Using CMake to resolve #include dependencies in an exotic language - include

I'm trying to use CMake as my build tool driver (I make it clear at once that I'm quite new to CMake since this my first on-my-own CMake project). My project is mainly C with some files in an exotic language I'll call here Z. These Z files must be processed by their Z compiler to produce .h and .c files.
I managed to tweak CMake handle Z compiling and dependencies between plain C files and generated Z -> .h header with
add_custom_command( ...*details omitted*... )
set_source_file_properties(*generated-.h* PROPERTIES GENERATED TRUE)
and CMake C-#include scanner does properly the rest of the job.
Now, Z files use something equivalent to C #include construct and I'd like to make profit of automatic recompilation when one of the Z-included files changes.
If a.Z includes inc.Z, I tried:
set_source_file_properties(a.Z PROPERTIES OBJECT_DEPENDS inc.Z)
but that doesn't trigger automatic recompilation of a.Z.
CMake manual says this property was introduced for this purpose and is no longer necessary for C/C++. However it is ineffective in my Z case.
If I modify my custom command as follows:
add_custom_command( ... DEPENDS inc.Z ...)
I get the desired result, but not all Z files depend on inc.Z (another one might depend on inc2.Z).
I then tried to generate dynamically the DEPENDS list with
get_source_file_property(dependencies ${filename} OBJECT_DEPENDS)
if(dependencies STREQUAL "NOTFOUND")
unset(dependencies)
else()
string(REPLACE ";" " " dependencies "${dependencies}")
endif()
add_custom_command(... DEPENDS $filename "$dependencies" ...)
and make errors out with
Make[2]: *** no rules to build target `-- content of dependencies variable --'
Note: error is the same with or without double quotes around variable substitution
I guess CMake interpreted my list of dependencies as a single (non-existent) filename and make was unable to handle that. Anyway, the files mentionned in the dependencies variable are not meant to be compiled, only to be included somewhere. They must make their way to make only as dependencies.
Which direction should I go to achieve the desired result?
Recall this is my first on-my-own CMake project and I certainly made newbie errors.
FWIW, my platform is Linux with CMake 2.8.9 (not bleeding edge but I'm only exploring) and KDevelop 4.x
Thanks for your help.

Related

Configure compilation options and compiler autoconf

I'm working on a personal project with Rust and tcl but i still want to use the classic makefile structure.
I know that to compile multifile I just need to declare mod second on main.rs and rustc automatically connect the modules. So I use
$ rustc main.rs -o output -C debuginfo=2
Now I tried to integrate autoconf and automake because I want to make a configure script to check for tcl, rustup etc... But I don't know how to edit to compile with rustc and its options insead of cc and c options (like trying a .o that doesn't compile because they don't have a main function).
for the configure.ac i used:
AC_CONFIG_SRCDIR([source/main.rs])
AC_CONFIG_AUX_DIR(config)
# I manually checked for rustup and tclsh
AM_INIT_AUTOMAKE
AC_CONFIG_FILES([Makefile])
AC_OUTPUT
for the Makefile.am:
AUTOMAKE_OPTIONS = foreign
bin_PROGRAMS = output
SUBDIRS = sources
output_SOURCES = sources/main.rs
I have the main directory with configure.ac and Makefile.am and the sources directory with all the stuff (and also the config directory for autoconf)
Now I tried to integrate autoconf and automake because I want to make a configure script to check for tcl, rustup etc...
The configure script is the responsibility of Autoconf. It is not obligatory to use Automake together with Autoconf, and you should consider whether it would be sensible for you to use Autoconf alone. That would give you complete control over the generated Makefile, as you would write a Makefile.in directly instead of relying on Automake to do that for you. Presumably, you would write a much simpler Makefile.in than Automake generates, and that's fine.
Automake is not necessarily out of the question, but its manual has this to say about language support:
Automake currently only includes full support for C, C++ (see C++
Support), Objective C (see Objective C Support), Objective C++ (see
Objective C++ Support), Fortran 77 (see Fortran 77 Support), Fortran
9x (see Fortran 9x Support), and Java (see Java Support with gcj).
There is only rudimentary support for other languages, support for
which will be improved based on user demand.
Some limited support for adding your own languages is available via
the suffix rule handling (see Suffixes).
The referenced section about suffix rules shows how you might use such a rule to teach Automake how to build Rust programs. It might look something like this:
.rs:
$(RUSTC) $< -o $# $(AM_RUSTFLAGS) $(RUSTFLAGS)
SUFFIXES = .rs
That assumes that configure will identify the Rust compiler and export its name as RUSTC. AM_RUSTFLAGS is for defining compilation flags internally in your project (typically in your Makefile.am), and RUSTFLAGS is for the builder to add or override compilation flags at build time.
But since the compiler does not produce intermediate object files (or so I gather), I would expect that defining sources in output_SOURCES would not yield a working Makefile, and that you would probably need the name of the top-level Rust source to match the name of the wanted binary (i.e. output.rs instead of main.rs). The single-suffix rule should, then, get your binary built without any sources being explicitly specified. You would also want to name all contributing Rust sources in the EXTRA_SOURCES variable, else they would be omitted from distribution packages built via make dist.
Note, too, that the above does not define all the build dependencies that actually exist if you're building multifile programs. I would suggest doing that by adding an appropriate prerequisite-only rule, such as
output: $(output_extra_sources)
(with no recipe) in multifile cases. This will ensure that make will recognize when output needs to be rebuilt as a result of a modification to one of its sources other than output.rs.

