I am new to WebRTC stuff. I cloned the webrtc_ios main branch, and I built the framework as instructed here with the python script for arm64. When I add this to my Xcode project as a framework, everything is fine. Project builds, I can import files using <WebRTC/...> syntax.
However, I need to use RTCMTLRendeder.h file. Building a framework with python script leaves some of the header files out. (When I take a look at WebRTC.h inside the built framework, I can see that this file is missing) How can I include all header files that actually exist inside /webrtc_ios/src/sdk/objc/components folder while building the framework? I can see RTCMTLRenderer.h and .mm files are in that folder before using the build script. When turned into a framework those files don't exist inside the framework anymore. Why? And is there any other way to actually copy those files into the project as well?
Turns out you need to create your own, long renderer class which does not inherit from RTCMTLRenderer at all if you want to render on Metal view manually in a Swift/ObjC hybrid project (at least this is how I solved it). That class does whatever RTCMTLRenderer does, and grabs pixel buffers from RTCVideoTrack through an RTCVideoView.
Related
A task that I thought would be pretty simple turns out a bit more complicated.
I have a .json file inside my Kotlin Multiplatform module and I would like to share it between iOS and Android (to not include it for both platforms separately). My first naive approach was to add it into a folder inside commonMain. Well that did not work (accessing the file from the two platforms results in "null", but I am not even sure that the resource was bundled with the module).
Then I checked these instructions: https://luisramos.dev/how-to-share-resources-kmm,
also did not work.
I know about this library: https://github.com/icerockdev/moko-resources but I would not want to include a library just to share a file.
Is there an easy way to say "include this folder with all its files" inside the Multiplatform module and make it accessible for all platforms? (Accessing the file can be platform dependent, but just the file should be at one place)
I have created a Razor Class Library to be able to distribute some global styles and views across projects, but I can't seem to import my .less files from the Razor Class Library.
In my RCL I have:
Styles
-Shared
-layout.less
-variables.less
In the project referencing the RCL I have:
Styles
-main.less
main.less only has:
#import "./Shared/variables.less";
When I run a build via webpack I get errors that it cannot resolve ./Shared/variables.less, but it works fine if I copy the Shared directory from the RCL into the project referencing it.
I have the BuildAction for the RCL .less files set to Content, is there anything I am missing, or is this something that is not possible?
It's some what possible, but not for specifically what you're trying to do here. An RCL is ultimately a DLL, so the only thing you can include in it, is things that can be "compiled" into that. I say compiled with quotes, because static files can be included as embedded resources, so while they're not themselves compiled, they are still literally being embedded into the resulting DLL. With the ManifestEmbeddedFileProvider, the app using your RCL can read from these embedded resources as if they were literally on the filesystem.
However, and importantly, they are not on the the filesystem, which means using things like webpack is a 100% no-go. What you'll need to do is actually do a webpack build as part of the RCL, and then embed the resulting static resources in the RCL. Your app, then, can have its own webpack build for it's own resources, but you won't be able to combine primitives from the RCL with primitives from your app.
You could possibly manually manage the build order and run a powershell script post build (making sure that the correct dll is building first) and interrogate the dll and extract the required files (?) into the correct folders in wwwroot, or wherever, before the webpack build. I think that happens after the projects are built, but I'm not an expert on webpack either and I haven't actually tried this.
Also technically if you want any static or view files to be embedded in the dll you would select "Embedded resource", well that is how I've done it in the past.
I'am a newcomer to cocoa developement and i have not yet got used to.
When use eclipse for java development, we can see all the added jars source file if we attached the source files.
But in xcode, we add a framework such as "Cocoa.framework",there is only one Cocoa.h file we can see.
eg:
I clone the 'Sonora' source code from github for study.
there is one line
#import "SNRFileCopyManager.h"
exist in SNRFileImportOperation.h file
Both the SNRFileImportOperation.h and SNRFileCopyManager.h can be found in Sonora/Classes directory.
But, i can see the SNRFileImportOperation.h file in xcode project only and they get work fine together, Why this non third party header file not display in it's own project?
I know something about the dependency setting adn header search paths.
Any one give me a guidance or some references of the code management|organization?
Thanks a lot.
The file SNRFileCopyManager.h is found because it is in one of the include search paths. If you select a header file in Xcode and look at the info panel on the right side you can see that it doesn't actually belong to any target, they are shown more or less regardless. Though, building a framework you can select header files to be included into the framework when building. So, you can just pull in your header files into Xcode.
I've got a 'Main app' and a 'Helper app' (sandboxed, if that matters) that share a private framework including some resources, nib files, sound files etc.
The framework gets called and used by both apps without issues. However from within in the Framework code I have a NSViewController that loads a nib file which is included in its resource folder. This seems to work as long as its called by the 'Main app'. Doing the same with the Helper app (a login item) however does not work and fails with an "Unable to find nib named" error.
