I'm trying compile the lastest WebRTC version for iPhone. I not need to compile the entire solution, I only need to compile the VAD module.
To do that, I have created a Xcode project and I have tried to compile the source necessary, but I have a problem with the *.s files and its assembler.
Like in the FFMPEG library, I know that I must "translate" the assembler code to an assembler code that the gcc for iPhone understand, but I don't know how I do this manually.
I have tried to create a configure file and set in it "as=gas-preprocessor.pl" (like in FFMPEG), but does not work.
Any idea? How do I run the gas-preprocessor.pl manually?
Thanks.
I'm just finishing it on iOS and has built standalone static libraries of NS/VAD/AECM and AGC, here's some tips for you, and hope you success:
1. Source File List
for standalone VAD build, you should make sure your project has all of these files(no .s file needed), and I'm not listed the header files here, you will get some header file can not be found errors, just fixing it and things will be done.)
webrtc_vad.c
vad_core.c
vad_filterbank.c
vad_gmm.c
vad_sp.c
real_fft.c
division_operations.c
complex_bit_reverse.c
cross_correlation.c
complex_fft.c
downsample_fast.c
vector_scaling_operations.c
get_scaling_square.c
energy.c
min_max_operations.c
spl_init.c
2. Adding a macro called WEBRTC_MAC
I'm not sure why the Xcode environment does not provide this macro, but it should be defined to ensure that WEBRTC_POSIX is enabled. To define this macro, adding it to a new header file or just define it in the webrtc-header-files.
3. Following these steps to build and setup a static library of WebRTC-VAD module on iOS
notice, do not use LLVM 2.0 to compile the VAD module(use GCC or LLVM GCC). cus' it can throw you lots of errors when compiling some webrtc variable declarations.
4. Using the libwebrtc_vad.a
if you got this far, things are easy to go, just include webrtc_vad.h and using the API provided by this module. and vad is working fine in my case.
hope i helped.
try to check this link https://groups.google.com/forum/?fromgroups=#!topic/discuss-webrtc/VJg-fk2-i_0 i believe you have to set inline assembly correctly. I am also onto this so let me know if you want to switch emails or something.
Related
I'm trying to build with the LAME files but I'm getting these errors on the machine.h file
What's going on?
Short answer : your defines are not what they are supposed to be. The BB10 SDK provides stdlib.h and string.h so you should either :
add a #define STDC_HEADERS 1 somewhere in the include path (inside version.h maybe, because it's seems to be included by everyone else)
add DEFINES += STDC_HEADERS to your .pro file.
You are trying to build your library inside Momentics.
That's usually the fastest way to go.
The issue is that you are skipping the whole configure part of Lame compilation, which was supposed to gather insight about the system you're trying to compile on by trial and error.
I haven't looked at Lame specificaly, but usually configure either creates an header file with all the right defines or add them in the Makefile it creates as arguments to the compiler.
Momentics, on the other way, compiles all .c|.cpp file and link them all together using qmake to handle all Qt specific bits. Momentics sets the right environment, and then there is a lot of scripts to handle all the BB10 processes (package, sign, ..).
So you'll have to provide the missing parts. Usually it's faster to create a new config.h from scratch, but sometimes you may want to use a console with the BB10 SDK environment and do a ./configure manually. Don't forget that the simulator is x86 and the real thing is ARM, so you will have take care of that too (Endianness/optimization issues).
I am building an iOS app where I want to be able to start my app in a demo mode with mocked connections. I use a static library for the mocking. However I don't want this library to be linked in my release build for the app store. I am looking for a solution to this problem.
The best solution I could imagine would be if I can define a compiler macro like this:
#define TEST_ENVIRONMENT 1
and if this macro is 1, the library would be linked. But I don't know how to setup the linking for this. Is there anything similar that would work?
I know I could create a new configuration in Xcode and manually add the -l option for the library for this configuration but this means that it can easily happen that one edits a build setting and forgets to edit the test configuration as well, or vise versa.
