Makefile: Run `postinstall -d /', not `postinstall -d //'.

This happens when doing `build.sh install=/'. The message is less
confusing and it makes postinstall's job simpler for detecting when
it's installing to / rather than somewhere else.

more NBUILDJOBS deprecation

The NBUILDJOBS option was deprecated in 2002;
there's no need to keep warning about it,
remove from params / show-params,
and only document as obsolete.

build.sh: add show-params operation

Add show-params operation to the makefile target, and reword
params operation to describe what it's done since 2012.

(In 2012-11-04 the Makefile's previous params target was renamed
to show-params and a new params target was added to save a 'params'
file, and build.sh wasn't adapted.)

Deprecate extsrc/, EXTSRCSRCDIR, MKEXTSRC

Remove support for building extsrc/:
1. Makefile: remove do-extsrc target.
2. build.sh: remove options -y and -Y extsrcdir.
3. distrib/sets: remove support for extsrc in various tools
including the options -L ext and -y, and the extsrc sets.
4. doc/BUILDING.mdoc: remove docs for extsrc/, EXTSRCSRCDIR, MKEXTSRC (etc)
5. bsd.own.mk and various Makefiles: remove support for extsrc/,
EXTSRCSRCDIR, MKEXTSRC.

As proposed on tech-kern and tech-userlevel on 2022-01-07
and followed up on 2022-08-21.

Makefile: fix location of postinstall program for MAKEVERBOSE > 2

If MAKEVERBOSE > 2, each shell command from a make target is echoed.
This resulted in two additional words ending up in the variable
_POSTINSTALL. Noticed by Brad Harder.

Before:
$ make -v _POSTINSTALL MAKEVERBOSE=3
echo .../usr.sbin/postinstall .../usr.sbin/postinstall/postinstall ...

After:
$ make -v _POSTINSTALL MAKEVERBOSE=3
.../usr.sbin/postinstall/postinstall ...

G/c spurious whitespace.

automatically remove old debug files for modules.

Enhance the support of LLVM sanitizers

Define _REENTRANT for MKSANITIZER build. This is needed for at least stdio
code. This caused new build issued with duplicated symbols in few places
and rump kernel code picking different code paths borrowed from libc.
Handle all this in one go.

Add bsd.sanitizer.mk to share common code used by programs and libraries.

Switch from realall to beforeinstall target in .syms files. This is more
reliable in MKSANITIZER.

Correct the build rules of LLVM sanitizers

Fix generation of .a.syms files in LLVM sanitizers. The old rules weren't
functional and sometimes they were fatal in the build.

Add the .a.syms files to distribution sets.
These files are generated only in certain .a files in 64bit version of
sanitizers.

Correct the location of files with LLVM sanitizers. Move the LLVM sanitizer
entries from base to comp files.

Rework the LLVM compiler_rt build rules

Register external/bsd/compiler_rt as a _SUBDIR. This will allow to handle
properly includes specific for compiler_rt.

Centralize build_install target for external/bsd/compiler_rt.

With the above changes MKOBJDIR is now respected properly for me.

Do not install LLVM sanitizers into TOOLDIR for !HAVE_LLVM.

Enhance MKSANITIZER support

Add i386 support for LLVM sanitizers.

Add new build phases do-sanitizer-* for building LLVM sanitizers after libs
and before other programs. This phase is important as sanitizers need
to link with libs and sanitized programs must link with sanitizer runtimes.

branches: 1.327.2;
Use the generated "postinstall" in the objdir for postinstall-check
and postinstall-fix targets - we do not have a simple postinstall script
in the source tree any more.

Fix editor mishap

.WAIT for tools to complete.
There's a lot of waits inside tools, but nothing that promises that we haven't
started doing the next subdir in the top level.

Shows up as warnings in the cleandir stage on releng builds:

cleandir ===> external/mit/xorg/server/drivers/xf86-video-suncg6
cleandir ===> lib/libpam/modules/pam_deny
sh: /home/builds/ab/HEAD-llvm/sparc64/201811091720Z-tools/bin/nbawk: not found
nbmake[9]: "../../Makefile.xf86-driver" line 23: warning: "/home/builds/ab/HEAD-llvm/sparc64/201811091720Z-tools/bin/nbawk '/^PACKAGE_VERSION=/ { match($1, "[0-9]+\\.[0-9]+\\.[0-9]+"); version = substr($1, RSTART, RLENGTH); } END { print version }' /home/source/ab/HEAD-llvm/xsrc/external/mit/xf86-video-ati/dist/configure" returned non-zero status
sh: /home/builds/ab/HEAD-llvm/sparc64/201811091720Z-tools/bin/nbawk: not found

branches: 1.324.2;
Re-add files that were accidentally deleted in my previous commit.

Synch with HEAD

include bsd.clean.mk so that we actually clean up the attempted "params"

branches: 1.321.2;
Include somewhat misleadingly named do-x11 into BUILDTARGETS even with
NOBINARIES set (subject to MKX11).

do-x11 target builds and installs only the tools and libraries, X11
programs are built during the regular recursive build in extsrc. Edit
its description accordingly and move it before do-build to match their
order in BUILDTARGETS.

While here, drop the "either" clause that has lost its "or" a few
years ago when xfree86 reachover was removed.

Add NOBINARIES, useful to build tools are libraries which is what's needed
for mknative.

don't descend into gnu

branches: 1.317.2; 1.317.4;
fix direct reference to texinfo, bleh

remove the xfree86 reachover makefiles and the vast majority of
the support in the rest of the source tree.

X11 sets could use some cleaning up perhaps (just deletion, as
we've never really marked the old X11R6 as obsolete for native
xorg using platforms so far either.)

Add MKCOMPATX11 knob for making X11 compat libraries.

Support building the xorg X11 libraries under MKCOMPAT.
Add XORG_MACHINE_ARCH to compat's bsd.*.mk
Turn xorg makekeys and makestrs into real host tools which are installed into
${TOOLSDIR}

PR 49870: pass the xsrc path to postinstall

Reorganise the way xorg is build. Move build of makestrs into src/tools
as it is needed for the "includes" phase. Make it symlink all necessary
headers in obj to decouple it from "includes" itself, breaking the
dependency cycle. Move the do-x11 target between do-lib and do-build, so
that libraries can get the benefit of build_install (correct .WAIT
behavior) and everything else like "includes" uses the plain SUBDIR
entry in external/mit.

iso -> images

remove obsolete BUILD_CC_LIB_BASEDIR and BUILD_CC_LIB_BASETARGET vars

branches: 1.309.2;
Run "postinstall fix obsolete_stand" just before checkflist,
if we are building with DESTDIR != /

Pass -m MACHINE -a MACHINE_ARCH to postinstall

Pass SED=${TOOL_SED:Q} to postinstall.

remove the build support for GCC 4.1, and any hacks i found for it.

there could still be more -- i only looked for lines matching
both "HAVE_GCC" and "4" as words (grep -w.)

branches: 1.305.2;
prepare for moving GCC 4.5 into gcc.old:

- convert to using ${EXTERNAL_GCC_SUBDIR}
- define base-external-gpl3-gcc* subdir as GCC_SUBDIR
- use <bsd.init.mk> over <bsd.own.mk> for a bunch of places; mostly
because it arranges for ../Makefile.inc to be included earlier, and
don't bother including the latter if the former is already included.
- move all .PATH: settings after <bsd.{own,lib}.mk> so that all
valid variables are set before it is evaluated
- rename mknative-gcc* to match their subdir name.

XXX the relationship between the Makefile.inc/Makefile.gcc_path files
is kind of sketchy, it would be great if this was fixed.

eliminate use of bsd.sys.mk from Makefiles

hp700 rescue/ldd fails on -j16 build if usr.bin/ldd/elf32/libldd_elf32.a
hasn't been made yet. Add an explicit .WAIT before rescue to avoid this.

Add src/etc/Makefile.params, containing the definition of the
RELEASEVARS variable, and commands related to printing the values of
the variables whose names are in RELEASEVARS.

Add an awk script to remove noise printed by "make -j" or high levels
of MAKEVERBOSE, so we get only the variables names and values. The
values are escaped so that variables containing embedded newlines,
quotation marks, and backslashes, are passed through safely.

Adapt src/etc/Makefile and src/Makefile to use the new ${PRINT_PARAMS}
command defined in src/etc/Makefile.params.

Now ${DESTDIR}/etc/release and the params file in the top-level
.OBJDIR should never contain unwanted noise, even after a build with
MAKEVERBOSE=4.

Move the build_install logic from lib/Makefile into share/mk and re-use
it for tools. The existing logic broke for the LLVM build with the
recent .ORDER changes in make.

Create a "params" file in the top-level .OBJDIR during "make build".
This file will contain the values of all variables that can affect
the build process. Take care not to update the file's timestamp if
the contents do not change.

${_SRC_TOP_OBJ}/params is referenced by ${_NETBSD_VERSION_DEPENDS} in
bsd.own.mk.