The actual 'Framework' is copied to the Main app's 'Frameworks' directory and I use a #rpath in the helper app to find the framework: #executable_path/../../../../Frameworks
This setup seems to work just fine however at runtime it seems the frameworks code tries to find the named Resource under the helper app's Resource folder and not under the Framework's resource folder. Is there a setting or some flag that I can set in xcode to make the framework's always look under the exact path where the framework's executable/library is installed?
It seems the only solution is to copy the framework to the 'Helper' app as well. Resources otherwise do not get loaded if the framework was just a symbolic link to the actual framework placed inside the main app.
What you can do is making your Framework dynamic or shared like described here Dynamic Library Programming Topics
Though it is a bit of a complex process, but a very nice feature.
What else can help you?
Perhaps editing the Library Search Paths or the Framework Search Paths under your Build Settings in Xcode. there you can specify additional search paths to look for.
Even though, I would not copy the Framework to the Main app's dir. I would leave it in one place on your disk, add them to your project (Main and helper) and add the specified search path.
By the way: How is your framework implemented? Is it a folder, is it compiled, or is it only code files?
I have created a sample application to perform bump distortion, using the GPUImage framework. I added this framework to my application, but I'm seeing the following error
Lexical or preprocessor issue 'GPUImage.h' file not found.
I have added the -ObjC flag to the Other Linker Flags, but I'm still seeing this error. How can I solve this problem and get my application to compile?
Adding GPUImage framework to XCode project could be tricky. So I haved added detailed step-by-step instructions w/ images on how to do it.
Static Compilation Method (detailed solution so we don't mess up)
This is Static compilation method. In this basically we will compile the framework using ./build.sh file. And simply add it to our XCode project, then configure XCode to properly use it.
Download GPUImage from Github and extract it (or just clone it).
Go to the GPUImagefolder in terminal
Run ./build.sh
Note: This will compile and create ready-to-use binary for all the sdks on your mac.
build.sh creates a folder called build and generates compiled binaries and dumps them to folders like: Release-iPhone, Release-iPhoneOS, Release-iphonesimulator etc folders.
For iPhone use Release-iphone (This also works for simulator).
Copy (not drag-drop) Release-iphone to your XCode project's root directory so that we have a local copy of framework.
Now iPhone drag-and-drop Release-iphone onto your XCode project. Make sure to check "Copy to .." option.
Note:
This Release-iphone folder contains two sub-folders: include and lib
include folder contains all the header .h files
lib folder contains compiled binary version file called libGPUImage.a
We now need to simply configure XCode to use .h and .a files.**
Select your project in the project explorer > Project name under Targets > select Build Phases > Expand Link Binary With Libraries
Add the libGPUImage.a to Link Binary With Libraries section. You may want to Right-click on libGPUImage.a then Open in Finder and finally drag-drop it.
While we are at it, also add the following GPUImage's dependent frameworks/ libraries
CoreMedia, CoreVideo, OpenGLES, AVFoundation, QuartzCore to Link Binary With Libraries section
Now, lets configure .h headers.
Select your project in the project explorer > Project name under Targets > select Build Settings > and type search paths to see search paths section.
Open Headers Search Paths by clicking on the value field.
Drag-and-drop the lib folder to that popup. Note: If it shows absolute path, change it to looks $(SRCROOT)/path/to/lib/. (You should have the framework relative to your xcode project see step 6).
Repeat 11 & 12 for Library Search Paths as well.
Additional tips: You can add .h files to Library Search Paths or Headers Search Paths, you can make them Recursive. I have a main root-folder called Dependencies folder where I keep all the dependencies like MySDK-framework including Release-iPhone. And I just have one search-path at the Dependencies (root folder) and made it recursive.
Did you follow all of the instructions from the Readme on the project page? From the installation instructions:
You'll also need to find the framework headers, so within your
project's build settings set the Header Search Paths to the relative
path from your application to the framework/ subdirectory within the
GPUImage source directory. Make this header search path recursive.
If you're seeing the above error, it means that you did not point the Header Search Paths at the right directory where you've installed GPUImage relative to your project, and / or did not click the checkbox to the left to make those search paths recursive.
I show some screenshots of where you need to go to set this in this answer, which explains something similar for the Core Plot framework. The same principles apply, only you need to find where you installed the GPUImage framework at.
I add relative path to "Header Search Paths", but there is still a error that is "not found".
Then, I add path to "User Header Search Paths", and it works.
May help you.
finally, I realized the KEY WORD is relativeļ¼
I put GPUImage source files in my project root dir:
before I solved the problem, I add Header Search Paths is:
$(PROJECT_DIR)/GPUImage/framework
$(PROJECT_DIR)/GPUImage/framework/iOS
and make them Recursive, but error is still there, So I changed Paths to:
GPUImage/framework and make it Recursive, then it works.
Hope it helps.