view libtests/libtest.mk @ 398:443c2939e84b api-inversion

off_t in ABI structures is a bad thing.

Why?  Because its size depends on the compile-time environment.  It was
OK, ish, when the only off_t was at the end of the struct, because then
we only stomped on uninitialized bits of memory; it is terrible when
there is more than one off_t kind of field.  Use uint64_t for those
fields instead.

EXECUTABLE=audioDB
SOVERSION=0
MINORVERSION=0
LIBRARY_FULL=lib$(EXECUTABLE).so.$(SOVERSION).$(MINORVERSION)
LIBRARY_VERS=lib$(EXECUTABLE).so.$(SOVERSION)
LIBRARY=lib$(EXECUTABLE).so
ARCH_FLAGS=


ifeq ($(shell uname),Darwin)
override LIBRARY_FULL=lib$(EXECUTABLE).$(SOVERSION).$(MINORVERSION).dylib
override LIBRARY_VERS=lib$(EXECUTABLE).$(SOVERSION).dylib
override LIBRARY=lib$(EXECUTABLE).dylib
ifeq ($(shell sysctl -n hw.optional.x86_64),1)
override ARCH_FLAGS+=-arch x86_64
endif
endif

all: $(LIBRARY_FULL) $(LIBRARY_VERS) $(LIBRARY) test1

$(LIBRARY_FULL):
	-ln -s ../../$(LIBRARY_FULL) $@

$(LIBRARY_VERS):
	-ln -s ../../$(LIBRARY_FULL) $@

$(LIBRARY): $(LIBRARY_VERS)
	-ln -s $< $@

test1: prog1.c ../test_utils_lib.h ../../audioDB_API.h
	gcc -g -std=c99 -Wall $(ARCH_FLAGS) -laudioDB -L. -Wl,-rpath,. -o $@ $<

clean:
	-rm $(LIBRARY_FULL) $(LIBRARY_VERS) $(LIBRARY)