I am trying to understand how compositors work on X (well basically because neither xcompmgr nor cairo-compmgr can draw shadow properly for my awesome wm~~~)
I have read part of the source code both xcompmgr and cairo-compmgr but I still don't really understand how they do that.
I want to know how they know where the shadow should be (well, arround the window for sure, but the shadow might be under other window and don't need to be drawn.), as well as where (on which layer/window) do they draw the shadow. Probably also how all those X extension are used (and what's for) and how cairo-compmgr use cairo to deal with low level X stuff.
It's a little hard for me to learn these from the source code because a lot of stuff (especially X extensions) is poor documented. It will also be helpful just to point out where I should look at.
The simpler you code it, the best it will work.
Get a list of the visible windows
Sort them by inverse z-order (from the bottom-most to the top-most)
Draw the shadow and then the window itself for each window
You need no black magic.
If you are wondering how it works, it's straightforward: you have to use the 'composite' X extension. It enables the overlay window, which is the only visible window on the screen once it is enabled, then you have to draw all the windows on it as you will be provided with a Pixmap for each window.
EDIT:
If you are seeking for documentation, you can use the linux manual (the man command), and the header files, they're the main (also best and perhaps the only real) source of documentation, as all the other sources/websites rely on them afaik.
Related
I'm new to cocoa development and wondering what the right approach would be to the following:
I have a window with controls who's size is dynamic. In other words, I don't know the width of the window in advance. I also want to give the window a padding of 15px. In the CSS world I would simply pad a div and make it inline to accomplish this. I can't seem to figure out how to do this with NSWindow/NSView.
Question is too vague to be sure, but IFF I'm understanding you, you want a subview of the window to resize itself to always be 15pixels from the left and right sides of the enclosing window (for example).
here are the two constraints (as set up in Interface Builder, but you can do the same thing from code):
That's one view whose width is set to adjust while keeping the left and right padding to the window frame at 15 pixels (points actually, but don't worry about that).
As your view hierarchy gets more complicated you'll use different techniques for different pieces of the layout. Which is why matt was suggesting that you haven't been specific enough to actually answer a question. (What specific part of what you are you trying to do isn't working and what did you try that didn't work?). I've shown you the first step of the simplest outer layer above and it reflects a principle that carries down through your view hierarchy, but there are other cases - two items with the same width, three items with the same horizontal space between them, and so on - that require different techniques.
It sounds reading the auto layout documentation all the way through would be helpful for you. It talks about different types of relations (Equal is what I used below) and some on techniques. Unfortunately Apple's documentation situation is a bit of a mess at the moment which makes this more difficult than it should be (sigh).
What used to be the definitive auto layout guide is "unmaintained" and the place that documentation seems to be heading is incomplete and certainly less accessible. The most useful part for this context is the NSLayoutContraint documentation, though that links to the "Unmaintained" document noted above (sigh).
I trying to write simple gtk+ application with pygobject that using gtk.TextView. In stock demo code I see that font sets up with Pango, but I can't find corresponding method for set up antialiasing-mode. By default font looks little ugly for me.
Does anybody know how to set anti-aliasing in gtk+ app (with python or pure C, Pango or other way)?
Well, it's not easy. Many factors interact to get the final font glyphs on your screen.
Maybe one of the easiest ways to change antialiasing globally would be to reprogram fontconfig. This article describes how to disable anti-aliasing, but it indicates where the config item is located.
Then, on a per-application basis, the actual work of drawing the characters is in the hands of Pango, but Pango may be programmed to use Xft (fontconfig) or Cairo to do the actual drawing. In case of Cairo (which seems to gather popularity), the anti-aliasing is done there. There's an article on that here on stackoverflow. Basically there's a function in the Cairo library, called 'cairo_set_aliasing'
Edit: Have a look at the XFCE window manager. In ths Settings | Appearance tool, you can change the anti aliasing 'globally', from 'none', 'slight', 'more' to 'full', and select subpixel in color or gray, horizontal/vertical and details. The code is contained in the xfce4-appearance-settings module of XFCE.
The aliasing setup is in xfce4-settings-4.12.0.tar.bz2, in dialogs/appearance-settings/main.c. I'm not sure if it actually activated there. I've quite a lot of work at the moment, but, if necessary can lend you a hand later.
I have created a basic MATLAB UI (without using GUIDE). I basically have a bunch of panels for various things, (sliders, axes, text boxes, etc).
The one thing I would like to do though, it make it so that they scale properly, when I resize the figure. Right now, I painstakingly have to make a re-scale function for every button, panel, sub-panel, etc etc to make it rescale correctly.
Is there an easy way to simply automate the re-scaling here?
Thanks.
Use the GUI Layout Toolbox from the MATLAB File Exchange. I haven't personally used dynamic resizing functionality, but that's one benefit of using this package.
It functions much like using uicontrols, except you can't use the inspect tool on these objects.
EDIT: If you're looking only to do resizing when the figure itself is resized, set the Units property for all your uicontrols to normalized.
You could also use the builtin, but undocumented uigridcontainer and uiflowcontainer.
They have the benefit of e.g. allowing to set contraints, such that e.g. your pushbuttons don't get increased in size, when the full figure does. Check the link for some examples:
http://undocumentedmatlab.com/blog/matlab-layout-managers-uicontainer-and-relatives/
I have to develop an application "Behavior like an Tetris game".
I have never used "OpenGL" for the iPhone application developement.
Application is something like this
Red / green / blue square boxes drop from top
Red + Red + Red = Points & boxes disappears
same way user has to make combination & get points
Different levels are there.
There are three buttons Left, Right for movement & bottom for speedy fall
For this kind of application should I use open GL or NOT?
i.e. Is it possible to develop entire application with view & it's animation?
