I have 2 frameworks created by me that use (both of them) a library also created by me.
The first framework initialize the library and makes all its workflow. After finishing the first framework, the second one must start.
But when the second one is going to start, after initializing the library, the app using both frameworks crashes with a exc_bad_access error.
Apparently the library is created correctly, but if i comment the line of code to initialize the library in the second framework, the workflow continues (it crashes later because it has no library initialization).
Is there anything I'm doing wrong? Should I use two separate libraries instead?
EDIT:
Imagine the situation:
Framework A has this methods: start, stop. And while it works it delegate to the methods: infoFromA,frameworkAFinished.
Framework B has this methods: start, stop. And while it works it delegate to the methods: infoFromB,frameworkBFinished.
Both start methods initialize the static library mentioned (lets call it problematicLibrary).
Both frameworks present some views to make its functionality. So lets make an example of the app workflow.
At the app view viewWillAppear method, I start the Framework A just using:
[FrameworkA start]; , this will initialize the library and present a view. With this view (using my problematicLibrary) some info will be delegated to the infoFromA delegated method. And after all the info is delegated, it will delegate to frameworkAFinished.
When the FrameworkA has delegated to the frameworkAFinished I start the next Framework: [FrameworkB start]. As the other Framework, it will initialize the library and present a view. While debugging, all the initialization of the library is done (create the instances of the needed objects and a new instance of the library is created) and while its presenting the view it goes through the viewDidLoad method and then it throws an exc_bad_access error at the problematicLibrary initialization line (which has been done before and continued to present the view!!) without going into any other method from the view.
I have checked if the initialization is doing well, and all the variables were at null value before the initialization, and a new memory address is given to the library object.
This sounds strongly like a symbol conflict to me. I'm just surprised the linker didn't catch it, but I assume that's because you're using a static library in both your frameworks instead of simply yet another framework.
Generally speaking, I'd warn that "this is a bad idea™". What you're trying to introduce into your design is basically dependency management. Like a lot of blog articles and specifically this SO answer suggest, you should avoid packaging frameworks (and by extension libraries) into frameworks.
What most likely happens in your scenario is this (I admit I'm guessing a bit here): You linked the library into Framework A. Thus, the library becomes a fixed part of it. Its symbols are in it, even if you did not expose them to the app in any header files or the like. As long as you use only that, everything works smoothly. Then comes Framework B, of which the library is also a fixed part. Even though you can't see it from your app, the very same symbols are inside it. This, however, clashes with the symbols that were already loaded by Framework A, hence a crash (as said, this would usually be caught by the linker, but I guess you "tricked" it by building the frameworks beforehand and packaged the library in them). Maybe somebody else can explain this in more detail, but that quickly becomes besides the point as to how you would go for a solution. From how I see it, it just won't work this way.
So here's a suggestion for how you can solve your problem:
If you really, really need to split like this (two frameworks used in your app using the same dependency), I'd suggest removing the library from the frameworks (i.e. make them depend on it, but not package the actual .a file in them) and documenting that properly. Then add the library to your app project, just like you added the frameworks.
If you want to make this fancy and easily installable into other apps, I'd recommend setting up a private CocoaPods repository and turn your frameworks into private pods. You can then easily define the library (or rather "new Framework C") as a dependency for Framework A and Framework B. When you then pod install in your app, cocoapods figures out the dependency and adds Framework C automatically to your project. This scenario is exactly what cocoapods (or any dependency manager for that matter) was designed for. It automates and helps in the project setup, so that the final build (the app) doesn't have to figure out dynamically what it can and can't use. The end result is the same.
Trying to duplicate that "in code" quickly becomes messy. Frameworks trying to figure out things of the surrounding app/project that uses them (like "is my dependency so-and-so already linked? if not, can I load my own version of the library?") can lead to a lot of pain.
Okay, in response to your comment I'll try my hand at a more detailed How-To for the non-cocoapods setup. As a preface, though, let me say that that's kinda hard to do on top of my head, as I have no ready sample project to toy around with. Since this is one of those "set it up once and then forget aout it for a long time" I have to admit my recollection of these things is a bit fuzzy, so consider this as a sort of "rough direction". There might be things you need to configure differently than how I recall them. Other SO user are thus hereby invited to edit and correct my answer here, too. :)
First, I have to say I am not exactly sure whether you need to convert your static library into a framework or not for this, I think you don't so I'll go from here (I have never used a static library in that way).
