For me, a library is a collection of classes that do useful things. Typically something, that can be useful in a lot of projects. Is that also the case in terms of objective-c? What exactly is a library there? Only classes that have methods? Or also collections of functions? And do they have to be compiled to be called a "library"? Where is the segregation between an "Framework"? Aren't bove the same thing?
According to Wikipedia: "Frameworks are functionally similar to shared libraries, a compiled object that can be dynamically loaded into a program's address space at runtime, but frameworks add associated resources, header files, and documentation."
A framework is essentially a shared library (binary, similar to a DLL) in a bundle that also includes all of the information needed to use that library (like header files, documentation, internationalization resources, etc). A framework without all of the extras is just a library.
There is no requirement that a framework be object-oriented in nature, though I assume that's the norm with Cocoa.
For Cocoa, the concept of a framework generally replaces (enhances) the concept of a library. However, the Objective-C toolchain imposes no such requirement. You can use source-only "libraries" or unix-style binary libraries (e.g. an .so file). I think of a "library" in these generic terms... it's just a collection of useful code, in source or binary form. A framework, on the other hand, is a specific thing with a specific meaning for OS X.
Assuming you are talking about a library that uses the Cocoa frameworks and not just one written in the plain old Objective-C language, a library (or framework) is a collection of classes that work together to perform a specific task. I would not organize an ObjC framework as a collection of functions since that totally goes against the paradigm of the language.
As for the difference between a library and framework, that's probably a bit subjective. To me, a library (in the context of your question) is something written in C that probably more closely resembles a non-OO collection of functions. A framework would be a full package of classes as I described above. So the Messaging framework on CocoaDev would be a framework, whereas the sqlite3 APIs you can access on the iPhone would be a library. Again, that's just me. Other people may interpret the terms differently.
Related
I want to bind a different programming language to the Godot game engine. Is there an instructional document or video on this topic? For example, how was this project done: godot-rust. If I can learn the basics, I can succeed in working in a different language. Thanks.
In this answer I show you the different approaches to add language support in Godot 3.x (the situation will be somewhat different with Godot 4.0 and GDExtension - which replaces GDNative and hopefully means less custom builds), and I mention some languages that are supported by each of these approaches. However, this is not an exhaustive list of the languages.
First of all, Godot has official build-in support for GDScript and Godot's VisualScript (and Godot's shading language and its visual counterpart if those counts for you).
There are a few ways to use C++:
You can use it to create GDNative scripts (which are basically a wrapper around native calls that allow you to use them as scripts in Godot).
Or you can create modules (which are static libraries you can add in a custom Godot build).
And since Godot source is in C++, you don't have to restrict yourself to making modules if you are making custom builds.
In web builds Godot can interface with JavaScript via the JavaScript class. However, this approach does not allow you to add JavaScript scripts to Nodes, and so on.
Then there are languages that can only be added in custom builds of Godot, which is currently the official support for C#.
There are other non-official custom builds that offer language binding for languages such as Lua, Kotlin, TypeScript and JavaScript (this time allowing you to make scripts).
If you need to add a runtime, you would probably do this.
Some language take advantage of the fact that Godot's has official Mono support in order to support C#. This way you can, for example, use F# and Clojure.
They start by adding a C# project and then modify it so it uses another language. This is viable if your language already compiles to .NET.
Some other languages can be added as plugins that implement the PluginScript class via GDNative. This is the case of Python and Lua (again) which you can get from the asset library.
This is the most user friendly way to add language support to Godot, but it is limited to what you can do with PluginScript.
Addendum: Gil Barbosa Reis, author of the aforementioned Lua bindings, has an article series about its implementation stuffed away in the repository (in English and Portugueses): godot-lua-pluginscript/extras/articles/. It is probably the most comprehensive tutorial to date.
Other languages are added by means of taking advantage of GDNative (They basically mimic what you would do with C++). This is the case of Nim, Rust, D, Haskell, Go, Swift…
So that's how godot-rust works: make native libraries using rust and the godot-rust create and add them as if they were made in C++. For any language for which there are the means to make native libraries already, this is a good option.
Finally there is another way to add support for a language: a transpiler from that language to GDScript, which can be automated with an addon that might also be written in GDScript. This is the case of Lisp.
This last approach is mostly used for domain specific languages.
The official docs here provide your answer:
Godot officially supports GDScript, C/C++, C#.
Some 3rd party languages that can be used are: Rust, D, Python, Nim, and Go.
I saw the two projects are quite related, but what are the differences between them? The official webpage doesn't tell much about it.
I know that ABI (Application Binary Interface) is used to provide low-level binary interface among different platforms. So is libc++abi used to provide different implementations for different platforms, and general interface for libc++?
Would be better go give some specific example, e.g. what are included in libc++abi and what in libc++.
Thanks.
The Application Binary Interface, or ABI for short, is intended to provide certain low level functions from which to build the C++ standard library. It is a supporting library that is a separate component from the actual standard library. Along with libcxxabi, you may also come across Pathscale's libcxxrt or GCC's libsupcxx.
