I'm using GCC to generate a dependency file, but my build rules put the output into a subdirectory. Is there a way to tell GCC to put my subdirectory prefix in the dependency file it generates for me?
gcc $(INCLUDES) -E -MM $(CFLAGS) $(SRC) >>$(DEP)
I'm assuming you're using GNU Make and GCC. First add a variable to hold your list of dependency files. Assuming you already have one that lists all our sources:
SRCS = \
main.c \
foo.c \
stuff/bar.c
DEPS = $(SRCS:.c=.d)
Then include the generated dependencies in the makefile:
include $(DEPS)
Then add this pattern rule:
# automatically generate dependency rules
%.d : %.c
$(CC) $(CCFLAGS) -MF"$#" -MG -MM -MP -MT"$#" -MT"$(<:.c=.o)" "$<"
# -MF write the generated dependency rule to a file
# -MG assume missing headers will be generated and don't stop with an error
# -MM generate dependency rule for prerequisite, skipping system headers
# -MP add phony target for each header to prevent errors when header is missing
# -MT add a target to the generated dependency
"$#" is the target (the thing on the left side of the : ), "$<" is the prerequisite (the thing on the right side of the : ). The expression "$(<:.c=.o)" replaces the .c extension with .o.
The trick here is to generate the rule with two targets by adding -MT twice; this makes both the .o file and the .d file depend on the source file and its headers; that way the dependency file gets automatically regenerated whenever any of the corresponding .c or .h files are changed.
The -MG and -MP options keep make from freaking out if a header file is missing.
The answer is in the GCC manual: use the -MT flag.
-MT target
Change the target of the rule emitted by dependency generation. By default CPP takes the name of the main input file, deletes any directory components and any file suffix such as .c, and appends the platform's usual object suffix. The result is the target.
An -MT option will set the target to be exactly the string you specify. If you want multiple targets, you can specify them as a single argument to -MT, or use multiple -MT options.
For example, -MT '$(objpfx)foo.o' might give
$(objpfx)foo.o: foo.c
You may like this briefer version of Don McCaughey's answer:
SRCS = \
main.c \
foo.c \
stuff/bar.c
DEPS = $(SRCS:.c=.d)
Add -include $(DEPS) note the - prefix, which silences errors if the .d files don't yet exist.
There's no need for a separate pattern rule to generate the dependency files. Simply add -MD or -MMD to your normal compilation line, and the .d files get generated at the same time your source files are compiled. For example:
%.o: %.c
gcc $(INCLUDE) -MMD -c $< -o $#
# -MD can be used to generate a dependency output file as a side-effect of the compilation process.
Detailing on DGentry's answer, this has worked well for me:
.depend: $(SOURCES)
$(CC) $(CFLAGS) -MM $(SOURCES) | sed 's|[a-zA-Z0-9_-]*\.o|$(OBJDIR)/&|' > ./.depend
This also works in the case where there is only one dependency file that contains the dependency rules for all source files.
Ok, just to make sure I've got the question right: I'm assuming you have test.c which includes test.h, and you want to generate subdir/test.d (while not generating subdir/test.o) where subdir/test.d contains
subdir/test.o: test.c test.h
rather than
test.o: test.c test.h
which is what you get right now. Is that right?
I was not able to come up with an easy way to do exactly what you're asking for. However, looking at Dependency Generation Improvements, if you want to create the .d file while you generate the .o file, you can use:
gcc $(INCLUDES) -MMD $(CFLAGS) $(SRC) -o $(SUBDIR)/$(OBJ)
(Given SRC=test.c, SUBDIR=subdir, and OBJ=test.o.) This will create both subdir/test.o and subdir/test.d, where subdir/test.d contains the desired output as above.
If there is an argument to GCC to do this, I don't know what it is. We end up piping the dependency output through sed to rewrite all occurrences of <blah>.o as ${OBJDIR}/<blah>.o.
[GNU] make gets angry if you don't place the output in the current directory. You should really run make from the build directory, and use the VPATH make variable to locate the source code. If you lie to a compiler, sooner or later it will take its revenge.
If you insist on generating your objects and dependencies in some other directory, you need to use the -o argument, as answered by Emile.