That means you can leave the project that builds your library as is. On second thought, you probably have to do this to be able to link against the library without making it a part of the framework that uses it. I will still refer to this code as "library" in the below points, but assume that you're able to turn it into a dynamic framework as well.
The projects building Framework A and Framework B should be changed. I don't know how you have this set up (as one project with various targets, whether you have a "development application" as part of it to test the frameworks on themselves, etc.), but the important thing is that in the target that builds a framework, the library should be linked (i.e. in the "Link Binary With Libraries" build phase), but not copied (i.e. it must not be in the "Copy Bundle Ressources" build phase). It might be tricky to set up any development/test target you use to run, depending on how you did that so far. For example you might have to manually edit Library Search paths in your build settings to be able to properly compile the framework.
Lastly, you need to change your final app's project settings, obviously. The library that was originally part of Framework A and B now needs to be linked to from its project directly, and, obviously, it needs to be copied into the bundle as well. Note that any projects that include either of your frameworks (A or B or both) in the future must do so, otherwise they won't work, because these frameworks expect the library to be present (but no longer "bring it with them").
In spite of this long-ish text wall, it shouldn't be that complicated, I think, but you may still want to check out how cocoapods can support you with something like this, perhaps it's helpful in the future. The linked article expects basic knowledge on how to use a pod and write one, but the other guides on the site explain that pretty well. I am just saying this because in the end, when using cocoapods in a project to add multiple libraries/frameworks that introduce dependencies, it basically sets up your project in the way I described above. It uses some fancy shell scripts that ensure everything is copied to the correct folders and such, but overall speaking that's how it works: A pod's podspec tells cocoapods what additional pods to include in your app for it to work (a dependecy the pod expects to be there, but doesn't "bring in" by itself).
Check if they are both compiling for the same target.
Check if they have access to the same target membership.
Check build phases to see that they are both there.
I think because the first library is not 'well' referencing the second one.
Also i think that you can't import a framework inside another one.
To make things easier, you can merge the two frameworks on a single one.
Also you can add a callback (using protocols or closures) that informs for the end of the first workflow, and you use this callback to initialize the second framework. This way the initialization will not be made automatically.
Moving from the old Effects Framework to the new world of SharpDX 3 (which has no effects framework) means I need to re-work my effect compilation. My old effect files have VS and PS entry points and the Shaderbytecode.CompileFromFile deals with this - but I'm trying to track down a bug where I have shaders which have dependent functions (which might be related to the bug - I'm just trying to eliminate possibilities).
Question : when compiling an entry point with Shaderbytecode.CompileFromFile does the resulting bytecode contain all the instructions for the dependent functions as well ? ShaderReflection on the bytecode doesn't help to find this out, so hoping someone will know the definitive answer.
To answer my own question for reference;
Yes - if you remove a dependent function from the HLSL source, the compiler will throw an error. Although this doesn't prove the bytecode contains the instructions for the function, its clearly being compiled.
We need to reduce the memory consumption of our title, initially for iOS and then later for Android.
One of the areas we're looking at is code stripping as suggested in the article Optimizing the Size of the Built iOS Player.
Testing the various stripping levels, assemblies and byte code, we are experiencing a crash at runtime. I have narrowed this down to using the AssetBundle.LoadAsync() method, replacing its usage with AssetBundle.Load() calls. Whilst this is has stopped the crash, it has severely broken parts of the game that at this late stage we don't have the time to repair.
So, taking a step back - the code stripping is removing a dependency of AssetBundle.LoadAsync() that causes the game to crash at runtime. In the linked article it says to use a link.xml file to specify additional dependencies (I've had to add System.Security.Cryptography to this list).
Q: What are the dependencies required for AssetBundle.LoadAsync()?
Also, are there any tips for working out dependencies? I had a peek in .NET Reflector however that didn't yield much information as it just calls an external DLL.
I am using an fmod plugin for Unity3D. Compiling to Windows and OSX is fine because I can dynamically load the DLL/dylib.