CMake: generate and compile architecture-specific source files in XCode

I have a CMake project that's used to generate an iOS-targeted XCode project that supports multiple CPU architectures (arm64 and armv7).
My CMake project includes some commands (defined with add_custom_command) that convert Lua scripts into C++ source files. These generated C++ files differ by architecture (the armv7 file should not be compiled for arm64 and vice versa).
The tool meant to be invoked like this:
./data_tool --input <script> --output <C++ source> --architecture <armv7 or arm64>
My (incorrect) CMake file currently looks something like this:
foreach(ARCHITECTURE ${TARGET_ARCHITECTURES})
string(
REPLACE ".lua" ".cpp" GENERATED_CPP
${GENERATED_SOURCE_DIRECTORY}/${ARCHITECTURE}/${INPUT_SCRIPT}
)
add_custom_command(
OUTPUT ${GENERATED_CPP}
COMMAND ${DATA_TOOL} --input "${INPUT_SCRIPT}" --output "${GENERATED_CPP}" --architecture ${ARCHITECTURE}
MAIN_DEPENDENCY ${INPUT_SCRIPT}
)
list( APPEND GENERATED_SOURCE ${GENERATED_CPP} )
endforeach()
Later, GENERATED_SOURCE is appended to the source file list passed to add_executable. This code is obviously wrong because both the armv7 and arm64 files are compiled when building for either architecture.
How can I tell CMake that each architecture compiles a different set of source files?
XCode doesn't have a great way to exclude files based on the architecture being built. While it is possible (see Disabling some files in XCode project from compilation), setting this up via CMake is going to be somewhat difficult.
Instead, I would suggest simply making your generation tool/script put preprocessor guards around the entire file, for the architecture that the generated file supports. That way, when XCode compiles them, they will be essentially empty, except for the architecture that they are meant for. In this answer (Determine if the device is ARM64), it shows how you can do a conditional compile based on arm64 (and use the reverse for armv7).
Well, don't put generated sources for different arches into the same list. Unwrap foreach body and just repeat these commands for each arch.
If you don't want to introduce code duplication, you can write a CMake function that creates that custom command and returns a list of generated sources. See this question for how to return values from functions.

