CVS log for src/share/mk/bsd.dep.mk

Up to [cvs.NetBSD.org] / src / share / mk

Request diff between arbitrary revisions

Keyword substitution: kv
Default branch: MAIN

bsd.dep.mk: move MKDEPINCLUDES to bsd.own.mk

Add default for MKDEPINCLUDES=no to to bsd.own.mk _MKVARS.no.
Simplify bsd.dep.mk check.

Pass W with IDU for lint and deps.

bsd.dep.mk: fix "make tags" (again)

[repeat revision 1.85]

Fix "make tags" to actually build a tags file:
- Use !commands() instead of !target(), so that the rule actually works
- Write to ${.OBJDIR}/tags for read-only source (don't know why ${.TARGET}
  isn't sufficient).
- Only match *.[cly] from ${.ALLSRCS} - just excluding *.h causes failures
  because of ${targ}: subdir-${targ} in bsd.subdir.mk.

Thanks to uwe@ for assistance.

bsd.dep.mk: revert 1.85 (for now)

Revert my recent 1.85 revision that fixed "make tags".  It causes too
much build breakage elsewhere in the tree that needs to be resolved first.

Issues include:
- Directories using TESTS_CXX with .cpp and .cxx extension instead of the
  default .cc extension (see bsd.prog.mk). Most of these have been fixed.
- external/gpl3/gcc build of .cc files. (No idea what's wrong there).

bsd.dep.mk: fix "make tags"

Fix "make tags" to actually build a tags file:
- Use !commands() instead of !target(), so that the rule actually works
- Write to ${.OBJDIR}/tags for read-only source (don't know why ${.TARGET}
  isn't sufficient).
- Only match *.[cly] from ${.ALLSRCS} - just excluding *.h causes failures
  because of ${targ}: subdir-${targ} in bsd.subdir.mk.

Thanks to uwe@ for assistance.

Sync with HEAD

Sync with HEAD

Most of the mv operations are to move temporary files to their final place.
Some use -f, others don't. This can lead to spurious build failures when
the user performing the build changes. Centralize, and always use -f.

Sync with HEAD.  (Note that most of these changes are simply $NetBSD$
tag issues.)

Don't remap mkdep, otherwise we add dependencies for /usr/include instead
of ${DESTDIR}/usr/include and things go south.

PR toolchain/9618: add ".d" to MKDEP_SUFFIXES so the depend files are rebuilt
if any of their dependencies (i.e. the files they refer to) are changed.
This allows changes to transitive included headers to be handled properly.

Rebase to HEAD as of a few days ago.

Rebase.

when calling mkdep, pass any -std=* flags for C and C++ source files
to ensure that any definitions this implies are set

sync with head.

for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.

this commit was splitted into small chunks to avoid
a limitation of cvs.  ("Protocol error: too many arguments")

add ${MKTARGET_CREATE} for tags

create mkdep output with a temporary file so that failures in
complation during .d creation don't leave an empty (stale) .d
file around to break future builds.

(hmm, probably could use .PRECIOUS here.)

add MKDEPCXX, that is used for C++ files depend creation.  also
add some missing ${COPTS.${.IMPSRC:T}} and ${CPUFLAGS.${.IMPSRC:T}}
to the dependancy rules.

this makes it possible to have C++ files .d files created with the
C++ compiler, and with the normal options for this file, which
means that we can build libstdc++-v3 from gcc 4.8.

resync from head

Add a rule to rebuild dependency files that contain stale entries automatically.

MKDEPINCLUDES support.

sync with (a bit old) head

Resync to 2012-11-19 00:00:00 UTC

Change some .if defined(SRCS) to .if defined(SRCS) && !empty(SRCS)

sync with head

Add support for --combine with gcc.

Move all "clean" and "cleandir" processing to a new bsd.clean.mk
file, as proposed in tech-toolchain.

For USETOOLS=yes, use the --sysroot support of the compiler to cut down
the number of hard-coded pathes and magic invocations.

Sync with HEAD

Allow HOSTLIB to use C++ by not specifying -nostdinc++.

Allow use of traditional CPP to be set on a per platform base in sys.mk.
Honour this for dependency processing in bsd.dep.mk.  Switch i386 and
amd64 assembly to use ISO C90 preprocessor concat and drop the
-traditional-cpp on this platform.

PR/42658: Geoff Wing: Quote directory variables to deal with spaces in the
source pathname.

Use ${TOOL_SED} instead if plain sed in Makefiles.

remove AINC support; it is now unnecessary

Cosmetically tweak how .PHONY is used by (generally) providing it as a source,
instead of using it as a target with all the rules as its sources.

Replace
	:M-[IDU]*
with
	:C/-([IDU])[  ]*/-\1/Wg:M-[IDU]*
so that arguments with whitespace after them work correctly.
Resolves [toolchain/18248] from Andrew White at Motorola.

