llvm

问题 In Linux, and specifically on Debian or Ubuntu, there are pre-packaged Mingw cross compilers which can generate Windows EXEs. But is there a similar cross compiler using Clang instead of GCC? (Or instructions on how to build such a beast.) 回答1: I've written a tool with which you can compile easily with clang for windows on linux. Check out: https://github.com/tpoechtrager/wclang. 来源： https://stackoverflow.com/questions/12704537/is-there-a-clang-mingw-cross-compiler-for-linux

问题 Why exactly is it that when I create an iOS static library project or framework project in Xcode, I don't need to link any iOS SDK frameworks to the project in order to make use of their headers and objects -- for example, I can #import <AudioToolbox/AudioToolbox.h> and put AudioToolbox code in the static library or framework without actually having AudioToolbox added under "Link Binary with Libraries" in build settings or having it present in the file navigator, and the project will build

问题 I want to find some debugging options for clang/LLVM which work like gcc's -fdump-tree-all-all -fdump-rtl-all-all -fdump-ipa-all-all . Basically I want to have an LLVM IR dumps before and after each optimization pass, also it can be useful to have all dumps of AST from clang and all phases of code generation (backend phases, Selection DAG, ISEL-SDNode, register allocation, MCInsts). I was able to find only the clang's -ccc-print-phases , but it will only print high-level phases names, e.g.

问题 I'm seeking clarification of what the best build options are for LLVM 3.2 on Raspbian I've fetched the LLVM source of 3.2, and my config is: cd llvm-3.2.src ./configure --build=armv6-unknown-linux-gnueabi \ --host=armv6-unknown-linux-gnueabi \ --target=armv6-unknown-linux-gnueabi --with-cpu=arm1176jzf-s \ --with-float=hard --with-abi=aapcs-vfp --with-fpu=neon \ --enable-targets=arm --enable-optimized --enable-assertions make REQUIRES_RTTI=1 sudo make install I've heard it takes quite a number

问题 What is the actual relation between assembly, machine code, bytecode, and opcode? I have read most of the SO questions about assembly and machine code, such as this, but they are too high level and do not show examples of actual assembly code being transformed into machine code. As a result, I still don't understand how it works at a deeper level. The ideal answer to this question would show a specific example of some assembly code, such as the snippet below, and how each assembly instruction

问题 Previously, this was impossible (you have to write it all out by hand / create a static array / put all the values into a dictionary and read them back ... etc) But I've noticed that the latest Xcode's lldb (4.6, maybe earlier versions too) is automatically converting enum-constants to strings. My problem is that we use a lot of libraries - including Apple's own! - which use annoying public enums with no "value-to-string" method offered. So I end up having to (many, many times over) do the

问题 I'm now playing with LLVM and it's JIT. I'm pretty interested in the JIT and then I wrote a small GTK+ hello world: #include <gtk/gtk.h> int main () { gtk_init(NULL, NULL); GtkWidget *win = gtk_window_new (GTK_WINDOW_TOPLEVEL); g_signal_connect (win, "delete-event", G_CALLBACK (gtk_main_quit), NULL); GtkWidget *lbl = gtk_label_new ("hello world"); gtk_container_add (GTK_CONTAINER (win), lbl); gtk_widget_show_all (win); gtk_main(); return 0; } I compiled it into Bitcode this way: clang -emit

问题 I write a simple llvm plugin pass that requires opt to load xxx.so file and run a ModulePass. The strange thing is that when I use deb package opt (e.g., from apt-get , let's call it opt-3.7 ), the plugin works fine (the drawback is that it is a Release build); however when I use the opt I build myself (simplify call it opt ), it frequently complains: Error opening 'xxx.so': xxx.so: undefined symbol: _ZNK4llvm12FunctionPass17createPrinterPassERNS_11raw_ostreamERKNSt7__cxx1112basic

问题 Recently I upgraded my project settings in Xcode 4.3 and now I use the latest llvm debugger: lldb However (sometimes) I have the impression the debugger is not giving me the correct info? Could this be possible? For example, The debugger says _documentsItem is nil (both in in the console and when mouse-over-ing the ivar). But I know it's NOT , that is why I can see it (an UIBarButtonItem) and more important that is why the app stopped at the shown breakpoint) if (_documentsItem) { ... In fact

问题 I'm using (Apple) LLVM 4.1 within Xcode 4.5 (but I saw the same thing in an earlier beta). I have it targeting C++11 and am using libc++ (the new, C++11 enabled, LLVM version of the std library). In a debug build, if I try to step over code that calls into a std library function (e.g. a std::vector constructor) the debugger stops in the std lib implementation and if I try to step out it steps even deeper! It can take several (sometimes 10 or more) step-outs before I get back to my own code