Rename the old "params" target to "show-params". This simply
prints the values of the variables that can affect the build.

branches: 1.299.2;
Implement build_install in src/lib/Makefile, which traverses the
subdirectories and invokes dependall/install. Do this in groups
delimited by .WAIT to ensure that depending libraries can use the
installed versions and don't need to know the locations in the source
tree.

Use this new target in src/Makefile to replace most of the adhoc library
logic with two special cases, src/lib and src/compat. Adjust
sys/Makefile to include the module directory when building them. Add
some necessary .WAITs in src/lib/Makefile to reflect the dependencies
from src/Makefile and also add the rump libraries here.

Add HOST_SH to _POSTINSTALL_ENV; needed for
usr.sbin/postinstall/postinstall revision 1.145.

Pass STAT=${TOOL_STAT:Q} through to postinstall.

TOOL_GREP is defined in share/mk/bsd.own.mk, but is not actually built
from tools/Makefile. So don't try to use it here.

Pass all relevant tool variables to postinstall, not just a subset.

Fix unbalanced quotes in .if ${MKRUMP} != "no"

exclude rump targets if MKRUMP = no

Add usr.bin to the exclude list for make includes.
This is no longer needed after flex was moved to the external framework.

branches: 1.291.2;
s/is the not root/is not the root/ in a comment - from Snader_LB

Add "live-image" and "install-image" target support to build.sh.

"live-image" target builds pre-installed disk images that can be used on
emulators or boot from USB memory sticks to try NetBSD without installation.
Currently amd64, i386, pmax, sparc, sparc64, sun2, sun3, and vax
(which have working emulators and don't require extra tools like preparing
msdosfs or partitioning MD label structures) support this target.

"install-image" target builds an bootable installation disk image that can
be used as an install CD but burned into USB memory sticks etc.
Current only amd64 and i386 (which would support USB boot) have this target.

For more details (and known issue) see following posts on netbsd-bugs@:
http://mail-index.NetBSD.org/netbsd-bugs/2011/08/06/msg023639.html
http://mail-index.NetBSD.org/netbsd-bugs/2011/09/23/msg024207.html
http://mail-index.NetBSD.org/netbsd-bugs/2011/12/07/msg025166.html
http://mail-index.NetBSD.org/netbsd-bugs/2011/12/08/msg025178.html

No particular comments about implementation, and
"go ahead" comments from mrg@, riz@ and christos@.

Closes PR toolchain/45153 and PR misc/45155.

branches: 1.289.2;
Move the code for creating src/BUILDING from "make build-docs" in
src/Makefile to "make regen" in src/doc/Makefile, and update a comment
in doc/BUILDING.mdoc to match.

Fix typo that prevented installmodules from working with -U

add an installmodules=<dir> command to build.sh

re-enable the gcc 4.5 crtstuff/libgcc builders now that they work.
also build the gmp, mpfr and mpc libraries in this case.

fix the parts that build libgcc/crtstuff, etc., for GCC 4.5.

remove all the code that supported HAVE_GCC=3. mostly from chuq.

build evbppc modules in 3 different flavours:

- oea (default)
- ibm4xx
- booke

this, together with earlier machine_module[] work, should allow evbppc
platforms to use modules.

branches: 1.282.2;
Pass down INSTALLSETS variable to installsets target. Use :Q to escape
shell meta-characters, as the variable contains a list of sets, separated
by spaces.

$ ./build.sh -V INSTALLSETS="base etc" install=idir

should now work as expected.

Reviewed by apb@ on tech-toolchain.

branches: 1.281.2;
Remove the do-external-lib and do-gnu-lib targets, along with
external/lib/Makefile and crypto/external/lib/Makefile, replacing
them all with SUBDIRs directly from lib/Makefile.

compat/compatsubdirs.mk becomes simpler now, as everything is built
from lib/Makefile, meaning all the libraries will now be built under
compat so update the set lists to account for that.

Centralise USE_COMPILERCRTSTUFF.

PCC might need help.

no need to create a do-lib-libdes target, the directory no longer exists

remove almost all the ability to build netbsd with an a.out target.
we're ELF now, and there are many missing checks against OBJECT_FMT.
if we ever consider switching, the we can figure out what new ones
we need but for now it's just clutter.

this doesn't remove any of the support for exec_aout or any actually
required-for-boot a.out support, only the ability to build a netbsd
release in a.out format. ie, most of this code has been dead for
over a decade.

i've tested builds on vax, amd64, i386, mac68k, macppc, sparc, atari,
amiga, shark, cats, dreamcast, landisk, mmeye and x68k. this covers
the 5 MACHINE_ARCH's affected, and all the other arch code touched.
it also includes some actual run-time testing of sparc, i386 and
shark, and i performed binary comparison upon amiga and x68k as well.


some minor details relevant:
- move shlib.[ch] from ld.aout_so into ldconfig proper, and cut them
down to only the parts ldconfig needs
- remove various unused source files
- switch amiga bootblocks to using elf2bb.h instead of aout2bb.h

descend into sys/rump/kern

adjust the do-compat-* rules now that src/compat/lib and friends no longer exist

Support "extsrc", externally added programs and libraries. Users can write
their own reach-overs, cross-build, install, and get set files just like base
and X11 / X.org. (These sets are not included as TNF releases.)

build & install librumpdev_netsmb and librumpfs_smbfs components

Let us use MKKMOD=yes to leave the kernel modules out of the
distribution.

Build modules before rump libs so that something else can be
unwarrantedly blamed for a change.

This commit follows the style guide and is entirely non-build-tested.

add crypto/external to the build.

Make clean_METALOG in distrib/sets early in the build, to ensure that
the METALOG does not grow without bounds on multiple MKUPDATE builds.
(Previously, sanitise_METALOG in distrib/sets/Makefile used to do
something like this near the end of a build.)

Remove an explicit BUILDTARGETS do-compat-external-lib before do-build
because now src/compat/Makefile descends there and no particular
dependency on 32 bit compat external libs during do-build.

Disable grep host tool and unbreak NetBSD build until I can come up with a
clean fix.

Use awk and grep host tools where required. 'build.sh release' now works
on Solaris (but only with HOST_CC=/usr/sfw/bin/gcc for now).

branches: 1.266.2;
Add the top-level do-compat-external-lib target that mimics
do-external-lib so that we can build the compat libraries for
ATF. Also add compat/external/lib/Makefile following the same
example of external/lib/Makefile, which descends into the
external libraries and builds them.

Build kernel modules the same time as the rest of the system libraries.

Invoke postinstall with the environment variable MAKE set up to the relevant
value, so that postinstall picks up correct values.

- retire MKXORG from public use
- introduce X11FLAVOUR to choose src/x11 vs src/external/mit/xorg
for the X11 to build if MKX11=yes is set. it takes the values
of either Xorg or XFree86.
- default to Xorg on alpha, i386, macppc, shark, sparc and sparc64
- remove MKXORG_WITH_XSRC_XSERVER, unused and never useful

branches: 1.262.2;
- enable 32-bit library builds on amd64 and sparc64

- add 3 new top-level targets:
do-compat-lib-csu
do-compat-libgcc
do-compat-lib-libc
and run them before "do-build" if ${MKCOMPAT} != "no"

Add a do-top-obj target, which builds the top-level obj directory.
This uses the new NOSUBDIR support in bsd.subdir.mk.

Insert do-top-obj near the beginning of BUILDTARGETS. The first few
items are now:
cleandir: cleans the tree.
do-top-obj: creates the top level object directory.
do-tools-obj: creates object directories for the host toolchain.
do-tools: builds host toolchain.
obj: creates object directories.

Also improve a comment and split a long line where SUBDIR is set from
_SUBDIR.

Defer the "make obj" step until after tools are built and installed.

There may be commands embedded in "!=" assignments that depend on tools
being available, and previously these would have printed error messages
during "make obj". They may still print error messages during "make
cleandir", which (if it is performed at all) is performed before tools
are built.

add do-sys-rump-net-lib

PR/39687 - Henning Petersen -- comment typo

Improve support for MKXORG!=no to enable builds into external/mit/xorg.
This will be mutually exclusive to MKX11!=no.

Build pcc and pcc-based crtstuff if using pcc compiler.

Install rump libraries and utilities to the base system and remove the
private non-installed build infrastructure from sys/rump.

breakdown of commit:
* install relevant headers into /usr/include/rump
* build sys/rump/librump/rumpuser and sys/rump/librump/rumpkern
from src/lib and install as librumpuser and librump, respectively
+ this retains the ability to test a librump build with just the
kernel sources at hand
* move sys/rump/fs/lib/libukfs and sys/rump/fs/lib/libp2k to src/lib
for general consumption, they are not kernel-space dwellers anyway
* build and install sys/rump/fs/lib/lib$fs as librumpfs_$fs
* add chapter 3 manual pages for rump, rumpuser, ukfs and p2k
* build and install userspace kernel file system daemons if MKPUFFS=yes
is spexified
* retire fsconsole for now, it will make a comeback with an actually
implemented version shortly

if MKXORG != no, use external/mit/xorg subdir instead of x11 subdir for
the mk-x11 target.

