation that allows the styles to be defined as relative unit as opposed to pixels. I would like to know if it's possible to use something relative then tie that unit to a ROOT font-size. In case I need to adjust for a particular screen size like a battery charger IOT device that has a small screen, while my main UI was designed for very large screens (27 inch monitor). So instead of me styling a button just for the IOT device, I shrink the base unit (technique similar to web responsive like REM) then everything like fonts, padding, buttons shrink proportionally. Is that something that QT supports?
To my knowledge vanilla Qt Widgets don't support free scaling at all. Recently high-DPI display support was added, but it only scales in fixed ratios (e.g. 1:2). There are other options to get scaling in Qt applications:
Use QML (didn't try this yet myself, but it's developed with adaption to different displays in mind)
Use QGraphicsScene with QGraphicsWidget
Use QGraphicsScene with custom graphics objects that implement the behavior you need
Roll your own widgets
I'd love to stand corrected on this answer - maybe in a future Qt version...?
I'm trying to make a wheel, which will be sliced into different coloured sections (kinda like a pie chart) but they may have extra elements on top of them, e.g., a stitched border across the borders at the side and an icon at the edge of each, etc.
So my question is do I just use css3 (examples: http://www.css3shapes.com/)
and then use something like http://css3pie.com/ for Internet explorer 6-9?
Or am I just better off doing it in Photoshop and slapping it on a div as a background image? And use absolute positioning for extra elements?
Which is the best practice?
I would (and do) use PHP's GD extension to create a pie chart based on input data, save the image generated by that, and optionally generate an imagemap so parts of the chart can be highlighted, have tooltips, etc.
I really wouldn't use CSS for drawing shapes. It's hacky at best, and can get really messy. Quite a few of the examples in the cssshapes site you linked to don't even work in Firefox 3.6 - which is still quite heavily used - let alone other older browsers, and in IE you'd really struggle; css3pie won't give you the kind of fine-grained control you'd need for a pie chart.
If you want to draw stuff on the browser using on the client side, there are a number of ways to do so without resorting to hacks like CSS shapes.
Both SVG and Canvas are relatively new technologies for placing and manipulating graphics on the browser. SVG draws vector graphics and Canvas is for pixel-based graphics.
Both of them are supported on a wide variety of browsers. The obvious exception being IE8 and earlier. However, IE does support an alternative graphics format called VML, and there are javascript tools for IE which allow it to convert both SVG and Canvas into VML, so you can use either of them and they will work in IE using these tools, which makes both SVG and Canvas effectively cross-browser compatible.
But if you're drawing pie charts, or similar sorts of things, my personal recommendation would be to use the Raphael library. This is a javascript library which works in all browsers, and can be used to draw and animate pretty much anything, including graphs and charts (it even has a separate graphing add-on library. See the demos on those two sites to see what it's capable of; I think you'll be impressed. (I certainly was)
Hope that helps.
After much digging I've realised there is no Qt-way to use 3 middle buttons in landscape mode in S50v5, however I don't like to waste this precious space. I've tried to place my widgets there but menu bar is on top and widgets aren't visible.
Is there any way to utilise this space without using native Symbian APIs?
Controlling the stock softkeys in Qt is a pain. The strategy I have used is to make the QMainWindow full-screen with showFullScreen() which allows you to use the softkey space yourself. You will have to make sure any widgets you create are ultimately parents of the main window, and be mindful of this bug when creating pop-ups.
I'm building an application, which is cross-platform (Linux, Win, OS X), and especially does graphics stuff via Cairo(mm), supported by a little bit of GTK+ scaffolding (i.e. DrawingArea). I now need some UI elements, which require a certain custom look. Ideally they should also be displayed as transparent overlay (i.e. different opacities for different parts of the UI).
As GTK+ is hard to customize (e.g. I want an edit-field with the suggestion-list above it), and me basically exactly knowing how I'd achieve this with HTML/CSS, the question popped up: Why not just let libwebkit handle the UI stuff?
I don't have much experience with WebKit, so what I need to know is:
Does my above reasoning make sense to anybody else?
CSS has opacity, can I interface WebKit in such way that it renders onto an RGBA-offscreen surface, with alpha-values inherited from the layout-processing of the CSS styles?
Even if 2. would not work, has anybody used libwebkit on an offscreen surface, which AFAIK requires redirection of keyboard and mouse interactions, probably via GTK+s provisions?
Is it possible to render web content over a clear background using WebKit?
is probably as good as it gets.
I am trying to use a vertical scrollbar in my Qt project.
The issue is when use the Cleanlooks widget style for the vertical scrollbar,it looks ok on the Linux Ubuntu machine,where Qt-4.3.3 is running.
But,when i run the same project on a Qtopia-4.3.3 Linux(ARM) machine,the scrolldown and scrollup images look totally jagged. They,look very bad. I tried using stylesheets,but without luck. Any suggestions regarding using stylesheets are most welcome.
Is there any way,to overcome this problem?
In my relatively small experience with styles on embedded platforms, some of the styles are designed so they use an image for certain things, and scale it as necessary. This produces very jagged graphics if the source is small and the target it large. One possibility is to inherit your own QStyle from the Cleanlooks style, and override the drawing of the arrow images with your own drawing code. It shouldn't be too hard to draw a proper arrow in code, rather than using an image.