Jump to content
  • Advertisement
Sign in to follow this  
Nanoha

OpenGL Implementing a 9-patch.

This topic is 877 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

I currently have an app which uses OpenGL. I have made it so I can use 9patch textures (that resize correctly to fit contents) BUT there is one tiny bit that I can't find a neat solution to. The 9-patch texture itself has  border that defines what parts can stretch and what parts are content and so you can use that to determine how much space is in there and scale accordingly. The problem is that the space is just space, it has no meaning at all. It's like saying a piece of string is 23 long - without saying what value that is it is meaningless (23mm, 23cm, 23m etc) and I have a similar issue with my 9patch.

 

I currently define all the values between 0-1. I know the texture dimensions so I could do something with that but different devices have different densities and so I need to load a different texture thus I cannot rely on that. My 'world' itself is define as -1 to +1 x/y with 0,0 being centre screen. I add a little bit to the top/sides depending on the screen orientation but I base everything relative to that -1 to +1. If I say some text is 0.5 high then it will take up 1/4 of the screen (landscape).

 

For the 9patch, right now I code in a conversion factor, I say this unscaled, as you load it patch would be 0.2 high this I can work out how big the borders are and scale the inside appropriately but I have issues with that too. Is there a clean way to do this?

 

I suppose I could assume the texture will be 1 pixel to 1 screen pixel, higher density would then make it smaller but since I then load a larger texture it would counter it.

 

the things I know are: screen resolution (pixels), screen size (dp), screen density (a bucket value), texture size (pixels). As a value of 0-1 where the inner size is and where the stretchable range is. Though this is Android it's an NDK app using opengl/c++ so I don't get any of the fancy Android ui stuff.

Share this post


Link to post
Share on other sites
Advertisement

In Android you would define a 9-patch in terms of display independent pixels (i.e. the size the pixels would be if the screen were the same physical dimensions, but only 160 dpi).

In a Java application you would use DisplayMetrics.density in your calculations. In the NDK you can write a JNI function to retrieve this value, or pass it in from a Java shim that bootstraps your application.

Share this post


Link to post
Share on other sites

I don't actually get any of the Android things since this is a native application. I know about display independent pixels and use them to an extent You can access these directly in NDK. The problem I have is working out a good way to go from an image to how big it represents in my space (-1 to +1) without having to hardcore something.

 

I may have a 9patch image that is 90 pixels by 90 pixels, the middle 9th might be the inner area. The problem is how big is that actual inner area? All I know is it is 30x30 pixels, how many dp that is depends on the density of the screen. On a mdpi screen the inner area would then be 30 dp but on a xxhdpi display that is suddenly only 10 dp. But then I can (and do) have larger images for larger densities so do I then make a xxhdpi image that is 270*270 so they then have the same inner size (in terms of dp).

 

I find this very confusing. Right now (I have a single 9patch) I just say - this image unscaled is 0.2 high (in my screen space) this the inner area I can work out would be 0.0667 and then I can stretch the border to fit the content. The issue I have wit that is it is hardcoded and if I end up with perhaps a larger 9patch image then it wouldn't scale as intended (unless I use another magic number..). I like using screen space (-1 to +1) since everything scales really well between devices, sometimes I do have to cap things so they don't look too big, that's where I start to use dp.

 

I did try using texture height and a 1:1 mapping with the screen min width as my reference height but my 9patch was getting stretched differently on different devices. 

 

I will try this next:

Load texture, get dimensions. Convert that to dp. then since I know the screen size in dp (min width) and the image size in dp I could convert that to a size in my screen space. Then use that to work out the inner size.

Share this post


Link to post
Share on other sites

I tried that and I get stretching happening differently again though all my textures report the same height in DP (though it is loading different textures depending on density) so that's good. I think I will try defining my sizes in dp instead of world space. My only issue with that is then they don't get larger for larger devices automatically.

 

Edit: I've used dp to define values and it occurs to me that it's really not that much of an issue if things don't get bigger in regards to text. 

Edited by Nanoha

Share this post


Link to post
Share on other sites
Sign in to follow this  

  • Advertisement
×

Important Information

By using GameDev.net, you agree to our community Guidelines, Terms of Use, and Privacy Policy.

We are the game development community.

Whether you are an indie, hobbyist, AAA developer, or just trying to learn, GameDev.net is the place for you to learn, share, and connect with the games industry. Learn more About Us or sign up!

Sign me up!