Be consisent in this syntax; previously we'd sometimes we'd just have :M-[ID]*
which would ignore -Ufoo ...

(This needs :C///W support in make(1) that Simon Gerraty added for me :)

Use -- to separate mkdep options from cc ones.
Avoids problems when the cc options start with -fpic.

Use modified mkdep(1) (not cat) to merge '*.d' into '.depend'.
Allows some of the postprocessing to be done there instead of a separate
sed script called from 'afterdepend'.

fix a typo in previous pointed out by lukem

allow -traditional-cpp to be overridden when making dependancies

Rework how MAKEVERBOSE operates:

    *	Don't bother prefixing commands with a line of  ${_MKCMD}\
	and instead rely upon "make -s".  This is less intrusive on
	all the Makefiles than the former.  Idea from David Laight.

    *	Rename the variables use to print messages.  The scheme now is:
	    _MKMSG_FOO		Run  _MKMSG 'foo'
	    _MKTARGET_FOO	Run  _MKMSG_FOO ${.TARGET}
	From discussion with Alistair Crooks.

Provide _MKMSG{BUILD,CREATE,COMPILE,FORMAT,INSTALL,LINK,LEX,YACC} to display
${_MKMSG} "  ...  {$.TARGET}"  and use appropriately.

Provide _MKMSGLINK.m to provide "link" message by itself (without ${.TARGET})

Replace _MKSHCMD with _MKSHECHO

Implement MAKEVERBOSE to control how "verbose" the standard make(1) rules are.
Supported values:
    0	Minimal output ("quiet")
    1	Describe what is occurring
    2	Describe what is occurring and echo the actual command ("verbose")
	This is the default, and is equivalent to the previous behaviour.

order .SUFFIXES similar to other *.mk files

Handle empty __DPSRCS.d, avoiding executing commands like "cat > .depend"
in directories where only non-C/C++ files are in SRCS (e.g.: the forth
code for sparc bootblocks).

Patch from Luke.

Rework how dependency generation is performed:

* DPSRCS contains extra dependencies, but is _NOT_ added to CLEANFILES.
  This is a change of behaviour.  If a Makefile wants the clean semantics
  it must specifically append to CLEANFILES.
  Resolves PR toolchain/5204.

* To recap: .d (depend) files are generated for all files in SRCS and DPSRCS
  that have a suffix of: .c .m .s .S .C .cc .cpp .cxx

* If YHEADER is set, automatically add the .y->.h to DPSRCS & CLEANFILES

* Ensure that ${OBJS} ${POBJS} ${LOBJS} ${SOBJS} *.d  depend upon ${DPSRCS}

* Deprecate the (short lived) DEPENDSRCS


Update the various Makefiles to these new semantics; generally either
adding to CLEANFILES (because DPSRCS doesn't do that anymore), or replacing
specific .o dependencies with DPSRCS entries.

Tested with "make -j 8 distribution" and "make distribution".

tweak ${DEPENDSRCS.d} target so that ${SRCS} isn't there.  this fixes the
"make: out of memory" problem when make dependencies in lib/libc.

XXX: this is a temporary workaround.  i'm working on cleaning up how
the DPSRCS stuff is used throughout the tree...

bsd.dep.mk:  add a dependency for *.d on ${SRCS} ${DPSRCS}.  this should fix
some build problems where DPSRCS contains .h files that need to exist to
build the .d file.

document that bsd.dep.mk should be included after anything that might
update SRCS or DPSRCS so that the above works correctly, and actually
ensure this is the case.

Set .NOPATH for all ${DEPENDSRCS} and not just .depend.
Should fix build problem noted by Geoff Wing on current-users@

Use the more "standard" `foo.d' instead of `foo.SUFFIX.dep'; the rules
are simpler and would have helped prevent the previous fubar.
Don't bother using -a with ${MKDEP} to build the individual .d files.

Make all the suffix rules with .c .c.dep (e.g not .cc .cc.cc.dep)

Simplify .*.dep generation by using suffix transforms.
This also fixes a problem in the previous where the .ALLSRC would contain
any other dependencies that .depend had, thus breaking the mkdep invocation.

Provide DEPENDSRCS which effectively contains ".depend *.dep", to be
used in other Makefiles that want to add dependencies for .depend.

Always use the "per source file".dep method to build the .depend file
(that was formerly only used for HOST_CYGWIN.).
Clean up implementation to minimise unnecessary variable munging.
Ensure all .dep files are deleted in cleandepend:.

This allows us to correctly support CPPFLAGS.<file> in make depend, and
whilst this method is marginally slower in the "no existing .depend file"
case, it should be faster in the "rebuilding depend for a subset of
source files" case.

introduce a new $CPPFLAG_ISYSTEMXX that is used to reference c++ system
includes.  for GCC3, set this to -I instead of -isystem.