If yes then, will it be more complex as compare to open gl?
What is the advantage of using open GL?
(I know that it gives good 2d, 3d look )
(But here my question means - easy coding?)
(Or open gl is more complicated as compare to objective-c?)
(I am just asking because I am not aware of it)
Basically your options are:
Using OpenGL
Using Quartz
Using UIKit
OpenGL is a fairly complicated beast, but is by far the best way to squeeze performance out of the iPhone. Do you need it for a Tetris game, though? Almost certainly not.
Quartz is the toolkit used in Mac OS X and the iPhone to draw images and do image effects. Because I come from an OpenGL background in other languages, I find Quartz strange and frustrating. However, it is probably easier for someone who is new to both.
You can do everything here using UIKit, and it will definitely be much much easier than other options. The main disadvantage is that it's rather slow in comparison, but once again doing a Tetris-like game shouldn't matter at all.
Before you go with UIKit, though, I recommend just checking out something like Cocos 2D, which will give you the advantages of OpenGL without the headache of dealing with all of its inner workings.
From the tone of your question it looks like you're confusing what OpenGL is and isn't with regard to Objective-C.
OpenGL is a library written in the C programming language (to put it simplistically) that excels at rendering shapes (especially 3D shapes) for display on a screen. It doesn't replace Objective-C inside your program, it merely assists you in drawing the shapes. If you don't use OpenGL, you'll need to write some sort of drawing/rendering code in your NSView (or subclass) to render the blocks. By using OpenGL, you will be provided a lot of helpful C methods for drawing shapes, which otherwise you'll have to implement yourself. On top of that OpenGL has thousands of man hours worth of drawing optimizations that you can take advantage of if you use it rather than trying to implement shape rendering yourself.
Having said that, OpenGL isn't all sunshine and roses. It works like a state machine and has its own assumptions about the way it will be used (like any API). Just because you know C and Objective-C doesn't mean that using OpenGL will be trivial. If you've never written any OpenGL code, I suggest you look into a reference like the venerable Red Book.
The thing to keep in mind is that OpenGL is not a language until itself (ignoring the OpenGL shading language). Its merely a set of C functions to aid you in rendering graphics.
You may well want to ask as well on http://iphonegamedev.stackexchange.com/, the new Stack Overflow variant just for iPhone gaming.
To learn & understand what you need.
Please go through following link.
it includes all necessary links for all kind of resources that you needed.
http://maniacdev.com/2009/04/8-great-resources-for-learning-iphone-opengl-es/
Edit :
After reading your question properly ( actually my question - By r & d I found solution).
I think - you need to develop a 2d application.
Go for the following link. Best option for 2d animation.
http://code.google.com/p/cocos2d-iphone/
Don't forget to visit following link, if you needed sample codes.
http://monoclestudios.com/cocos2d_whitepaper.html
This isn't strictly a programming question, but I'm asking it here because it's certainly a software development question, if you take "software development" to include all aspects of creating a software system.
I am an independent iPhone developer. Except for translations, I handle all aspects of my apps myself—graphics included. I have to create icons, buttons, and UI elements of all sorts on a regular basis. I've learned a few tricks along these lines, and while they're certainly not works of art, I can effectively use gradients, shadows, border strokes, transparency, and textures to create minimalistic, attractive effects.
So far, I've used a vector drawing tool called VectorDesigner for all of my development, with occasional raster postprocessing by Pixelmator. It's worked mostly okay so far, but VectorDesigner has a host of issues:
It uses a package format for its files, which interferes with the use of Subversion.
It is very much a print tool, and I have to be very careful not to end up with objects on fractional pixel values which cause antialiasing.
While you can take the union or intersection of shapes, or add and subtract them, curves tend to deform with repeated boolean operations, sometimes quite dramatically.
And it offers very little control over strokes, to the point where I barely use them.
So I'm looking for a better tool for this specific purpose: shape-based drawing of simple icons, buttons, and UI elements on a Mac by someone without graphic design training. Good functions for exporting would be a plus—ideally it should be almost as easy to export a PNG to the place it goes in my project as it is to save (not save as) the file.
The perfect tool for me would be one that would allow you to define an object's shape by stacking up areas and masks defined by primitive shapes (which would remain separately editable), then define properties on those objects like transforms and strokes. I have no idea if something like this exists, though.
Adobe's tools generally strike me as very heavyweight, and are usually expensive, but I suppose they're a possibility. (Fireworks, with its emphasis on screen design, seems like it might be particularly suitable, but I don't know that much about it.) But what else is out there? If you're in a position like me, what do you use? What do you recommend?
Edited to add: Of course a graphic designer could get better results from an ancient copy of MacPaint than I could from Illustrator CS5. No tool can replace skill and taste, and many programmers have little of either. I'm aware of that. But I'm fortunate enough to have at least some taste—enough that my users compliment my apps' appearance in their reviews. I'm not hugely talented, but I do know my limitations, and I don't let myself produce anything ugly. Given my budget, that will have to do for now.
Try Opacity. A little rough on the edges, but one of the coolest and most unique features they have is export as source code (in Quartz, Cocoa, Cocoa Touch, or Canvas)
I'd suggest OmniGraffle
OmniGraffle is easy to use, can save as a PNG, can create binary non-package files (it's an option in the interface). You can also set the units to pixels to ensure exact alignment. (Canvas Size -> Ruler Units)
Finally, the Graffletopia website has some nice iPhone stencils for getting it right:
http://graffletopia.com/search/iphone
I think the only acceptable answer here should be "hire a designer".
But it sounds like Pixelmator/Inkscape are your best bets.
Though if you do find something better, that'd be really cool. Like a jQueryUI but for native.