Descend into external.
Add 'do-external-lib' target to build external/lib "early"

branches: 1.252.2; 1.252.4;
Consistently document RELEASEDIR/RELEASEMACHINEDIR.

branches: 1.251.4;
Add the tests.tgz set

This adds a new tests.tgz set to releases which includes all the tests
for the system. It is important to note that this set does not rely on
comp.tgz: a user of the system can run the tests without having the
development tools installed, which can be useful in a production machine.

Do not abort on exit code 1 from postinstall check, this is expected.
Reported, fix suggested and tested by David Holland.

(Why does "make build" invoke postinstall check on destdir is another
question, it does not seem to have much sense.)

Provide 'do-libgcc' to build gnu/lib/crtstuff${LIBGCC_EXT} (if necessary)
and then gnu/lib/libgcc${LIBGCC_EXT}.
Makes it easier to cut&paste from the list of targets built with 'make build'.

Add an iso-image-source option which does the same thing as iso-image but
includes source sets as well. The infrastructure for this was already
there, although it needed a bug fix. Will look at adding arbitrary
directories next.

branches: 1.247.4;
Add an iso_image target (iso-image as a target name doesn't play well with
bsd.subdirs.mk) in distrib/makefile, which builds an iso image for $MACHINE
with binary sets, stored in ${RELEASEDIR}/iso. The image is bootable for:
alpha, amd64, cats, i386, pmax, sgimips, sparc, sparc64, sun3, vax.
mac68k/macppc no there yet because of missing feature in makefs.

call iso_image in distrib/ for iso-image in the top Makefile.

branches: 1.246.2;
don't filter out .WAIT when converting from _SUBDIR to SUBDIR

remove support for building (with) GCC 2.95. also:
- always install <stdbool.h>
- don't generate a fake one for vax / gettext.

retire HAVE_GCC3/HAVE_GCC4 and introduce HAVE_GCC that is set to 2, 3 or 4.

* Add an "iso-image" target in the top level Makefile,
which invokes make iso-image in the etc directory.
* Add an "iso-image" action in build.sh, which
invokes make iso-image.
* Document the above in doc/BUILDING.mdoc.
* Re-generate BUILDING.

Approved by christos

Add .MAKE to all targets that use ${MAKEDIRTARGET} otherwise make closes
the job-token pipe and the sub-makes have their own '-j n' processes.
As well as giving unexpected parallelism, it makes it impossible to stop
the build promptly on errors.

Add a new "syspkgs" target, which simply chains to the "syspkgs" target
in distrib/sets/Makefile.

Reviewed by agc

Modify "release" to invoke "make distribution" instead of "make build",
so that the "obsolete file removal" functionality of "distribution" is used.
Fixes build problems numerous people have experienced with MKUPDATE=yes
and the "release" target.

use usr.sbin/postinstall/postinstall instead of etc/postinstall

branches: 1.238.2;
correct the comment describing the build order

DO do-gnu-lib-crtstuff${LIBGCC_EXT} only when ${HAVE_GCC3} != "no",
since we don't have crtstuff subdirectory.

Switch to using GCC's supplied crtbegin / crtend files.

The do-lib-libdes BUILDTARGET is unnecessary now we have the DEPLIBS
support that I added to src/lib/Makefile.inc.

Move '${MAKEDIRTARGET} x11 distribution' from Makefile::distribution to
etc/Makefile::distribution, and only perform if ${MKX11}!=no.
Should fix a problem experienced by Alistair whilst doing a "make release".

Perform 'make distribution' in ./x11/ as part of the distribution target.

Convert lines of the form
(cd ${.CURDIR}/somedir && ${MAKE} AVAR=value atarget)
to
${MAKEDIRTARGET} somedir atarget AVAR=value
which results in a "prettier" display of these operations

Be more gcc like in the way we deal with libgcc so that sh3/hppa millicode
and exception handling have a chance of working properly.

- creates libgcc, libgcc_eh and libgcc_s

- updates LIBGCC_SPEC to use them appropriately.

There's a hack in here at the moment with respect to libgcc_so in that it
is preferable to link against libgcc_so will only when -shared-libgcc is
specified (the c++ frontend does this automatically.) Configurations where
LINK_EH_SPEC is defined already do this. The gcc configuration for
NetBSD/alpha and another NetBSD platform (I forget which) actually define
LINK_EH_SPEC probably by accident rather than design.

- updates share/mk to use the compiler's knowledge of what needs linking into
libraries and executables. This removes an hppa hack.

- updates the sets for the newly created libgcc* files.

- support for linking against the _pg version of libgcc has been removed.

MAKEVERBOSE support for build-docs

implement params with cd etc && make params

Apply the .PHONY attribute where appropriate.
Fixes [toolchain/21453] from rafal@.

Implement do-x11, to run "make build" in ./x11/
If MKX11 != no, add "do-x11" to BUILDTARGETS.

Say ``hello'' to Imakeicide.

rename release-info to etc-release
rename install-release-info to install-etc-release

Run "cd etc && make install-release-info" at the end of the "build" target.
Based on Hubert Feyrer's suggestion on tech-userlevel@.

only build do-lib-libdes if ${MKCRYPTO}!=no

Consistently check ${TOOLCHAIN_MISSING} against "no".

fix some comments

netbsd.org->NetBSD.org

maybe build gnu/lib/libgcc3 instead of libgcc[2]

document do-lib-libdes

Make and install libdes before the rest; libkrb needs it.
XXX there must be a better way.

tweak how postinstall-fix-obsolete is made. (may fix various problem reports)

redo previous (functionality rollback of 1.213) to retain the style of 1.213

Undo revision 1.213 now that grotty is fixed. Tested by lukem, thanks!

Add MKUNPRIVED; if not no has the same semantics as if UNPRIVED was defined.
Replace defined(UNPRIVED) tests with ${MKUNPRIVED} != "no"

Add MKUPDATE; if not no has the same semantics as if UPDATE was defined.
Replace defined(UPDATE) tests with ${MKUPDATE} != "no"

Improve documentation for these and other make flags.

grotty from groff 1.19 needs -c for -b, -o or -u to work.

This is damn annoying, because it means the old method of "groff -P-bou"
won't suppress escape codes, unless -P-c is also provided.
(Mmm, quality software...)

Add postinstall-fix-obsolete (to invoke "postinstall fix obsolete" -
funny that :), and call in "distribution" just before the "make checkflist".
Should minimise a bunch of the "I use make UPDATE= and checkflist failed"
errors, but not all of them.

Rename a large chunk of the make(1) variables which refer to a
program/tool from "FOO" to "TOOL_FOO". The new variables are:
TOOL_ASN1_COMPILE TOOL_CAP_MKDB TOOL_CAT TOOL_CKSUM TOOL_COMPILE_ET
TOOL_CONFIG TOOL_CRUNCHGEN TOOL_CTAGS TOOL_DB TOOL_EQN TOOL_FGEN
TOOL_GENCAT TOOL_GROFF TOOL_HEXDUMP TOOL_INDXBIB TOOL_INSTALLBOOT
TOOL_INSTALL_INFO TOOL_M4 TOOL_MAKEFS TOOL_MAKEINFO TOOL_MAKEWHATIS
TOOL_MDSETIMAGE TOOL_MENUC TOOL_MKCSMAPPER TOOL_MKESDB
TOOL_MKLOCALE TOOL_MKMAGIC TOOL_MKTEMP TOOL_MSGC TOOL_MTREE
TOOL_PAX TOOL_PIC TOOL_PREPMKBOOTIMAGE TOOL_PWD_MKDB TOOL_REFER
TOOL_ROFF_ASCII TOOL_ROFF_DVI TOOL_ROFF_HTML TOOL_ROFF_PS
TOOL_ROFF_RAW TOOL_RPCGEN TOOL_SOELIM TOOL_SUNLABEL TOOL_TBL
TOOL_UUDECODE TOOL_VGRIND TOOL_ZIC

For each, provide default in <bsd.sys.mk> of the form:
TOOL_FOO?= foo
and for the ${USETOOLS}=="yes" case in <bsd.own.mk>, provide override:
TOOL_FOO= ${TOOLDIR}/bin/${_TOOL_PREFIX}foo

Document all of these in bsd.README.

This cleans up a chunk of potential (and actual) namespace collision
within our build infrastructure, as well as improves consistency in
the share/mk documentation and provision of appropriate defaults for
each of these variables.

add MAKECONF to vars that "params" target shows

Add "params" target, to display the value of various variables.
This is part of a broader plan to improve diagnostics & support
to end-users of build.sh.

update "targets invoked by make build" to match reality ...

build libgcc early, if at all

Add "sourcesets" argument/target, which builds source sets into
RELEASEDIR/source/sets

.WAIT before distrib, since distrib may need things built earlier
in the build process.

libcrypto no longer depends on libcrypt, so there is no need to build
it early.

Fix merge botch in last.