Related
I am new to C++ and am trying to get the hang of build systems like make/CMake. Coming from Go, it seems that there is a constant risk that if you forget to do a little thing, your binaries will become stale. In particular, I can't find a best practice for remembering to keep dependencies/prerequisites updated in make/CMake. I'm hoping I am missing something obvious.
For example, suppose I have a basic makefile that just compiles main.cpp:
CFLAGS = -stdlib=libc++ -std=c++17
main: main.o
clang++ $(CFLAGS) main.o -o main
main.o: main.cpp
clang++ $(CFLAGS) -c main.cpp -o main.o
main.cpp:
#include <iostream>
int main() {
std::cout << "Hello, world\n";
}
So far so good; make works as expected. But suppose I have some other header-only library called cow.cpp:
#include <iostream>
namespace cow {
void moo() {
std::cout << "Moo!\n";
}
}
And I decide to call moo() from within main.cpp via `include "cow.cpp":
#include <iostream>
#include "cow.cpp"
int main() {
std::cout << "Hello, world\n";
cow::moo();
}
However, I forget to update the dependencies for main.o in makefile. This mistake is not revealed during the obvious testing period of running make and rerunning the binary ./main, because the whole cow.cpp library is directly included in main.cpp. So everything seems fine, and Moo! is printed out as expected.
But when I change cow.cpp to print Bark! instead of Moo!, then running make doesn't do anything and now my ./main binary is out of date, and Moo! is still printed from ./main.
I'm very curious to hear how experienced C++ devs avoid this problem with much more complicated codebases. Perhaps if you force yourself to split every file into a header and an implementation file, you'll at least be able to quickly correct all such errors? This doesn't seem bulletproof either; since header files sometimes contain some inline implementations.
My example uses make instead of CMake, but it looks like CMake has the same dependency listing problem in target_link_libraries (though transitivity helps a bit).
As a related question: it seems like the obvious solution is for the build system to just look at the source files and infer dependencies (it can just go one level in and rely on CMake to handle transitivity). Is there a reason this doesn't work? Is there a build system that actually does this, or should I write my own?
Thanks!
First of all you will need to reference the dependencies file in your Makefile.
This can be done with the function
SOURCES := $(wildcard *.cpp)
DEPENDS := $(patsubst %.cpp,%.d,$(SOURCES))
wich will take the name of all *.cpp files and substitute and append the extension *.d to name your dependency.
Then in your code
-include $(DEPENDS)
- tells the Makefile to not complain if the files do not exist. If they exist they will be included and recompile your sources properly according to the dependencies.
Finally the dependencies can be created automatically with the options: -MMD -MP for the rules to create the objects file. Here you can find a complete explanation. What generates the dependencies is MMD; MP is to avoid some errors. If you want to recompile when system libraries are updated use MD instead of MMD.
In your case you can try:
main.o: main.cpp
clang++ $(CFLAGS) -MMD -MP -c main.cpp -o main.o
If you have more files it is better to have a single rule to create object files. Something like:
%.o: %.cpp Makefile
clang++ $(CFLAGS) -MMD -MP -c $< -o $#
You can take a look also at this 2 great answers:
one
two
In your case a more suitable Makefile should look like the following (there might be some errors but let me know):
CXX = clang++
CXXFLAGS = -stdlib=libc++ -std=c++17
WARNING := -Wall -Wextra
PROJDIR := .
SOURCEDIR := $(PROJDIR)/
SOURCES := $(wildcard $(SOURCEDIR)/*.cpp)
OBJDIR := $(PROJDIR)/
OBJECTS := $(patsubst $(SOURCEDIR)/%.cpp,$(OBJDIR)/%.o,$(SOURCES))
DEPENDS := $(patsubst $(SOURCEDIR)/%.cpp,$(OBJDIR)/%.d,$(SOURCES))
# .PHONY means these rules get executed even if
# files of those names exist.
.PHONY: all clean
all: main
clean:
$(RM) $(OBJECTS) $(DEPENDS) main
# Linking the executable from the object files
main: $(OBJECTS)
$(CXX) $(WARNING) $(CXXFLAGS) $^ -o $#
#include your dependencies
-include $(DEPENDS)
#create OBJDIR if not existin (you should not need this)
$(OBJDIR):
mkdir -p $(OBJDIR)
$(OBJDIR)/%.o: $(SOURCEDIR)/%.cpp Makefile | $(OBJDIR)
$(CXX) $(WARNING) $(CXXFLAGS) -MMD -MP -c $< -o $#
EDIT to answer comments
As another question, is there any problem with rewriting the DEPENDS definition as just DEPENDS := $(wildcard $(OBJDIR)/*.d)?