XXX this should really be either not required or a -isystem-cxx, but for
now this is required to build libstdc++-v3, which otherwise complains
heavily about STL headers not being C code(!)

whitespace cleanup

Add missing } for rev. 1.43.

Also recognize cpp files as C++ files for CYGWIN case.

recognize .cpp as c++ files.

Back this out until I know why it does not work.

Add .cpp dependency rules.

Fix _HOST_CYGWIN lossage.

Hacks for limited command and environment space (John Gordon)

The traditional C preprocessor is fixed in GCC 3.3 now, so no need
to skip -traditional-cpp for HAVE_GCC3.

Oops, AINC actually is still used; put it back.

The MIPS toolchain now defines __ABICALLS__ when abicalls are in use,
and the sources now use that define, so there is no need for us to define
ABICALLS.  Since that was the only use for the AINC variable, garbage-collect
it.

Don't use -traditional-cpp if HAVE_GCC3.

If DESTDIR is set, do a similar nostdinc dance for the standard C++ include
directory.

Improve readability and reduce redundant redundancy in share/mk; add some
descriptive comments.  The guts of hostprog, lib, and prog are mostly
unchanged; this will be done in another pass.

Make several things work properly with UNPRIVILEGED set, and make the
install-time "cmp" logic work for MANZ.  Also reimplement INCS{DIR,NAME}_foo
(requested by wiz) in a backwards compatible way.

Reviewed by christos.

Rather than explicitly saying "clean cleandir:" all the time, make clean
a dependency of cleandir (so that "make cleandir" implies "make clean").

Make "distclean" a REAL synonym for "cleandir" by actually aliasing it
in <bsd.own.mk>, rather than sprinkling it in all the .mk files (and
some Makefiles throughout the tree).

Make `dependall' play nicely with SUBDIR, using a hack similar to `realinstall'
for both `depend' and `all'.
Also, remove the old hack that invokes an extra shell for .depend, now that
make(1) has been fixed.

Pull up to last week's -current.

Add missing "\" in two line shell command.

Pass -traditional-cpp to mkdep for assembler files,
as in the actual compilation (${COMPILE.S} rule in sys.mk).

correct previous fix:
- add CLEANDEPEND variable for stuff which should be cleaned out
in "cleandepend" phase.
- add it to the files removed during the cleandepend phase.

Change CFLAGS to AFLAGS for mkdep of .S/.s files.  Fixes "make depend"
on pmax kernels (and doesn't break "make depend" on entire tree).

Use :Q to quote arguments to "echo" so that quote marks don't fall out of
the printed mkdep command.

Our make can't actually turn off implicit rules per-target, and the
"depend:" rule had no actual commands to execute, so it used the implicit
".c:" rule in gas.new (which has a depend.c).  Add a command here -- just
"true" -- that prevents use of implicit rules.  This may need a new
.-option in make(1) to be done Right.

Kill spaces in the $$files check introduced by borrowing the C++ rule for
ObjC.

Add dependency rule for Objective-C files.

add distclean as a synonym for cleandir

Automatically add the .c files corresponding to .l and .y files to DPSRCS.
Automatically add DPSRCS to CLEANFILES.

Remove cruft.

Set the .NOPATH flag for the .depend target.

Don't use _SUBDIRUSE outside of bsd.subdir.mk.

Always define cleandepend.  Don't use a .depend target if SRCS is empty.

Shuffle some .PHONYs around, to work better with `-t'.

Add a pile of missing .PHONYs.

Allow mkdep to exist as a different name (fixes PR#3302)

- make sure the afterdepend rule fires `after' .depend gets created.
  [fixed for parallel make]
- make depend prettier printing.

Remove `-+' arg to mkdep.

Add CPPFLAGS to mkdep lines, for cross-compilation.

RCSID frobbing, and minor sanity adjustments.

don't be careless with _SUBDIRUSE

cleandepend should happen on 'cleandir' not 'clean', from michaelv@iastate.edu

do .S files, too

Depend .s files.

Don't do tags and cleandepend if no SRCS.

Makefile: Use `install -d' to create directory.
bsd.dep.mk: Add defaults for `beforedepend' and `afterdepend'.
sys.mk: 8 tabs --> space

Move man rules to sys.mk.
Use bsd.subdir.mk for building subdirectories.
Don't forget to remove x.cc and x.C when using SHAREDSTRINGS.
Make ${PROG} always depend on ${LIBCRT0}.
Don't duplicate code between cleandir and clean.
Move tags target to bsd.dep.mk.
Move STRIP to bsd.own.mk.
Remove redundant definitions of {BIN,MAN}{DIR,GRP,OWN,MODE}.
Move {LIB,DOC}{DIR,GRP,OWN,MODE} defaults to bsd.own.mk.

Abstract out dependency stuff, and add RCS identifiers.