# for it to operate as expected. When included the module generates
# the requested rules based on the contents of its control variables.
-# This works out very well and allows a good degree of flexability.
-# To accomidate some of the features we introduce the concept of
+# This works out very well and allows a good degree of flexibility.
+# To accommodate some of the features we introduce the concept of
# local variables. To do this we use the 'Computed Names' feature of
# gmake. Each module declares a LOCAL scope and access it with,
# $($(LOCAL)-VAR)
-# This works very well but it is important to rembember that within
-# a rule the LOCAL var is unavailble, it will have to be constructed
-# from the information in the rule invokation. For stock rules like
+# This works very well but it is important to remember that within
+# a rule the LOCAL var is unavailable, it will have to be constructed
+# from the information in the rule invocation. For stock rules like
# clean this is simple, we use a local clean rule called clean/$(LOCAL)
# and then within the rule $(@F) gets back $(LOCAL)! Other rules will
# have to use some other mechanism (filter perhaps?) The reason such
# lengths are used is so that each directory can contain several 'instances'
# of any given module. I notice that the very latest gmake has the concept
# of local variables for rules. It is possible this feature in conjunction
-# with the generated names will provide a very powerfull solution indeed!
+# with the generated names will provide a very powerful solution indeed!
# A build directory is used by default, all generated items get put into
# there. However unlike automake this is not done with a VPATH build
# (vpath builds break the distinction between #include "" and #include <>)
-# but by explicly setting the BUILD variable. Make is invoked from
+# but by explicitly setting the BUILD variable. Make is invoked from
# within the source itself which is much more compatible with compilation
# environments.
ifndef NOISY
error-out-and-die
else
error-all/environment.mak:
- echo Can't find the build directory in $(BUILD_POSSIBLE) -- use BUILD=
+ echo Can not find the build directory in $(BUILD_POSSIBLE) -- use BUILD=
error-out-and-die
endif
OBJ := $(BUILD)/obj/$(SUBDIR)
DEP := $(OBJ)
DOC := $(BUILD)/docs
+PO := $(BUILD)/po
+LOCALE := $(BUILD)/locale
+PO_DOMAINS := $(BUILD)/po/domains
# Module types
LIBRARY_H = $(BASE)/buildlib/library.mak
DEBIANDOC_H = $(BASE)/buildlib/debiandoc.mak
MANPAGE_H = $(BASE)/buildlib/manpage.mak
PROGRAM_H = $(BASE)/buildlib/program.mak
+PYTHON_H = $(BASE)/buildlib/python.mak
COPY_H = $(BASE)/buildlib/copy.mak
-YODL_MANPAGE_H = $(BASE)/buildlib/yodl_manpage.mak
+PO4A_MANPAGE_H = $(BASE)/buildlib/po4a_manpage.mak
+FAIL_H = $(BASE)/buildlib/fail.mak
+PODOMAIN_H = $(BASE)/buildlib/podomain.mak
+
+include $(BUILD)/environment.mak
ifdef STATICLIBS
LIBRARY_H += $(BASE)/buildlib/staticlibrary.mak
# Source location control
# SUBDIRS specifies sub components of the module that
-# may be located in subdrictories of the source dir.
+# may be located in subdirectories of the source dir.
# This should be declared before including this file
SUBDIRS+=
# Header file control.
-# TARGETDIRS indicitates all of the locations that public headers
+# TARGETDIRS indicates all of the locations that public headers
# will be published to.
# This should be declared before including this file
HEADER_TARGETDIRS+=
# Options
-include $(BUILD)/environment.mak
CPPFLAGS+= -I$(INCLUDE)
LDFLAGS+= -L$(LIB)
+# Directors to create
+MKDIRS := $(BIN)
+
# Phony rules. Other things hook these by appending to the dependency
# list
-.PHONY: headers library clean veryclean all binary program doc
+.PHONY: headers library clean veryclean all binary program doc dirs
.PHONY: maintainer-clean dist-clean distclean pristine sanity
all: binary doc
binary: library program
echo Very Clean done for $(SUBDIR)
clean:
echo Clean done for $(SUBDIR)
-
+dirs:
+ mkdir -p $(patsubst %/,%,$(sort $(MKDIRS)))
+
# Header file control. We want all published interface headers to go
-# into the build directory from thier source dirs. We setup some
+# into the build directory from their source dirs. We setup some
# search paths here
vpath %.h $(SUBDIRS)
$(INCLUDE)/%.h $(addprefix $(INCLUDE)/,$(addsuffix /%.h,$(HEADER_TARGETDIRS))) : %.h
# Dependency generation. We want to generate a .d file using gnu cpp.
# For GNU systems the compiler can spit out a .d file while it is compiling,
# this is specified with the INLINEDEPFLAG. Other systems might have a
-# makedep program that can be called after compiling, that's illistrated
+# makedep program that can be called after compiling, that's illustrated
# by the DEPFLAG case.
# Compile rules are expected to call this macro after calling the compiler
+ifdef GCC3DEP
+DFILE = $(DEP)/$(basename $(@F)).d
+else
+DFILE = $(basename $(@F)).d
+endif
ifdef INLINEDEPFLAG
define DoDep
- sed -e "1s/.*:/$(subst /,\\/,$@):/" $(basename $(@F)).d > $(DEP)/$(@F).d
+ sed -e "1s/.*:/$(subst /,\\/,$@):/" $(DFILE) > $(DEP)/$(@F).d
+ #sed -e "1s/.*:/$(subst /,\\/,$@):/" $(DEP)/$(basename $(@F)).d > $(DEP)/$(@F).d
-rm -f $(basename $(@F)).d
endef
else
PARALLEL_RUN=no
endif
-ifndef PARALLEL_RUN
- PARALLEL_RUN=yes
- .EXPORT: PARALLEL_RUN
- # handle recursion
- ifneq ($(NUM_PROCS),)
- MAKEFLAGS += -j $(NUM_PROCS)
- endif
-endif
+# mvo: commented out, lead to build failures in the arch-build target
+#ifndef PARALLEL_RUN
+# PARALLEL_RUN=yes
+# .EXPORT: PARALLEL_RUN
+# # handle recursion
+# ifneq ($(NUM_PROCS),)
+# MAKEFLAGS += -j $(NUM_PROCS)
+# endif
+#endif
projects / apt.git / blobdiff