Nice question, it took me a while to see your point
From here
$(wildcard pattern…) This string, used anywhere in a makefile, is
replaced by a space-separated list of names of existing files that
match one of the given file name patterns. If no existing file name
matches a pattern, then that pattern is omitted from the output of the
wildcard function.
So wildcard return a list of the file names matching the pattern. patsubst acts on strings, it does not care about what are those strings: it is used as a way to create the file names of the dependencies, not the files themselves. In the Makefile example that I posted DEPENDS is actually use in two cases: when cleaning with make clean and with include so in this case they both work because you are not using DEPENDS in any rule. There are some differences (I tried to run and you should too to confirm). With DEPENDS := $(patsubst $(SOURCEDIR)/%.cpp,$(OBJDIR)/%.d,$(SOURCES)) if you run make clean dependencies *.d that do not have a correspondent *.cpp file will not be removed while they will with your change. On the contrary you might include dependencies not relevant to your *.cpp file.
I asked this questions: let's see the answers.
If the .d files get deleted from carelessness but the .o files remain, then we are in trouble. In the original example, if main.d is deleted and then cow.cpp is subsequently changed, make won't realize it needs to recompile main.o and thus it will never recreate the dependency file. Is there a way to cheaply create the .d files without recompiling the object files? If so then we could probably recreate all the /.d files on every make command?
Nice question again.
Yes, you are right. Actually it was an error of mine. This happens because of the rule
main: main.o
$(CXX) $(WARNING) $(CFLAGS) main.o -o main
actually should have been:
main: $(OBJECTS)
$(CXX) $(WARNING) $(CXXFLAGS) $^ -o $#
so that it got relinked (the executable is updated) whenever one of the objects change and they will change whenever one their cpp file change.
One problem remains: if you delete your dependencies but not the objects, and change only one or more header files (but not the sources) then your program is not updated.
I corrected also the previous part of the answer.
EDIT 2
To create the dependencies you can also add a new rule to your Makefile:
here is an example.
I am working on a makefile for a C++ project that needs to support a few configurations, i.e. debug , release and maybe a few more customized ones in the future.
Currently, my naming convention for generated .o files is $(SOURCE_FULLPATH).$(CONFIGURATION).o. For instance, ABC.cpp generates ABC.cpp.debug.o in debug mode.
Now I would like to write the pattern rule for generating those object files in a configuration-independent way. What I did was: from each XX.o filename, I strip the .debug or .release suffix from XX, and use the remaining part of XX as the source filename.
%.o: $$(basename %)
$(CC) $(CC_FLAGS) $(INCLUDE_FOLDERS) -c -o $# $<
With this trick, I can build the executable correctly, except that I get one warning message from make:
make: Circular makefile.o <- makefile dependency dropped.
I am puzzled because I do not list makefile or makefile.o as a target or dependency anywhere in my makefile. I did a search on SO, but most questions about Circular dependency is on a specific user source file, rather than the makefile itself. Can anyone help me understand what causes the circular dependency, and how to get rid of this warning message?
A sample makefile that can reproduce this issue is listed below.
.SECONDEXPANSION:
PROJECT := helloworld
CC := clang++
BUILD_FOLDER := Build
OBJ_FILE_SUFFIX := .o
# Source
CPP_FILES :=\
Source/hello.cpp \
Source/mysqrt.cpp \
INCLUDE_FOLDERS := \
-IInclude
# MMD outputs the dependency files (".d" files). These files will be used by
# this makefile to allow for dependency checking on .h files.
CC_FLAGS += -MMD
EXISTING_OBJ_FILES = $(wildcard $(addsuffix *.o, $(basename $(CPP_FILES))))
##--------------------
## Targets definition
##--------------------
.PHONY:default
default: all
.PHONY:all
all: debug release
.PHONY:debug release
# Add a 'debug'/'release' suffix to the name of the object file
# e.g. hello.cpp -> hello.cpp.debug.o
debug release: OBJ_FILES=$(addsuffix .$#$(OBJ_FILE_SUFFIX), $(CPP_FILES))
debug release: $${OBJ_FILES} # Use Secondary Expansion to get the obj names
$(CC) $^ -o $(BUILD_FOLDER)/$(PROJECT)_$#
# Strip configuration name from the end of the object file name
%.o: $$(basename %)
$(CC) $(CC_FLAGS) $(INCLUDE_FOLDERS) -c -o $# $<
## clean: remove executable, all object files, and all dependency files
.PHONY:clean
clean:
-rm -f $(BUILD_FOLDER)/$(PROJECT) $(EXISTING_OBJ_FILES) $(EXISTING_OBJ_FILES:.o=.d)
# Include the dependent files so that in later builds, modified .h files
# will cause all .cpp dependent on them to rebuild
-include $(OBJ_FILES:.o=.d)
The folder structure is
makefile
Source
- hello.cpp
- mysqrt.cpp
Include
- mysqrt.h
The full output of make debug is
make: Circular makefile.o <- makefile dependency dropped.
clang++ -MMD -IInclude -c -o Source/hello.cpp.debug.o Source/hello.cpp
clang++ -MMD -IInclude -c -o Source/mysqrt.cpp.debug.o Source/mysqrt.cpp
clang++ Source/hello.cpp.debug.o Source/mysqrt.cpp.debug.o -o Build/helloworld_debug
Everything is good except for the first line.
I would also really appreciate it if anyone can point to me if there is any bad practice in my makefile (I am still a newbie in makefile). Thank you in advance!
GNU Make always attempts to update the makefile(s) it has read before
making anything else. If it finds rules and prerequisites that tell it
to update makefile(s), then it does so and then starts again from scratch -
including attempting to update the makefile(s). See 3.5 How Makefiles Are Remade.
In your recipe:
%.o: $$(basename %)
$(CC) $(CC_FLAGS) $(INCLUDE_FOLDERS) -c -o $# $<
you have provided make with a rule for making makefile.o from makefile.
It is also the inverse of the rule in the builtin recipe
%: %.o
$(LINK.o) $^ $(LOADLIBES) $(LDLIBS) -o $#
which makes an executable from a single object file. So your recipe has introduced the circularity:
makefile.o <- makefile <- makefile.o
when make is considering makefile itself as a target.
You could suppress the circularity by expressly deleting the builtin inverse rule,
by writing the empty rule:
%: %.o
in the makefile. Then you could observe the following confusion on the part of the
compiler:
$ make makefile.o
clang++ -c -o makefile.o makefile
clang: warning: makefile: 'linker' input unused
And the same would occur if you attempted to make any target that depended
on makefile.o.
It is probably safe to assume that you will have no targets that depend on
makefile.o. Nevertheless a rule that would attempt to
compile foo.o from any existing file foo is clearly more sweeping that you
want or need. For the particular pattern of dependency that you wish to capture:
foo.cpp.{debug|release}.o: foo.cpp
You'd be better off with:
%.o: $$(basename $$(basename %)).cpp
$(CC) $(CC_FLAGS) $(INCLUDE_FOLDERS) -c -o $# $<
Note, BTW, that in GNU Make conventions - the conventions that are
assumed by GNU Make's builtin rules - CC denotes your C compiler while
CXX denotes your C++ compiler. Likewise flags for the C compiler are
denoted CFLAGS and flags for the C++ compiler are denoted CXXFLAGS.
Flags for the preprocessor are denoted CPPFLAGS, and -Ipath options
- which are preprocessor options - are conventionally be passed through CPPFLAGS.
Does anyone have a complete makefile that can do the following:
Rebuilds the project if a HEADER file changes
The cpp files are listed in the makefile
The header files are NOT listed in the makefile
The header files are allowed to have different names than the cpp files
Some of the cpp files do not have header files
I have seen instructions for figuring out how to make the make tool figure out header dependencies, but none of these instructions provide anything remotely resembling a working example. A simple example would be as follows: main.cpp C1.cpp C1.h C2.cpp C2.h
CXX = g++
OBJECTS := main.o C1.o C2.o
all: $(OBJECTS)
%.o : %.cpp
$(CXX) $(CPPFLAGS) -Wall -MMD -c $< -o $#
-include *.d
EDIT: As TobySpeight points out, this won't work if you build an object file, rename or delete one of the prerequisite source or header files, then try to rebuild the object file; the .d file will still require the missing file, and the build will fail. I neglected to include lines to deal with that case:
%.h: ;
%.cpp: ;
(This is effective, but crude. The more precise approach is to put some sed commands in the %.o rule, so as to add specific null rules to the .d file, one for each prerequisite, but the sed commands are ugly, and the approach above is good enough for almost all cases.)