The problem comes when I compile for iOS. I use
[DllImport("__Internal")]
Because iOS requires statically linked libraries. When I compile though I get a
SystemException: Duplicate native method found : FMOD_System_CreateSound. Please check your source carefully.
I am quite sure I don't duplicate the symbol. I think this might be due to the fact that Unity imports FMODs itself and that the two might be colliding... But if this is the case, I am surprised that FMOD_System_CreateSound is the first one to get caught. Is there a way around this? thx!
As always, I will be happy to provide any additional details!
Here is a sample project that will cause the error:
Sample Unity Project with FMod
EDIT:
The conflict was caused by iOS not allowing functions to have the same name even though they don't have the same signature. After removing the same-named functions (thus removing some FMOD features that I didn't need), I can compile to iOS, but as expected, I still get an error when Initializing because FMOD is already initialized by Unity.
Unity3d already has a limited version of FMOD that is bundled with it, which is causing the conflict you are seeing. Unfortunately, it doesn't seem possible to disable it at this time, so that you can use the full version of FMOD
In reference to your edit and after looking at the sample, it is true that you cannot have two methods of the same name as the compiler will not recognize which to link to.
The easy fix is to obviously name them differently.
As for the initialization, if you can access the FMOD that Unity 3D already created, then you don't have to reinitialize it.
I assume that a pointer to that object will be sufficient to remove the duplicate initialization. Hope this is clear.
I'm revisiting an an older project and converting to ARC, my first time through Xcode's conversion tool (Edit -> Refactor -> Convert to Objective-C ARC...), and I'm seeing a couple things that I'm not sure are real issues or red herrings somehow.
I get, as expected a big list of things that the tool finds that prevent it from completing, but:
Many (all?) instances of retain/release/autorelease appear to be flagged as errors e.g. "release is unavailable: not available in automatic reference counting mode". Am I really supposed to get rid of all these myself? I thought that's what the tool did.
In many of my classes, I'm seeing a bunch of errors that look like phantom parse/build errors that have nothing to do with ARC. E.g. in a simple class that apparently has no ARC-related issues, I'll get an "undeclared identifier" on some arbitrary method implementation, and then a bunch of "Parse error: expected }" at the end of the file, etc. These are not real-- the project builds fine, and I don't see any proximate cause or resolution for the errors.
There are "real" issues in the list as well (expected bridging issues that need to be explicitly clarified in code) but there are so many random errors of the above variety that it's hard to even find the signal in the noise. This seems wrong to me.
Am I misunderstanding what this tool is really doing? Apple's docs say this:
Xcode provides a tool that automates the mechanical parts of the ARC
conversion (such as removing retain and release calls) and helps you
to fix issues the migrator can’t handle automatically
Thanks.
The tool does not get rid of them for you, but simply adds retain/release code as need under the hood at the time of compile.
Those problems very well may go away when you get rid of old reference counting code.
EDIT: Further explanation:
In Xcode 4.2, in addition to syntax checking as you type, the new
Apple LLVM compiler makes it possible to offload the burden of manual
memory management to the compiler, introspecting your code to decide
when to release objects. Apple’s documentation describes ARC as
follows:
“Automatic Reference Counting (ARC) is a compiler-level feature that
simplifies the process of managing object lifetimes (memory
management) in Cocoa applications.”
In other words, ARC does not "strip" reference counting from your code, but rather does it on it's own under the hood. You no longer have to type release or retain or dealloc again. One thing the ARC needs to work is for it to do the reference counting entirely on it's own (with no user reference counting to "get in the way").
Took a long time to resolve, but both of these issues seemed to stem from some custom macros I was using. I had a macro for release-and-set-to-nil that I was using frequently, like this:
#define RELEASENIL(x) [(x) release]; \
(x) = nil;
I'm still not sure why, but for some reason, the ARC conversion tool didn't take this in stride, and choked on it, throwing the release warnings and the parse errors. (Some interaction with the preprocessor?) When I changed the macro to remove the release line, the conversion proceeded much more in line with my expectations.
And yes, it does of course remove the messages for you. (I'm answering my own question on the off chance that someone else ever has this issue.)