Generate list files with CMake

I hope this is a simple question and I'm just missing something fundamental.
I'm trying to emulate a binary build manager for an embedded Cortex-M0 target using a CMake project. I'm having some trouble figuring out how to generate list files for each dependency of my executable target.
The current build system, when building a file called main.c passes -Wa,-alh=.\CortexM0\ARM_GCC_493\Debug/main.lst as an argument to gcc. I can't figure out how to get CMake to use the current filename without the extension to save the file.
I've looked at the get_filename_component command, but it appears only to get the filename of the output:
add_executable(TestExe main.c)
get_filename_component(curr_name TestExe NAME_WM)
message(${curr_name})
As expected, this prints TestExe instead of the hoped for main
Is there a simple variable I'm overlooking that I could put in my toolchain file's CMAKE_C_FLAGS like -Wa,-alh=${CURR_SOURCE}.lst? Or some other method that I'm not seeing?
System info:
Windows 10
Msys shell
CMake 3.7.2
arm-none-eabi-gcc v4.9.3
You can use Expansion Rules and extend CMAKE_C_COMPILE_OBJECT:
set(CMAKE_C_COMPILE_OBJECT "${CMAKE_C_COMPILE_OBJECT} -Wa,-alh=<OBJECT>.lst")
But there is unfortunately
no Expansion Rule that does give the current source file without path and extension
so you will get in the above example main.c.o.lst as an output name
Footnote: In CMake generated makefile projects, if you just need the assembly file can just do make main.s or for the pre-processed file make main.i.

Compile subdir-objects multiple times

Automake 1.14 is causing us a few issues. At first, automake errored with the complaint:
warning: source file 'X' is in a subdirectory but option 'subdir-objects' is disabled
So I enabled subdir-objects, but now it isn't recompiling some files. For example, lets say
src/a/foo.c is compiled in SUBDIR a but in src/b, I would like to compile it again with different preprocessor flags, however since ../a/foo.o already exists, make doesn't rebuild it. This is because subdir-objects changes am_b_OBJECTS to look for ../a/foo.o instead of foo.o. Is there a way I can get around the original complaint and instruct make to build the file a second time with the appropriate preprocessor flags? This all worked on previous versions of automake.
I would settle for executing rm ../a/foo.o before compiling src/b but I don't know how to edit the Makefile.am to make that happen.
This happens if you're using subdir-objects under the same tree from different Makefile.am files. As automake can't see you're using the same source file with different parameters it'll assume it was rebuilt correctly.
The proper solution to this is to not use separate Makefile.am files and instead rephrase the build system as non-recursive automake and so in that case it would then build foo.c as foo-a.o and foo-b.o.

What is the difference between make and gcc?