You can also use CMake for this. Everything you need to write is:
add_executable (exec main.cpp C1.cpp C2.cpp)
I have couple of cpp and hpp files in directory ./src. I compile all cpp files in one binary file, say ./bin/run. I want to re-compile only if I need i.e it or one of its header was changed.
I, probably, can create Makefile where file will be recompiled if and only if it was changed, but it's quite uncomfortable because big part of my code is in the headers. (It's not going to be changed, because the product is header itself and cpp files are tests).
I want to store temporary .o files in ./build
I know about g++ -MM function but I'm not sure how to use it.
I'll glad to see solutions that use not necessary make but any other system possible if they are easy enough.
UPD
I'll try to clarify, what's the problem is:
New cpp's maybe created, includes may be added or gone, etc. I don't want to edit my makefile each time.
To solve the problem I mentioned (-include is not a good solution), I use something like this:
build/%.o: %.cpp
#$(CC) -MD -c -Wall -o $# $<
#cp build/$*.d build/$*.P
#sed -e 's/#.*//' -e 's/^[^:]*: *//' -e 's/ *\\$$//' \
-e '/^$$/ d' -e 's/$$/ :/' < build/$*.P >> build/$*.d
#rm build/$*.P
-include build/*.d
No `%.d rule is needed.
EDIT:
#JackKelly has [*cough*, *choke*] shown me a better way to get effectively the same dependency file:
build/%.o: %.cpp
#$(CC) -MD -c -Wall -o $# $<
#$(CC) -MM -MP -Wall -o $*.d $<
-include build/*.d
Ye, you can have multiple rules for the same target, as long as only one of them has commands; the prerequisites accumulate. The idea is to get a file like this:
file.o: file.cpp headerfile.h
headerfile.h:
The second line (headerfile.h:) is a rule for headerfile.h that has no prerequisites or commands. It does nothing, but it's a rule, so if headerfile.h is missing, Make is satisfied.
You mention g++ -MM, which can do what you're trying to do:
include $(ALLOBJ:%.o=%.d)
%.d: %.cxx
#echo making dependencies for $<
#g++ -MM -MP $(CXXFLAGS) $< -o $#
#sed -i 's,$*\.o,& $# ,g' $#
Basically this defines a rule that creates .d files from .cxx files. The .d files are, in turn, required by the include statement, which requires one for each .o file in ALLOBJ.
The last line in the dependency rule is the 'sed magic' that makes the dependency files regenerate themselves. If you think regular expressions are hacks at best, and evil more often than not, you can use the -MT flag.
The outline of a solution is as follows:
Use auxiliary dependency files for each source file (that is, create foo.dep for foo.c, bar.dep for bar.c etc)
Use gcc -MM to create the dependency files
In order to force make to do this automatically, use foo.c as a prerequisite for foo.dep and foo.o; this requires some minor sed magic on the output of gcc -MM
Include all the dependency files in your main makefile; this is a key step that makes this approach possible.
The last step is written as follows:
-include $(dependency_files)
This is very tricky but possible; see the GNU make manual for more information.
You can do this with Make, you just need to specify the headers in your rule's sensitivity list. For example
myfile.o: myfile.cpp
gcc -c myfile.o myfile.cpp ${LDFLAGS} # This is optional. make can infer this line.
Turns into
myfile.o: myfile.cpp myfile.h
gcc -c myfile.o myfile.cpp ${LDFLAGS} # Again, optional.
Now, whenever myfile.h changes, myfile.cpp will be rebuild. More headers can be chained in a similar way.
I'm currently trying to build a proper Makefile.
What I want is full control of what's happening, so I don't want any third party software.
My current attempt seems logic to me, but since the dependency generation is not valid, I'm kind of stuck.
For better readabilty, the full Makefile is broken into little pieces. I would appreciate any comment on any section if there's something to improve.