On the other hand, libc++ is an implementation of the C++ standard library that can be built using either of the 3 mentioned ABIs.
Background: I want to develop a component therefore building a class library.
This component should be usable with many higher-order languages such as C, C#, C++, VB, Java, Haskell, Ruby, Erlang, ...
I do not want to exclude any users which are not using my development language.
Are there principles or tools which supports my project?
I searched a little bit and found Haxle for compiling into different languages, but it supports very few of them.
I would even develop parallel in all n languages to be supported, but if I want to change or fix something I have to maintain all other n-1 and the code is possibly distributed...
This is not what I know about clean code design and maintainability. So how to manage edits on code for the different languages?
What is the proper way to solve this? I am surely not the first one which want to build a library for multiple languages.
I want to develop this library (in this case for complexity reduction) in one target language but this time for multiple platforms (Unix, Win, Mac ...).
How to manage this? In fact there will be appreciated about 90% of code which is platform independent and 10% which differs for every OS.
What is the best way to control the changes in the platform dependent code? (The independent is easy ...)
What if I change so things in the part which uses Unix dependent code, then I have to trail all other platform and the code is possibly distributed.
I think #IFDEF is no option ...
Are there any experiences or hints?
I would be delighted if there are existing solutions to these problems, which are quite similar.
What you want is only possible if you provide interfaces for every language you want to support. Some kind of wrapper which transforms between the client language and ypu library.
This is possible but not practical in most cases for standalone libraries.
Take a look at webservices or or message orientad middleware. In this case your application will be provided within a special container which itself provides interface mechanism e.g.
SOAP, XML-RPC to call your application.
For 1), I would use an interface that all these languages can use. A common approach is something based on networking/TCP, like protobuffer, REST, SOAP. Many languages support these in their standard libraries, and interfaces designed with that are normally language agnostic.
There are basically two options - you could develop a network server or you could develop a JVM-based library which could be shared between some of the JVM languages like JRuby or Jython.
Update from #millimoose: you could also develop your library in C and create bindings for all other languages.
For multi-language:
I thought about TCP/middleware/webservices/REST/ which seems to be the recommended proceeding.
But I think it's all to much at runtime for only using a library.
Also the functionality is a little bit time-critical and so direct procedure calls are more fitting (instead of networking even on localhost). And the library user hasn't to construct an access component only to use the library functions.
So I think the way to go seems to be developing the library in a core language which is widely supported (C/C++, ...) and provide wrapper interfaces for the different target languages.
For multi-platform (mono-language):
No real answer has been provided also not in my mind.
Of course I could simply use Java (what I am familiar with) but what about other languages?
I am surely not the first one having this/these problem(s) ...
We'd like to offer a compiled library that implement a protocol layer to be imported into C/C++ source code project for microcontrollers. And eventually expose a sort of compiled function to the source code project. let's say a sort of "dll". Is there any know technique to realize something of similar?
While it is possible to provide functions via a library, generally in the microcontroller/embedded realm it quickly becomes impractical.
Each microcontroller core will have a unique instruction set. Further, micros from the same family may have a variety of extensions which are either supported or not... So you're left with providing a library file for each individual microcontroller (from each vendor) that you'd like to support.
But...
In my experience, calling conventions between compilers are not the same. So a library compiled by one toolchain will not be able to be linked to object files created by another toolchain.
That leads you to then provide a library for each individual micro from each vendor for each toolchain someone might use. Ick. Oh, and don't rely on an OS calls either, as you don't know what you'll be linked with...
A more conventional approach is to use the same approach RTOS vendors tend to use: provide the source, and protect your IP with licensing terms. The reality is that if your end users want to, they can step through the assembly and figure out exactly what is happening, so you're not hiding your implementation that carefully anyway.
There are lots of oop languages, but I couldn't find any that has conveniences like garbage collection, but compiles natively to machine code. Sort of like between C and java/c#. One interesting language I found was Vala, but that's limited to the GNOME platform and is not that well-known
Go is probably closest.
But why on earth do you want it natively compiled anyway?
JIT compilation of portable bytecode has proved to be an extremely effective strategy. It compiles down to native code at runtime (so you get up to the performance of native code after the first few iterations) and it avoids the issues of having to build and manage platform-specific compiled binaries.
Are you thinking about C++? That is in high usage and can be compiled on nearly any (major) platform.
In the case you want to use an oo language that compiles down to native code you will "always" have to use header files and stuff as the elf format doesn't support oo (There is no class information in elf).In case you want to use classes from external libs you need to make the compiler aware somehow about the fact that there are classes, functions, etc. that are declared outside of your project. In C++ this is solved by the use of header files. So that's, I think, a major drawback in native object oriented languages. To resolve that issue a few tweaks would need to be made to elf/loader/linker in order to support the kind of features (like "linking" against "classes") you might expect. Even though mechanism for garbage collection could be implemented even in native languages. But that seems no good for os implementation.
There are C++ libs to do that for userspace apps like:
Boehm collector
Smart pointers