Always build tools/compat, so that random host helper programs
can use libnbcompat, if necessary.

Link libcrypto against libcrypt to make sure crypt() is still there -- and
therefore build it earlier in the build process as well.

all:
Use /var/db/obsolete instead of /etc/obsolete

etc/Makefile:
Create separate target "install-obsolete-files" to populate
/var/db/obsolete, instead of using "install-etc-files".

Makefile:
Add do-obsolete target, to run "cd etc && make install-obsolete-files",
and add this to BUILDTARGETS.
This moves the "obsolete files" creation from "distribution" to "build".


Per discussion with Andrew Brown.

- Add "sets" top-level Makefile target and build.sh operation.
[Latter suggested by Julio Merino in private email]
- Add "help" build.sh operation.

Rework how KERNOBJDIR functions; now it's always determined with
cd ${KERNSRCDIR}/${KERNARCHDIR}/compile && ${PRINTOBJDIR}
This is far simpler than the previous system, and more robust with
objdirs built via BSDOBJDIR.

The previous method of finding KERNOBJDIR when using BSDOBJDIR by
referencing _SRC_TOP_OBJ_ from another directory was extremely
fragile due to the depth first tree walk by <bsd.subdir.mk>, and
the caching of _SRC_TOP_OBJ_ (with MAKEOVERRIDES) which would be
empty on the *first* pass to create fresh objdirs.

This change requires adding sys/arch/*/compile/Makefile to create
the objdir in that directory, and descending into arch/*/compile
from arch/*/Makefile. Remove the now-unnecessary .keep_me files
whilst here.

Per lengthy discussion with Andrew Brown.

Add an obj rule to create ${KERNOBJDIR} here, at the end of the "make obj"
pass. (etc/Makefile doesn't create this any more).

Install release files under "${RELEASEDIR}/${MACHINE}/..." instead of
"${RELEASEDIR}/...".

${RELEASEDIR} is never cleaned , and ${RELEASEDIR}/${MACHINE} is only
cleaned if UPDATE is not defined.

The obj target for src/etc is invoked too early for it to function
properly in a clean tree (bsd.subdir.mk makes the obj target in
.CURDIR depend on the recursive targets). If "make obj" is invoked a
second time, the problem will correct itself. To work around this:

(1) Before including bsd.subdir.mk, create a target called "obj-etc"
so that the "regular" one (from bsd.subdir.mk) won't be instantiated.

(2) After including bsd.obj.mk, create a "real" obj-etc target that
does the required work with .USE as a "source" so that we can merely
cons it onto the actual obj target. Note that this is only done if
the obj target has commands (ie, NOOBJ was not set and MKOBJ was not
no and whatever else you might have come up with) and we would have
recursed into etc anyway (etc is in the SUBDIR list).

This makes the obj target in src/etc get invoked *after* the obj
target in the root of the source tree so that KERNOBJDIR will have the
correct value.

Reviewed by lukem.

fix DESTDIR checks if !defined(DESTDIR)

* Add two new (experimental) targets:

buildworld Builds a distribution into DESTDIR as per
"make distribution", except that DESTDIR
*must* be defined and *must not* be "" or "/".

installworld Copies (using distrib/sets/maketars -i ...) from
${DESTDIR} to ${INSTALLWORLDDIR} (defaults == `/'),
and then runs "postinstall-check" against
${INSTALLWORLDDIR}.
If ${INSTALLWORLDDIR} == "/", ensure that
`uname -s` == "NetBSD" and `uname -m` == ${MACHINE},
to prevent nasty accidents...

* Tweak some comments

improve previous; show times at end of distribution/release/snapshot as well

show the start time at the end of the build as well.
inspired by [misc/18173] from Julio Merino.

Add do-{target} aliases for: cleandir obj includes
to be consistent with the other do-TARGET targets for make build.
(These are not used by "make build" per se, they're just to make my
life easier)

MKDYNAMICROOT for a.out

move BUILDING.mdoc -> doc/BUILDING.mdoc

* Eliminate the USE_NEW_TOOLCHAIN variable. Instead, split it into
two variables:

TOOLCHAIN_MISSING -- set to "yes" on platforms for which there is
no working in-tree toolchain (hppa, ns32k, sh5, x86_64).

EXTERNAL_TOOLCHAIN -- if defined by the user, points to the root of
an external toolchain (e.g. /usr/local/gnu). This enables the cross-build
framework even for TOOLCHAIN_MISSING platforms.

If TOOLCHAIN_MISSING is set to "yes", MKGDB, MKBFD, and MKGCC are all
unconditionally set to "no", since the bits are not there to build.

If EXTERNAL_TOOLCHAIN is set, MKGCC is unconditionally set to "no",
since the external toolchain's compiler is not in-sync with the
in-tree compiler support components (e.g. libgcc).

* Set MACHINE_CPU much earlier in bsd.own.mk, so that more tests in
that file can use it.

Add a do-lib-libc target which is processed after do-lib-csu. This
ensures that the results of the libc build are available to other
lib builds (notably, llib-lc.ln).

print date when make {distribution,release,snapshot} finished

add "do-ld.elf_so" target, which builds & installs libexec/ld.elf_so.
call after do-gnu-lib but before do-build.

descend into "rescue"

Fix botch in previous. Actually do "build" for "distribution".

After running the body of the "distribution" target, do a "checkflist", so
that people doing fresh builds into NFS root areas will know if their builds
are incomplete.

find the Makefile for postinstall-check in the presence of ${.OBJDIR}s

Fix what rev. 1.176 broke, in particular, doing a priv'd build into
an NFS root area for a diskless client. In such a situation, you
DO want the postinstall-check to be run.

The circumstances under which postinstall-check is now run:
* UNPRIVED is not set.
* The original target invoked was not distribution, release, or
snapshot.

Fix whitespace usage in comment.

Change directory to .CURDIR before invoking ${MAKE}.

In afterinstall, only run postinstall check if DESTDIR == / (or equivalents).
Fixes [misc/17275] from Gregory McGarry <g.mcgarry@ieee.org>

branches: 1.175.2;
Change -s srcdir and -d destdir so that they don't contain the "/etc" portion.
srcdir now defaults to /usr/src (from /usr/src/etc).
destdir now defaults to / (from /etc).

Move the METALOG from ${_SRC_TOP_OBJ_} to ${DESTDIR}.

There were too many synchronisation problems with using the former;
including situations such as a "make clean" performed between two
installs to the same DESTDIR would result in a truncated METALOG and
the resultant sets would be missing stuff such as include files that
don't get reinstalled if they haven't changed, even with !UPDATE.

Correct BUGS in BUILDING.mdoc now that most ports are USE_NEW_TOOLCHAIN

Obsolete NBUILDJOBS; build.sh just passes -jN through to make(1),
which inherits it cooperatively through the build tree. Fix
documentation so that it's shown to be deprecated.

If you use build -jN, please save full build logs so that errors due
to missing dependancies can be analyzed and corrected.

Use printf rather than echo -n, since the former works on all POSIX systems.

we might as well provide a handy "postinstall-fix" target whilst here

move postinstall from afterinstall to separate "postinstall-check" target.
requested by Matthias Scheler <tron@zhadum.de>.

run "./etc/postinstall check" in afterinstall:

Revert previous so crtbeginS and crtendS are available.

We don't need to do "install" of lib/csu specially. Just install it with
the rest of lib.

(None of lib/* needs csu installed in order to link. In the USETOOLS!=yes
case, csu may even depend on a newer libc, so installing csu before
compiling host tools in lib/ could break things even worse.)

Split the _SUBDIR assignment into two assignments so you can do things
like "grep ^_SUBDIR Makefile".

Build whatis.db and infodir-meta from the commands list of the afterinstall
target. Remember: inserting `dummy' targets is not equivalent of some
convenient macro expansion; it has side effects!

Restructure infodir-meta to call a proper submake. Put the code into the
install-info gnu subdir as it's the program updating the dir file. This
all has to happen because METALOG depends on _SRC_TOP_OBJ which is now
not defined in the top level Makefile per tv's last commit. Which is fine
as this rule should have gone into a program's Makefile anyways just like
the whatis.db one does.

If _SRC_TOP_OBJ_ gets set here, we will end up with a directory that may
not be the top level objdir, because "make obj" can happen in the *middle*
of "make build" (long after <bsd.own.mk> is calculated it). So, pre-set
_SRC_TOP_OBJ_ here so it will not be added to ${.MAKEOVERRIDES}.

Bomb if NBUILDJOBS is set, as it is badly broken in some way (as described
in toolchain/14837). This is the root of some problems thought to be of
different causes by users who forget to say that they are using
"build.sh -j"....

remove METALOG in clean

Add a wrapper for the "distribution" target, similar to the wrapper
for the "release" target.

Add a new target (infodir-meta:) that synthesizes the share/info/dir
file's metadata if UNPRIVED and MKINFO != no. Make afterinstall depend
on it as well as whatis.db.
XXX I wish there was a more rational place to synthesize the metadata,
but I can't think of one. The share/info/dir file is built by all the
world's install-info executions.

