@ben-clayton GLFW, interesting =)
@ben-clayton Hi, I have a question. I'm a python programmer and use wxPython for GUI programming. For an audio application I needed speed and took my chances with Go and that worked quite well. So far I like working with Go. Now I would like to do the GUI work with GXU but as I discovered GXUI does all the positioning and sizing by using layouts and has no SetSize or SetPosition method - only ProgressBar has a SetDesiredSize method. Is there a possibility to add a SetSize and SetPosition method to the controls to get more control over the layout of the components in the window?
We have had a number of requests for more explicit and advanced layout of controls. This should be addressed by implementing more Containers.
I haven't had much time to work on GXUI as of late, but it is in my TODO list
I am however happy to accept pull requests if anyone fancies implementing more types of Containers.
@umren Hi, I work with "portaudio-go". I use raw audio pcm to mix and process with portaudio-go. You can still find it at https://code.google.com/p/portaudio-go as long as it's available. I couldn't find it at github.
Gentleman, I looked through library seams like all OpenGL code is abstracted into "Driver". Im interested to build application that uses OpenGL for 3D visualization, I been interested whether there is an abstraction for OpenGL calls that would preserve state within a GXUI Control. So I could implement my code within Control for 3d visualization, and used GXUI for everything else. Possible? what do you guys think about it?. Some examples would be very appreciated.
Hi @eflauzo. You are correct that all the GL logic is abstracted away - this is intentional so that other non-GL implementations could be written, and I'd like to keep it that way. That said, you are not the first person to request this (see #142) and I've been thinking about how we can expose this functionality.
I think the right way to do this is to keep all the OpenGL logic still in the drivers/gl package, but perhaps have the gl package expose a special Control, which requires the driver to be one created with gl.StartDriver()
This special control would have some callbacks for performing the GL logic, but set up to be drawn within the control's bounds
Considerations we'd have to make:
• GXUI lazily redraws, and I'm assuming that most OpenGL requests are for interactive animations. This should be doable, but some optimisation might be required
• The custom GL control would likely have full access to all the GL state - I'm not sure whether we'd be able to sanely 'preserve state' like you suggest, to avoid the custom GL logic from breaking other GXUI draws. It's possible we could create a new GL context for each of these control, and display that in the GXUI context, but we're venturing into extension territory there, which is likely to cause compatibility issues.
• GXUI lazily redraws, and I'm assuming that most OpenGL requests are for interactive animations. This should be doable, but some optimisation might be required
• The custom GL control would likely have full access to all the GL state - I'm not sure whether we'd be able to sanely 'preserve state' like you suggest, to avoid the custom GL logic from breaking other GXUI draws. It's possible we could create a new GL context for each of these control, and display that in the GXUI context, but we're venturing into extension territory there, which is likely to cause compatibility issues.
• We'd also want to avoid breaking compatibility with the WebGL backend. To support WebGL, we have another layer of abstraction in the https://github.com/goxjs/gl package. As far as I'm aware, this exposes a significant subset of desktop GL in order to be able to transparently support desktop GL and WebGL (and GLES possibly in the future). @shurcooL would be able to describe more on this.
Thanks for comment. I personally would be happy with bare bones GLES / WebGL. I would guess that most of the uptakes of this library are people who want potentially run their software in web so WebGL sounds the way to go. As for preservation of state I would guess maybe not most performant, but most safe way would be each (GLCanvas Widget) would have it own context, where it would draw his content to glFrameBuffer -> Texture which would be then binded to that GLWIdget. If i think about it, I would think that most of the people who would needed it would need some visualization where it is not the main goal of application, so they would be willing to sacrifice on performance, otherwise they would write whole thing in WebGL.
As for updates, it widget just handle onPaint event, scheduling would be still done by gxui
Specifically, if I wanted to change the color of a button when it is pressed, I have to modify the theme, which means modifying gxui. It seems like a theme could be specified by the user, since the light and dark themes are really a collection of colours.
Is there a reason why users of the library were not allowed to specify a theme at runtime?
Hi, If I want to open a sub window while in the main window, what is the right way to connect the sub window to the main window? Also catching the close event of the sub window so that the adjustments in the sub window can be processed. I couldn't find any examples. If anyone could give me an example that would be nice. Thanks.
of controls - the original goal was that the dark package provided a very simple example of implementing a theme
The light theme was pretty much a fork of that
I'm not opposed to making the theme colors controllable at runtime, but it sounds like you're wanting to control a button's colors on a per-instance basis
In that case, I'd probably recommend 'deriving' your own version of Button and overriding the Paint method as you see fit
Theme has CreateWindow() to spawn new windows. You can add listeners to that window's OnClose event.
You probably know all this, so I'm probably just not understanding your question
What is the correct method to overwrite text (for completions, snippets, electric braces, or other forms of automatic code writing) in a mixins.CodeEditor without access to mixins.CodeEditor.TextBox.controller? I see the code editor using the controller field to apply code suggestions and indentation, but the field isn't exported and doesn't have any getters, so far as I can see. I suppose I can use SetText, but that seems overkill.
On a similar note, how bad an idea would it be to export access to that controller field? I find myself wanting access to it more and more as I go on; but that may just mean that I should be implementing my own gxui.CodeEditor without using the mixins.
@ben-clayton When I put my question on the board I didn't but in the meantime I found out how. But now a strange problem pops up. On my linux labtop everything run as it should be but when compile for Windows the main window is fine but the second window stays black. The controls are missing. Could it be a memory problem? The windows pc has 2GB of ram memory, that should be enough for such a simple program is my guess.
guess I missed the font, seems font is an interface, what's easiest way to add the font?
@kelwang - there is a default font loaded in the theme here: https://github.com/google/gxui/blob/5c8a378579b9c8bb96535e45f41a87a43154c8a9/themes/dark/theme.go#L16
This default font is Roboto Regular:
https://github.com/google/gxui/blob/5c8a378579b9c8bb96535e45f41a87a43154c8a9/gxfont/gxfont.go#L24
https://github.com/google/gxui/blob/5c8a378579b9c8bb96535e45f41a87a43154c8a9/gxfont/gxfont.go#L24
You can use custom fonts by using Driver.CreateFont:
https://github.com/google/gxui/blob/master/driver.go#L29
https://github.com/google/gxui/blob/master/driver.go#L29
@S0lll0s - sorry I haven't got around to uploading screenshots. I see you've found the PR, I think adding them to another branch is a good idea. There is the
gh-pages branch which is what gxui.io points to. Perhaps that is the most sensible place for them
www.gxui.io that is