First of all, I have the following static definitions
CXX = g++
CXXFLAGS = -Wall \
-Wextra \
-Wuninitialized \
-Wmissing-declarations \
-pedantic \
-O3 \
-p -g -pg
LDFLAGS = -p -g -pg
DEPFLAGS = -MM
Afaik this should be fine. It would be perfect to make the profiling flags optional but that's not important.
SRC_DIR = ./src
OBJ_DIR = ./obj
SRC_EXT = .cpp
OBJ_EXT = .o
TARGET = ./bin/my_target
SRCS = $(wildcard $(SRC_DIR)/*$(SRC_EXT))
OBJS = $(subst $(SRC_DIR), $(OBJ_DIR), $(SRCS:$(SRC_EXT)=$(OBJ_EXT)))
DEP = depend.main
Basically, this should just extract all the *.cpp files out of the subfolder src and additionally replace ./src with ./obj and .cpp with .o as names of the objects.
.PHONY: clean all depend
all: $(TARGET)
$(TARGET): $(OBJS)
#echo "-> linking $#"
#$(CXX) $^ $(LDFLAGS) -o $#
$(OBJ_DIR)/%.$(EXT_OBJ):
#echo "-> compiling $#"
#$(CXX) $(CXXFLAGS) -c $< -o $#
Afaik, this block - provided a valid dependency file exists - should do all necessary compiling and linking.
clean:
#echo "removing objects and main file"
#rm -f $(OBJS) $(TARGET)
Should be self-explanatory and correct, or am I missing something here?
$(SRC_DIR)/%.$(SRC_EXT):
$(CXX) $(DEPFLAGS) -MT \
"$(subst $(SRC_DIR),$(OBJ_DIR),$(subst $(SRC_EXT),$(OBJ_EXT),$#))" \
$(addprefix ,$#) >> $(DEP);
clear_dependencies:
#echo "-> (re-)building dependencies";
#$(RM) $(DEP)
depend: clear_dependencies $(SRCS)
This is the non-functional part. What I intend to do is using the g++ Compiler flag -MM to auto-create dependencies and using -MT to use a different path than the default one.
The resulting dependency should look like
./obj/main.o: ./src/main.cpp ./src/some_header_file.h
Unfortunately, this will never be called and I lack the knowledge why this is the case.
In a similar question, user Beta gladly provided a temporary solution by adding a .Phony but this has the side effect on rebuilding every object without any change.
Finally, there is just the one line
-include $(DEP)
to include the dependency file, once created.
Any answer providing some hints about any part are very welcome.
So my question is: What can I do better or maybe "cleaner" and why doesn't the dependency generation work?
Here goes.
Assign simply expanded variables where possible:
SRCS := $(wildcard $(SRC_DIR)/*$(SRC_EXT))
From GNU Make manual:
Another disadvantage [of recursively expanded variables] is that any functions referenced in the definition will be executed every time the variable is expanded. This makes make run slower; worse, it causes the wildcard and shell functions to give unpredictable results because you cannot easily control when they are called, or even how many times.
Use substitution references or patsubst function to convert sources into objects:
OBJS := $(SRCS:$(SRC_DIR)/%$(SRC_EXT)=$(OBJ_DIR)/%$(OBJ_EXT))
Specify proper prerequisites in compilation pattern rule. This is mandatory to get Make keeping your object files up to date and updating them on source changes.
$(OBJ_DIR)/%$(OBJ_EXT) : $(SRC_DIR)/%$(SRC_EXT)
#echo "-> compiling $#"
#$(CXX) $(CXXFLAGS) -o $# -c $<
Compile sources and generate dependency files for them at the same time. Use -MMD -MP flags to get things work (just append them to CXXFLAGS).
CXXFLAGS += -MMD -MP
-include $(OBJS:$(OBJ_EXT)=.d)
From GCC manual:
-MD
-MD is equivalent to -M -MF file, except that -E is not implied. The driver determines file based on whether an -o option is given. If it is, the driver uses its argument but with a suffix of .d, otherwise it takes the name of the input file, removes any directory components and suffix, and applies a .d suffix.
-MMD
Like -MD except mention only user header files, not system header files.
-MP
This option instructs CPP to add a phony target for each dependency other than the main file, causing each to depend on nothing. These dummy rules work around errors make gives if you remove header files without updating the Makefile to match.
Also consider studying this article of Paul Smith (he is a maintainer of GNU Make). It gives a rather good overview of different autodep-generation approaches.