No, really revert previous. Recursing to etc works, and I'm in the middle
of fixing distrib right now. If something's broken, please report it via
the proper channels.

No, really take etc and distrib out of the default subdir
list. We've waited too long for the null target sweep.

Warn when "make build" is started with USE_NEW_TOOLCHAIN on a platform for
which USE_NEW_TOOLCHAIN is not verified to work [is not on by default].

Reverse the order of "make cleandir" and "make obj" back to the way they
were historically. Discrepancy noted by atatat.

Rather than build in `tools' and `lib' twice, build `tools' and `lib'
completely and install them, then skip recursing into those directories
on the final dependall/install.

The top level has an objdir now, so make sure to cd to ${.CURDIR} when
recursing in-place.

Major overhaul of src/tools and host toolchain. Summary of changes:

* Rewrite src/tools Make logic to work like the rest of the tree wrt
"dependall" and "install". The old "make build" hack is gone.

* Remove the MKTOOLS logic. This was linked to the "make build" hack,
and was only needed because TOOLDIR originally had no writable default.

* Redo the GNU configure/make logic to make it fit reasonably in a
BSD make wrapper. Use new ${.ALLTARGETS} variable to scan for
targets in $(srcdir), and mark them with .MADE: to prevent rebuilding.

* Only build cross tools in src/tools; remove some messy logic in
src/usr.* and src/gnu/usr.* that would do target filename rewriting
(improves consistency and readability).

* Add the ability to build cross gdb at tool build time by setting
MKCROSSGDB (default no) to "yes" in mk.conf.

* Add src/tools/groff and set up paths to work with this cross groff.

Create an objdir here for storage of timestamps, build-wide data, and so
forth.

By popular demand, one preformatted version of BUILDING, plaintext, no CRs.

Revert previous. <bsd.own.mk> already defaults MKOBJDIRS to "no". Also
clean up the evaluation of ${MKOBJDIRS} later -- no need for :Uno.

Add back original behavior of MKOBJDIRS defaulting to "no" so it doesn't
force people to set yet another variable to get the expected behavior

Add back logic which forces DESTDIR=/ for the distrib-dirs rule if it's not set
or it's blank. The new toolchain rules set it to blank which can run into
the sanity checks in etc/Makefile

* Always descend into "tools" regardless of USETOOLS setting. Instead,
just don't build the tools if USETOOLS!=yes. This permits objdirs to
be created even when USETOOLS=no.

* Clean up use of "-m" by specifying it to .MAKEFLAGS: if needed.

Work around parallelism lossage, because .ORDER doesn't work yet
(noted in PR bin/14404). This will be fixed back to using .ORDER later.

Add rules for generating plaintext and HTML versions of the "BUILDING"
document automatically (to be checked in when BUILDING.mdoc is changed).

Re-add intermediate dependall/install targets (do-lib-csu, do-lib,
do-gnu-lib) to permit restarting broken builds in the middle.

Add shortcut rules to avoid going into distrib on unneeded (and in some
cases, currently broken) targets.

Revert previous. etc and distrib do now belong in _SUBDIR; the correct fix
for broken builds is to fix the relevant Makefiles as was done for sparc.
This was done in the interest of simplicity of the system build (and the
fact that there may be a need to descend into these directories in the
near future).

Remove distrib and etc from _SUBDIR. These should never get added to the
subdir list as anything building into them will access them directly (ala
the make release rule or the mtree rule). There were good reasons for
controlling this at the top level before and those should remain in place.

On cleandir and obj builds these should get run through but that's it for
the "normal" rule sets. (This was breaking the builds as make includes doesn't
function inside of distrib for instance).

Add a couple more shortcut rules for the "tools" subdir to prevent recursion.

Change
${USETOOLS} != "no"
to
${USETOOLS} == "yes"
to allow USETOOLS to be tri-state: "yes", "no", or "never". The use of
these options will be documented in src/BUILDING.

Overhaul and simplify the top level Makefile, and add some sanity checking.

- "make build" cannot be invoked from the command line at the same time as
any standard recursive target except "obj" and "cleandir", as things like
"make all release" can lead to unexpected results.

- Put everything in SUBDIR, and auto-weed missing directories. This allows
the standard targets to work as expected (including in tools, etc, and
distrib).

- Leverage .for loops to simplify the prerequisite dependall/install phase
of compile-time dependencies (lib/csu, lib, gnu/lib). These are now all
distilled into one component rule ("do-build").

- Use the actual targets of obj, cleandir, and includes; no need for double
indirection goop in this case.

- GC all the deprecated "domestic" cruft. We are officially using (and
fixing, where appropriate) the cryptosystem integrated into src/crypto.

- Collapse the whatis.db build rule into a dependency of "afterinstall",
and nuke the "_BUILD" cruft.

Use @true instead of @${TRUE} in includes-foo targets, since there is no
variable TRUE defined in our makefile system.

This prevents "make includes" from breaking with older bsd.subdir.mk, and is
more consistent with the uses of "true" in the rest of the tree.

Do not descend into `tools' on make clean{dir,} if USETOOLS=no, as doing
so without TOOLDIR set will blow up.

Add "tools" to the list of subdirs for an explicit "make obj" or
"make cleandir". Fixes PR bin/14177.

"make build"'s logic and the conventional <bsd.subdir.mk> logic probably
need to be split apart here at some point to prevent these kinds of things
from causing odd problems in the future....

Use ${.CURDIR}/share/mk to include bsd.own.mk and bsd.subdir.mk into this
file in order to get proper definitions for a build. (Temporary; this
Makefile will be moved to a new name and replaced with a bootstrap Makefile
in the near future.)

UPDATING (1.38) says:

i386 platform builds must now set TOOLDIR in mk.conf or the
environment in order to build the system. This is a pathname
where host build tools will be installed, and must NOT be
the same as src/tools.

(Optionally, USETOOLS=no can be set in the environment if you
wish to avoid using the separate host toolchain for building a
subtree. Otherwise, this will default to "yes" automatically.)

However, building with USETOOLS=no set and no defined TOOLDIR ends up
with breakage, as the do-make-tools target is run regardless of whether
or not USETOOLS is set.

Hence, a wrapper around the body of the do-make-tools target:

.if ${USETOOLS} != "no"
...
.endif

...which makes the build work when USETOOLS=no, with no TOOLDIR set. This
would seem to more accurately match the description put forth in UPDATING.

Put do-make-tools before creating distrib-dirs (as binstall has to be
available before making distrib-dirs).

Always use src/share/mk as the source of the system .mk files for a build.

- Remove the include of Makefile.tools; this will be done elsewhere.
- Skip the "includes" phase of bin, games, libexec, regress, sbin, and usr.sbin.

Make sure DESTDIR is always set when calling the distrib-dirs target.
In the new build world order, DESTDIR is set to "" (in Makefile.tools)
if not already set, to ensure correct include/library handling. This
caused the top-level to not set DESTDIR=/, which confused distrib-dirs,
since it doesn't include Makefile.tools, and thus didn't see the same
DESTDIR that the top-level did.

The final step -- hook src/tools into the main build. If USE_NEW_TOOLCHAIN is
set, will descend into "tools" to build a toolchain, and use that toolchain
explicitly for the build (via tools/Makefile.tools).

While here, GC all the "egcs-detection" logic; it's not needed anymore, since
nothing is on gcc 2.7 any longer.

Wrap a long line.

clarify some variable documentation; from cagney

make previous work with -j. idea from cagney

fix PR#12832: split `make build' up into these steps (from new the
comments in src/Makefile):

# Sub targets of `make build,' in order:
# buildstartmsg: displays the start time of the build.
# beforeinstall: creates the distribution directories.
# do-force-domestic: check's that FORCE_DOMESTIC isn't set (deprecated.)
# do-share-mk: installs /usr/share/mk files.
# do-cleandir: clean's the tree.
# do-make-obj: create's object directories if required.
# do-check-egcs: check's that we have a modern enough compiler (deprecated.)
# do-make-includes: install include files.
# do-lib-csu: build & install startup object files.
# do-lib: build & install system libraries.
# do-gnu-lib: build & install gnu system libraries.
# do-dependall: builds & install the entire system.
# do-domestic: build & install the domestic tree (deprecated.)
# do-whatisdb: build & install the `whatis.db' man database.
# buildendmsg: displays the end time of the build.



this is something i've wanted to do for years....

parallelize the obj pass, too

if building w/o a DESTDIR, grab the system make files from the source
tree until they've been installed.

Descend into distrib on "make cleandir", too. This will let "make build"
with "MKOBJDIRS" set do the right thing in distrib.

Document the cool NOCLEANDIR/NOINCLUDES options that somebody added.

branches: 1.118.2;
Merge a bunch of things from crypto-us and crypto-intl into basesrc,
adding support for Heimdal/KTH Kerberos where easy to do so. Eliminate
bsd.crypto.mk.