The last sentence in the article caught my eye
[F]or C/C++ developers and
students interested in learning to
program in C/C++ rather than users of
Linux. This is because the compiling
of source code is made simple in
GNU/Linux by the use of the 'make'
command.
I have always used gcc to compile my C/C++ programs, whereas javac to compile my Java programs. I have only used make to install programs to my computer by configure/make/make install.
It seems that you can compile apparently all your programs with the command make.
What is the difference between make and gcc?
Well ... gcc is a compiler, make is a tool to help build programs. The difference is huge. You can never build a program purely using make; it's not a compiler. What make does it introduce a separate file of "rules", that describes how to go from source code to finished program. It then interprets this file, figures out what needs to be compiled, and calls gcc for you. This is very useful for larger projects, with hundreds or thousands of source code files, and to keep track of things like compiler options, include paths, and so on.
gcc compiles and/or links a single file. It supports multiple languages, but does not knows how to combine several source files into a non-trivial, running program - you will usually need at least two invocations of gcc (compile and link) to create even the simplest of programs.
Wikipedia page on GCC describes it as a "compiler system":
The GNU Compiler Collection (usually shortened to GCC) is a compiler system produced by the GNU Project supporting various programming languages.
make is a "build tool" that invokes the compiler (which could be gcc) in a particular sequence to compile multiple sources and link them together. It also tracks dependencies between various source files and object files that result from compilation of sources and does only the operations on components that have changed since last build.
GNUmake is one popular implementation of make. The description from GNUmake is as follows:
Make is a tool which controls the generation of executables and other non-source files of a program from the program's source files.
Make gets its knowledge of how to build your program from a file called the makefile, which lists each of the non-source files and how to compute it from other files.
gcc is a C compiler: it takes a C source file and creates machine code, either in the form of unlinked object files or as an actual executable program, which has been linked to all object modules and libraries.
make is useful for controlling the build process of a project. A typical C program consists of several modules (.c) and header files (.h). It would be time-consuming to always compile everything after you change anything, so make is designed to only compile the parts that need to be re-compiled after a change.
It does this by following rules created by the programmer. For example:
foo.o: foo.c foo.h
cc -c foo.c
This rule tells make that the file foo.o depends on the files foo.c and foo.h, and if either of them changes, it can be built by running the command on the second line. (The above is not actual syntax: make wants the commands indented by a TAB characters, which I can't do in this editing mode. Imagine it's there, though.)
make reads its rules from a file that is usually called a Makefile. Since these files are (traditionally) written by hand, make has a lot of magic to let you shorten the rules. For example, it knows that a foo.o can be built from a foo.c, and it knows what the command to do so is. Thus, the above rule could be shortened to this:
foo.o: foo.h
A small program consisting of three modules might have a Makefile like this:
mycmd: main.o foo.o bar.o
$(CC) $(LDFLAGS) -o mycmd main.o foo.o bar.o
foo.o: foo.h bar.h
bar.o: bar.h
make can do more than just compile programs. A typical Makefile will have a rule to clean out unwanted files:
clean:
rm -f *.o core myapp
Another rule might run tests:
check: myapp
./myapp < test.input > test.output
diff -u test.correct test.output
A Makefile might "build" documentation: run a tool to convert documentation from some markup language to HTML and PDF, for example.
A Makefile might have an install rule to copy the binary program it builds to wherever the user or system administrator wants it installed.
And so on. Since make is generic and powerful, it is typically used to automate the whole process from unpacking a source tarball to the point where the software is ready to be used by the user.
There is a whole lot of to learn about make if you want to learn it fully. The GNU version of make has particularly good documentation: http://www.gnu.org/software/make/manual/ has it in various forms.
Make often uses gcc to compile a multitude of C or C++ files.
Make is a tool for building any complex system where there are dependancies between the various system components, by doing the minimal amount of work necessary.
If you want to find out all the things make can be used for, the GNU make manual is excellent.
make uses a Makefile in the current directory to apply a set of rules to its input arguments. Make also knows some default rules so that it executes even if it doesn't find a Makefile (or similar) file in the current directory. The rule to execute for cpp files so happens to call gcc on many systems.
Notice that you don't call make with the input file names but rather with rule names which reflect the output. So calling make xyz will strive to execute rule xyz which by default builds a file xyz (for example based on a source code file xyz.cpp.
gcc is a compiler like javac. You give it source files, it gives you a program.
make is a build tool. It takes a file that describes how to build the files in your project based on dependencies between files, so when you change one source file, you don't have to rebuild everything (like if you used a build script). make usually uses gcc to actually compile source files.
make is essentially an expert system for building code. You set up rules for how things are built, and what they depend on. Make can then look at the timestamps on all your files and figure out exactly what needs to be rebuilt at any time.
gcc is the "gnu compiler collection". There are many languages it supports (C, C++, Ada, etc depending on your setup), but still it is just one tool out of many that make may use to build your system.
You can use make to compile your C and C++ programs by calling gcc or g++ in your makefile to do all the compilation and linking steps, allowing you to do all these steps with one simple command. It is not a replacement for the compiler.
'gcc' is the compiler - the program that actually turns the source code into an executable. You have to tell it where the source code is, what to output, and various other things like libraries and options.
'make' is more like a scripting language for compiling programs. It's a way to hide all the details of compiling your source (all those arguments you have to pass the compiler). You script all of the above details once in the Makefile, so you don't have to type it every time for every file. It will also do nifty things like only recompile source files that have been updated, and handle dependancies (if I recompile this file, I will then need to recompile THAT file.)
The biggest difference is that make is turing complete (Are makefiles Turing complete?) while gcc is not.
Let's take the gcc compiler for example.
It only knows how to compile the given .cpp file into .o file given the files needed for compilation to succeed (i.e. dependencies such as .h files).
However, those dependencies create a graph. e.g., b.o might require a.o in the compilation process which means it needs to be compiled independently beforehand.
Do you, as a programer want to keep track of all those dependencies and run them in order for your target .o file to build?
Of course not. You want something to do that task for you.
Those are build tools - tools that help making the build process (i.e. building the artifacts like .o files) easier. One such tool is make.
I hope that clarifies the difference :)

Resources