Best regards,
Michael
I am looking for a simple library ( and/or app - eventually want both ) example ( like a math library or whatever) for iOS which has a makefile for it that I can use as a template to make other makefiles from and learn. Static of course, (and dynamic if iOS supports it so I can have 2+ apps that share common code)
There is lots of incomplete and cryptic info out there but so far I havn't found any nice concise "with these source files" you create a makefile this way to build an iOS "fat" library I can import into other projects.
This would be on a Mac with the ios4 sdk installed.
It is always great to start with something that basically works.
I have created complex makefiles before for unix and windows and for other devices.
thanks.
The first link pictorially represents the process step by step that you've asked then the second link contains a package that allows a programmer to compile a make file based project
click me
click me
If you want to build a static library using a make file and link against it every time you build your Xcode project, you can add a "run script" build phase in your project before all the others, which runs this make file, and then add the built library to your linking phase. If you want a make file that builds the entire iOS project, I don't think it's posible (you can use the command line to compile the project without Xcode opened though).
I have an existing iPhone application, and I just wanted to make a static library out of the code, so that I can use it by a separate test application (within the workspace). The application compiles fine, but when I try to compile the library version (which has the same code files) I get the following error:
Lexical or Preprocessor Issue - "MyClass.h" file not found
However, the application still runs, even through there is this error. In fact when I compile the application (which uses the static library) it initially says no errors, and then after a second or two the above mention error then appears.
I'm basically doing:
Create a new target in the project using the Coco Touch Static Library
Add the .m files in the targets "Compile Sources"
Adding the .h files in the targets "Copy Headers"
Adding the same set of libraries in the "Link binaries with libraries"
Go into the Profile and modified the Target to be RELEASE
Any suggestions? Am I missing a step here?
I'll have a stab at this :-) I think it might be that MyClass.h has been set to private. Thus it is not added to header directory when you create the static library. You should check to see if this is so by checking the output header directory. If the header file is missing, then it's pretty sure thats whats going on.
Secondly, why use a seperate app for testing and therefore require building a static library at all? I have several static libraries and apps which I unit test with GHUnit. I do it by simply adding a new app target and configuring it to run the GHUnit iOS app. I can then add the classes I want to test and any testing frameworks such as OCMock without having to create separate projects. Basically it sounds like your testing methodolody is perhaps more complex than necessary. :-)
So, the issue turned out to be I had "-ObjC" set in the "Other Link Flags" option, which I'd put there as at one stage trying to get things working I read this was required - Dereks advice to review the compilation log worked well here
I have found an existing open source library that I would like to include in my iPhone project (the unrar source code found here: http://www.rarlab.com/rar_add.htm ).
I have compiled this source as a linked library on my Mac with "make lib" which creates the libunrar.so file just fine.
These are the makefile settings for that target:
lib: WHAT=RARDLL
lib: $(OBJECTS) $(LIB_OBJ)
#rm -f libunrar.so
$(LINK) -shared -o libunrar.so $(LDFLAGS) $(OBJECTS) $(LIB_OBJ)
Obviously I can't use this on iPhone but I imagine I should be able to compile with different options to make a static library that I can use.
Do I:
Compile it on the Mac with different make options then drag the resulting library (some kind of .a ?) into my xcode project?
or
Drag all of the source code into my xcode project and create special targets of some kind to create it?
or
Something else entirely different?
I've been working on solving my unrar problem for a couple of weeks now and I believe using this library will give me the best results but I just don't know the final steps to make use of it.
Thanks for all advice.
I think that you'll probably want to do the easiest thing first, create an empty iPhone project and bring all of the code into it, just to see if it will compile.
If so, you may want to use the empty project to aid in the development of an Objective-C wrapper, once that is complete you can bring the code into your existing project for integration and testing.
If it looks like this is code that you would like to use with other projects, or you are feeling giving and would like to provide an iPhone port of the code, go ahead and pull this together into a library.
Here is a site that documents the process.