There is still a bunch more work to do, but crypto is now more-or-less
fully merged into the base NetBSD distribution.

branches: 1.117.2;
remove two of the previous ${_M}; they were invoked before $DESTDIR/share/mk may have existed. add one more.

Add ${_M} where it was missing in a few places.

HAVE_GCC28 -> HAVE_EGCS, and make it match gcc 2.9, as well.

if MAKEOBJDIRPREFIX defined, include etc in make obj

use MKSHARE=no instead.

fix long standing bug in `make build' with $DESTDIR:
- during `make build' after /usr/share/mk has been installed, call
make with `-m $DESTDIR/usr/share/mk'.

At least one crypto-intl lib has an info file(s) - add MKINFO=no to the
early build of ../src/lib ala ../src/gnu/lib.

Handle new make variable BUILD_DONE which creates an empty build target if
it is set.

When you type `make obj', make obj dirs in distrib whether or not you
have MKOBJDIRS variable set.

Add a buildmsg target for build to depend upon instead of using an ifmake block. "make snapshot" now echos the start of build time.

share/tmac is now installed by `make includes', so don't do it again.

Make 'make build' hook into crypto-tree for crypto specific build-time
dependencies.

make sure include and include/sys are populated before starting to
make includes in lib

After ${MAKE} obj, explicitly make obj dirs in ${.CURDIR}/distrib as well.

Revert previous but move "cd share/tmac" later down so that any
generated files will go in obj dirs if they are created. After
discussion with the two Bill's on tech-userlevel.

Move the "${MAKE} obj" above the share/tmac stage so that the generated
files in that directory don't pollute the source tree before the obj
dirs are created.

Document MKOBJDIRS here, and sort the variable list in the comments
into alphabetical order.

Add MKOBJDIRS make variable to bsd.own.mk, defaulting to "no" but
overrideable in mk.conf

Document in bsd.README; this is distinct from "MKOBJ", which controls
whether "make obj" does anything.

In the top-level makefile, if MKOBJDIRS != "no", do a "make obj" at an
appropriate point during a "make build".

branches: 1.99.4;
use new dependall target instead of depend then all in make build

Remove all of the domestic/crypto-us glue from this file. It is handled
differently now.

Handle crypto-us just like domestic

Add the FORCE_DOMESTIC flag to force the 'all', 'includes',
and 'install' targets to descend into the domestic subtree. This is
generally a bad idea unless one knows precisely what they're doing,
but we'll provide the rope anyway (with the appropriate warnings).

Document both EXPORTABLE_SYSTEM and FORCE_DOMESTIC at the top of the file.

branches: 1.95.2;
Improve documentation.

Make a note of a subtle lossage mode pointed out by Mycroft.

Make the includes and install targets consistent, and shove the
ugly hack into the etc/Makefile distribution target instead of
the top-level build target. Problem noted by Mycroft.

Back out rev 1.90: The need for texinfo is a one-shot event. This was
announced in a reasonable forum. Further, it adds significant overhead to
the build process, and doesn't work correctly when a non-root DESTDIR
build is attempted (unless texinfo is already available, which of course
makes this change unnecessary).

make -> ${MAKE}

Build and install texinfo before building and installing the world.

Don't try to run makeinfo when doing preliminary GNU library builds.

Revert 1.83; it doesn't work. A different fix for his problem is in
the pipeline.

Add release and snapshot targets. (They do the same thing.)

convert from NOxxx= to MKxxx=no.
include <bsd.own.mk> if testing a MKxxx variable.

Back out the info dir changes; /usr/share/info was not in pkgsrc's info
path, which confused me to no end.

If updating from a system which has a /usr/pkg/info/dir, move that to
/usr/share/info/dir and make a symlink in /usr/pkg/info. This allows
systems with the gtexinfo pkg installed to work "properly".

Build the domestic libraries first, before building programs. Unfortunately
some programs like /usr/libexec/ftpd, need the kerberos etc libraries.
XXX: In reality programs like these should be moved to the domestic tree!
I would have assumed that if I make install in the exportable tree, I will
not install any domestic binaries!!!

Once more, from the beginning: we really, really don't want to descend into
the domestic tree if we are building an EXPORTABLE_SYSTEM.

Translate the conditional to English so as to avoid future problems
when people are tempted to modify it.

If doing a "domestic" build, make sure whatis.db is built _after_ the
domestic tree.

Sigh. The last change was too clever. Instead of using EXPORTABLE_SYSTEM,
which has side effects, use _BUILD to determine whether we ought to
install things in the domestic subtree. Otherwise, the core idea from
the last revision is unchanged.

When I split the build target into the top-level and "domestic" pieces, I
failed to account for how a 'make distribution' would get the proper bits
into the distribution tree. To correct this, we set EXPORTABLE_SYSTEM
while doing a 'make build' in order to avoid trying to install not-yet-
generated headers and binaries. I've also made the required adjustment
to the condition that adds "domestic" to SUBDIR.

This is too hairy and needs to die, but I don't see how it can until
we integrate the crypto bits into the main body of the tree.

Don't build domestic bits directly; use the domestic 'build' target
to do them for us.

Even if EXPORTABLE_SYSTEM is defined, descend into domestic for targets
obj, clean, cleandir and distclean. This reduces the chance of problems
(and the need to re-run make obj) if one is switching back and forth
between EXPORTABLE_SYSTEM and not, as US/Canadian developers tend to do.

Add a 'NBUILDJOBS' variable which control the number of jobs started in
parallel for a 'make build' (using make's -j flag). Only CPU-intensive jobs
are started in parallel.
Document the variables useable at the begining of the makefile
(NBUILDJOBS NOMAN NOSHARE UPDATE DESTDIR). Feel free to add documentation
for the ones I forgot !

GC old gcc 2.7.2.2.

Let make do the NOMAN= variable assignments instead of the shell so us
csh/tcsh weenies can cut and paste "make -n build" lines.

It is no longer necessary to run ldconfig in the build: target. It is
handled by bsd.lib.mk.

Only ldconfig if it's there (ELF systems don't use ldconfig).

Do not build libgcc 3 times.
Instead:
* If upgrading the compiler, build it *first*, before anything else.
* If upgrading the compiler and DESTDIR is set, barf.
This guarantees that libgcc is built during the library stage.

When doing a `make build' without DESTDIR, run ldconfig(8) after updating the
libraries.

Fix libgcc on alpha; shouldn't build 2.7 libgcc when bootstrapping.

don't install in /usr/share if NOSHARE is defined

Test gcc2.8 and egcs properly

No need to shield alpha from gcc 2.7 libgcc any longer. Default is egcs.

Only build the egcs libgcc if the current ${CXX} is actually egcs.

Re-enable libgcc from egcs and RE-ADD the warning & libgcc rebuild at the
end of the build - it is needed for gcc 2.7->egcs bootstraps!

Turn off the egcs bits while the 1.1 import is taking place.

Don't try and build non-EGCS libgcc on Alpha and PowerPC.

During the initial library pass, do not install man pages.
Also, GC the weird egcs stuff in here.

Do not blow away my egcs libgcc at the beginning of a build.

Fix flawed logic: I'm not all here today. (.if's werten't doing what the
commit comment said they would do.

Fix the rebuild of libgcc:
- If USE_EGCS is set, rebuild egcs's libgcc and install it
(unless DESTDIR is set and system compiler is not gcc 2.8,
in which case print a warning message and do nothing).
- Do not rebuild gcc 2.7's libgcc. egcs can build this fine.

build libgcc during make build; also, don't build lex and yacc first any more.

Revert 1.44 and associated changes; you can't prebuild anything without
includes, which is the case if you are building to an empty $DESTDIR.
Fixes PR 5497.

NOTE: This reverts us to the 1.3 behavior of requiring the secr
set to be installed for non-exportable builds. While unfortunate,
this allows a user to start with existing binary sets, rather than
requiring special knowledge of how to build a domestic tree to
$DESTDIR.

Move (re)build of libgcc to end of build list (after gcc/egcs would be
installed/updated).

Gack, don't try to build libgcc if it wouldn't be appropriate.

Build the appropriate libgcc, so that build-to-destdir works.

Make sure system includes are installed before descending into gnu.

Do the cleandir before the includes, so rpcgen doesn't build includes
twice. (Should build and install rpcgen before the includes, but it's not
too clear whether to do this before or after the cleandir.)

Print date after build completes as well as before. Previous code failed because make build depends on beforeinstall but not afterinstall.

fix make build for non-usa domestic a better way

don't assume domestic/usr.bin/compile_et exists for make build (eg: the user has a non-USA domestic subtree)

In the build targer, make sure the troff macros are installed before
attempting to build manual pages.

Domestic library build fix: libtelnet depends on libdes and libkrb. libkrb
depends on libcom_err. So, build libdes, libcom_err, libkrb in that
order before building domestic/lib. (Geez, this stuff is a mess.)

add date printout at beginning and end of 'make build's

Prebuild some things that the build depends on *before* they are needed so that fresh builds don't die prematurely

Make the "build" target depend on "beforeinstall", ensuring that all
directories have been created for the installation of the include files.

branches: 1.42.2;
Back out BUILDDIR and NOINSTALL changes.

Add a cleandir to make build, to make it more like it was before.

These updates to the build allow building against include files
and libs in the object tree, if you use a separate object tree,
while maintaining backward compatability with other build methods.
See the notes in src/share/mk/bsd.README for full details. Note
that the `make includes' target now only installs the include files
in the build directory (if you use one--otherwise they go in DESTDIR
just like before); `make install' will install include files in
DESTDIR.

Don't cleandir on make build when UPDATE is set.

when doing 'make build', build lib/csu & install it before building
lib itself. On the Alpha (and other ELF systems), shared library builds
require files built/installed by the csu build, so if lib/csu isn't
built/installed before lib is, DESTDIR builds won't work and normal builds
can mistakenly use old versions of some csu files.

All of domestic/lib now builds without errors, so we build it all.

allow builds to work if directory games is missing. Closes pr misc/1921 and
misc/2294

Headers in src/domestic/include are now installed by the includes target

Headers in src/include are now installed by the includes target

Added "${MAKE} includes" to build target

pull in bsd.own.mk earlier, to get configuration variables.
if people define EXPORTABLE_SYSTEM, they REALLY don't want to do stuff
in 'domestic'. Pay attention to it when doing 'make build'.
in 'make build', install share/mk _first_, because include dirs may
require new make templates.
in 'make build', install domestic includes after normal includes, so
they can previously-installed includes if necessary.
Note that all of domestic/lib should be built in 'make build' during the
library build stage.

behaviour of 'make build' should not change if DESTDIR is set; this is an artifact of an ancient mistake of mine.

a futile nod to ease-of-upgrade: make build now installs new rules from share/mk.

removed obsolete src/kerberosIV reference, albeit commented out

fix typo

libcrypt has moved

get krb4 includes installed, do distrib-dirs if DESTDIR not set

New-style RCS id.

Comment out last change as it interferes with distribution building.

stick distrib-dirs in beforeinstall -- should be cleaner?

work in the face of a /usr/src/obj directory

add 'domestic' to subdirs list, if it exists.

explicitly pull in bsd.own.mk, so that it's there for afterinstall.

if not installing man pages, don't remake whatis db

light style cleanup

don't use 'make' explicitly

kill typo

add some new targets to DTRT when installing/building

do "lib" first. otherwise you can be *really* screwed by a 'make install'
on a system with shared libraries...

Make regress only if directory exists. It can be optional.

allow toplevel makefile to descend into sys

change from rcs "Header" to "Id"

oops, clean that up a bit

add regress, but not for install. also, add "regress" target

put all of the gnu stuff in one place

added gnulibexec gnugames

added `usr.gnubin' for new gnu binaries necessary for the distribution.

added gnulib, for gnu libraries that for whatever reason need to go
into the main tree

change "Id" to "Header" in rcsid, and noted why etc was not in subdirs

added rcs ids to all files

after 0.2.2 "stable" patches applied

branches: 1.1.1;
Initial revision

more NBUILDJOBS deprecation

The NBUILDJOBS option was deprecated in 2002;
there's no need to keep warning about it,
remove from params / show-params,
and only document as obsolete.

build.sh: add show-params operation

Add show-params operation to the makefile target, and reword
params operation to describe what it's done since 2012.

(In 2012-11-04 the Makefile's previous params target was renamed
to show-params and a new params target was added to save a 'params'
file, and build.sh wasn't adapted.)

Deprecate extsrc/, EXTSRCSRCDIR, MKEXTSRC

Remove support for building extsrc/:
1. Makefile: remove do-extsrc target.
2. build.sh: remove options -y and -Y extsrcdir.
3. distrib/sets: remove support for extsrc in various tools
including the options -L ext and -y, and the extsrc sets.
4. doc/BUILDING.mdoc: remove docs for extsrc/, EXTSRCSRCDIR, MKEXTSRC (etc)
5. bsd.own.mk and various Makefiles: remove support for extsrc/,
EXTSRCSRCDIR, MKEXTSRC.

As proposed on tech-kern and tech-userlevel on 2022-01-07
and followed up on 2022-08-21.

Makefile: fix location of postinstall program for MAKEVERBOSE > 2

If MAKEVERBOSE > 2, each shell command from a make target is echoed.
This resulted in two additional words ending up in the variable
_POSTINSTALL. Noticed by Brad Harder.

Before:
$ make -v _POSTINSTALL MAKEVERBOSE=3
echo .../usr.sbin/postinstall .../usr.sbin/postinstall/postinstall ...

After:
$ make -v _POSTINSTALL MAKEVERBOSE=3
.../usr.sbin/postinstall/postinstall ...

G/c spurious whitespace.

automatically remove old debug files for modules.

Enhance the support of LLVM sanitizers

Define _REENTRANT for MKSANITIZER build. This is needed for at least stdio
code. This caused new build issued with duplicated symbols in few places
and rump kernel code picking different code paths borrowed from libc.
Handle all this in one go.

Add bsd.sanitizer.mk to share common code used by programs and libraries.

Switch from realall to beforeinstall target in .syms files. This is more
reliable in MKSANITIZER.

Correct the build rules of LLVM sanitizers

Fix generation of .a.syms files in LLVM sanitizers. The old rules weren't
functional and sometimes they were fatal in the build.

Add the .a.syms files to distribution sets.
These files are generated only in certain .a files in 64bit version of
sanitizers.

Correct the location of files with LLVM sanitizers. Move the LLVM sanitizer
entries from base to comp files.

Rework the LLVM compiler_rt build rules

Register external/bsd/compiler_rt as a _SUBDIR. This will allow to handle
properly includes specific for compiler_rt.

Centralize build_install target for external/bsd/compiler_rt.

With the above changes MKOBJDIR is now respected properly for me.

Do not install LLVM sanitizers into TOOLDIR for !HAVE_LLVM.

Enhance MKSANITIZER support

Add i386 support for LLVM sanitizers.

Add new build phases do-sanitizer-* for building LLVM sanitizers after libs
and before other programs. This phase is important as sanitizers need
to link with libs and sanitized programs must link with sanitizer runtimes.

branches: 1.327.2;
Use the generated "postinstall" in the objdir for postinstall-check
and postinstall-fix targets - we do not have a simple postinstall script
in the source tree any more.

Fix editor mishap

.WAIT for tools to complete.
There's a lot of waits inside tools, but nothing that promises that we haven't
started doing the next subdir in the top level.

Shows up as warnings in the cleandir stage on releng builds:

cleandir ===> external/mit/xorg/server/drivers/xf86-video-suncg6
cleandir ===> lib/libpam/modules/pam_deny
sh: /home/builds/ab/HEAD-llvm/sparc64/201811091720Z-tools/bin/nbawk: not found
nbmake[9]: "../../Makefile.xf86-driver" line 23: warning: "/home/builds/ab/HEAD-llvm/sparc64/201811091720Z-tools/bin/nbawk '/^PACKAGE_VERSION=/ { match($1, "[0-9]+\\.[0-9]+\\.[0-9]+"); version = substr($1, RSTART, RLENGTH); } END { print version }' /home/source/ab/HEAD-llvm/xsrc/external/mit/xf86-video-ati/dist/configure" returned non-zero status
sh: /home/builds/ab/HEAD-llvm/sparc64/201811091720Z-tools/bin/nbawk: not found

branches: 1.324.2;
Re-add files that were accidentally deleted in my previous commit.

Synch with HEAD

include bsd.clean.mk so that we actually clean up the attempted "params"

branches: 1.321.2;
Include somewhat misleadingly named do-x11 into BUILDTARGETS even with
NOBINARIES set (subject to MKX11).

do-x11 target builds and installs only the tools and libraries, X11
programs are built during the regular recursive build in extsrc. Edit
its description accordingly and move it before do-build to match their
order in BUILDTARGETS.

While here, drop the "either" clause that has lost its "or" a few
years ago when xfree86 reachover was removed.

Add NOBINARIES, useful to build tools are libraries which is what's needed
for mknative.

don't descend into gnu

branches: 1.317.2; 1.317.4;
fix direct reference to texinfo, bleh

remove the xfree86 reachover makefiles and the vast majority of
the support in the rest of the source tree.

X11 sets could use some cleaning up perhaps (just deletion, as
we've never really marked the old X11R6 as obsolete for native
xorg using platforms so far either.)

Add MKCOMPATX11 knob for making X11 compat libraries.

Support building the xorg X11 libraries under MKCOMPAT.
Add XORG_MACHINE_ARCH to compat's bsd.*.mk
Turn xorg makekeys and makestrs into real host tools which are installed into
${TOOLSDIR}

PR 49870: pass the xsrc path to postinstall

Reorganise the way xorg is build. Move build of makestrs into src/tools
as it is needed for the "includes" phase. Make it symlink all necessary
headers in obj to decouple it from "includes" itself, breaking the
dependency cycle. Move the do-x11 target between do-lib and do-build, so
that libraries can get the benefit of build_install (correct .WAIT
behavior) and everything else like "includes" uses the plain SUBDIR
entry in external/mit.

iso -> images

remove obsolete BUILD_CC_LIB_BASEDIR and BUILD_CC_LIB_BASETARGET vars

branches: 1.309.2;
Run "postinstall fix obsolete_stand" just before checkflist,
if we are building with DESTDIR != /

Pass -m MACHINE -a MACHINE_ARCH to postinstall

Pass SED=${TOOL_SED:Q} to postinstall.

remove the build support for GCC 4.1, and any hacks i found for it.

there could still be more -- i only looked for lines matching
both "HAVE_GCC" and "4" as words (grep -w.)

branches: 1.305.2;
prepare for moving GCC 4.5 into gcc.old:

- convert to using ${EXTERNAL_GCC_SUBDIR}
- define base-external-gpl3-gcc* subdir as GCC_SUBDIR
- use <bsd.init.mk> over <bsd.own.mk> for a bunch of places; mostly
because it arranges for ../Makefile.inc to be included earlier, and
don't bother including the latter if the former is already included.
- move all .PATH: settings after <bsd.{own,lib}.mk> so that all
valid variables are set before it is evaluated
- rename mknative-gcc* to match their subdir name.

XXX the relationship between the Makefile.inc/Makefile.gcc_path files
is kind of sketchy, it would be great if this was fixed.

eliminate use of bsd.sys.mk from Makefiles

hp700 rescue/ldd fails on -j16 build if usr.bin/ldd/elf32/libldd_elf32.a
hasn't been made yet. Add an explicit .WAIT before rescue to avoid this.

Add src/etc/Makefile.params, containing the definition of the
RELEASEVARS variable, and commands related to printing the values of
the variables whose names are in RELEASEVARS.

Add an awk script to remove noise printed by "make -j" or high levels
of MAKEVERBOSE, so we get only the variables names and values. The
values are escaped so that variables containing embedded newlines,
quotation marks, and backslashes, are passed through safely.

Adapt src/etc/Makefile and src/Makefile to use the new ${PRINT_PARAMS}
command defined in src/etc/Makefile.params.

Now ${DESTDIR}/etc/release and the params file in the top-level
.OBJDIR should never contain unwanted noise, even after a build with
MAKEVERBOSE=4.

Move the build_install logic from lib/Makefile into share/mk and re-use
it for tools. The existing logic broke for the LLVM build with the
recent .ORDER changes in make.

Create a "params" file in the top-level .OBJDIR during "make build".
This file will contain the values of all variables that can affect
the build process. Take care not to update the file's timestamp if
the contents do not change.

${_SRC_TOP_OBJ}/params is referenced by ${_NETBSD_VERSION_DEPENDS} in
bsd.own.mk.

Rename the old "params" target to "show-params". This simply
prints the values of the variables that can affect the build.

branches: 1.299.2;
Implement build_install in src/lib/Makefile, which traverses the
subdirectories and invokes dependall/install. Do this in groups
delimited by .WAIT to ensure that depending libraries can use the
installed versions and don't need to know the locations in the source
tree.

Use this new target in src/Makefile to replace most of the adhoc library
logic with two special cases, src/lib and src/compat. Adjust
sys/Makefile to include the module directory when building them. Add
some necessary .WAITs in src/lib/Makefile to reflect the dependencies
from src/Makefile and also add the rump libraries here.

Add HOST_SH to _POSTINSTALL_ENV; needed for
usr.sbin/postinstall/postinstall revision 1.145.

Pass STAT=${TOOL_STAT:Q} through to postinstall.

TOOL_GREP is defined in share/mk/bsd.own.mk, but is not actually built
from tools/Makefile. So don't try to use it here.

Pass all relevant tool variables to postinstall, not just a subset.

Fix unbalanced quotes in .if ${MKRUMP} != "no"

exclude rump targets if MKRUMP = no

Add usr.bin to the exclude list for make includes.
This is no longer needed after flex was moved to the external framework.

branches: 1.291.2;
s/is the not root/is not the root/ in a comment - from Snader_LB

Add "live-image" and "install-image" target support to build.sh.

"live-image" target builds pre-installed disk images that can be used on
emulators or boot from USB memory sticks to try NetBSD without installation.
Currently amd64, i386, pmax, sparc, sparc64, sun2, sun3, and vax
(which have working emulators and don't require extra tools like preparing
msdosfs or partitioning MD label structures) support this target.

"install-image" target builds an bootable installation disk image that can
be used as an install CD but burned into USB memory sticks etc.
Current only amd64 and i386 (which would support USB boot) have this target.

For more details (and known issue) see following posts on netbsd-bugs@:
http://mail-index.NetBSD.org/netbsd-bugs/2011/08/06/msg023639.html
http://mail-index.NetBSD.org/netbsd-bugs/2011/09/23/msg024207.html
http://mail-index.NetBSD.org/netbsd-bugs/2011/12/07/msg025166.html
http://mail-index.NetBSD.org/netbsd-bugs/2011/12/08/msg025178.html

No particular comments about implementation, and
"go ahead" comments from mrg@, riz@ and christos@.

Closes PR toolchain/45153 and PR misc/45155.

branches: 1.289.2;
Move the code for creating src/BUILDING from "make build-docs" in
src/Makefile to "make regen" in src/doc/Makefile, and update a comment
in doc/BUILDING.mdoc to match.

Fix typo that prevented installmodules from working with -U

add an installmodules=<dir> command to build.sh

re-enable the gcc 4.5 crtstuff/libgcc builders now that they work.
also build the gmp, mpfr and mpc libraries in this case.

fix the parts that build libgcc/crtstuff, etc., for GCC 4.5.

remove all the code that supported HAVE_GCC=3. mostly from chuq.

build evbppc modules in 3 different flavours:

- oea (default)
- ibm4xx
- booke

this, together with earlier machine_module[] work, should allow evbppc
platforms to use modules.

branches: 1.282.2;
Pass down INSTALLSETS variable to installsets target. Use :Q to escape
shell meta-characters, as the variable contains a list of sets, separated
by spaces.

$ ./build.sh -V INSTALLSETS="base etc" install=idir

should now work as expected.

Reviewed by apb@ on tech-toolchain.

branches: 1.281.2;
Remove the do-external-lib and do-gnu-lib targets, along with
external/lib/Makefile and crypto/external/lib/Makefile, replacing
them all with SUBDIRs directly from lib/Makefile.

compat/compatsubdirs.mk becomes simpler now, as everything is built
from lib/Makefile, meaning all the libraries will now be built under
compat so update the set lists to account for that.

Centralise USE_COMPILERCRTSTUFF.

PCC might need help.

no need to create a do-lib-libdes target, the directory no longer exists

remove almost all the ability to build netbsd with an a.out target.
we're ELF now, and there are many missing checks against OBJECT_FMT.
if we ever consider switching, the we can figure out what new ones
we need but for now it's just clutter.

this doesn't remove any of the support for exec_aout or any actually
required-for-boot a.out support, only the ability to build a netbsd
release in a.out format. ie, most of this code has been dead for
over a decade.

i've tested builds on vax, amd64, i386, mac68k, macppc, sparc, atari,
amiga, shark, cats, dreamcast, landisk, mmeye and x68k. this covers
the 5 MACHINE_ARCH's affected, and all the other arch code touched.
it also includes some actual run-time testing of sparc, i386 and
shark, and i performed binary comparison upon amiga and x68k as well.


some minor details relevant:
- move shlib.[ch] from ld.aout_so into ldconfig proper, and cut them
down to only the parts ldconfig needs
- remove various unused source files
- switch amiga bootblocks to using elf2bb.h instead of aout2bb.h

descend into sys/rump/kern

adjust the do-compat-* rules now that src/compat/lib and friends no longer exist

Support "extsrc", externally added programs and libraries. Users can write
their own reach-overs, cross-build, install, and get set files just like base
and X11 / X.org. (These sets are not included as TNF releases.)

build & install librumpdev_netsmb and librumpfs_smbfs components

Let us use MKKMOD=yes to leave the kernel modules out of the
distribution.

Build modules before rump libs so that something else can be
unwarrantedly blamed for a change.

This commit follows the style guide and is entirely non-build-tested.

add crypto/external to the build.

Make clean_METALOG in distrib/sets early in the build, to ensure that
the METALOG does not grow without bounds on multiple MKUPDATE builds.
(Previously, sanitise_METALOG in distrib/sets/Makefile used to do
something like this near the end of a build.)

Remove an explicit BUILDTARGETS do-compat-external-lib before do-build
because now src/compat/Makefile descends there and no particular
dependency on 32 bit compat external libs during do-build.

Disable grep host tool and unbreak NetBSD build until I can come up with a
clean fix.

Use awk and grep host tools where required. 'build.sh release' now works
on Solaris (but only with HOST_CC=/usr/sfw/bin/gcc for now).

branches: 1.266.2;
Add the top-level do-compat-external-lib target that mimics
do-external-lib so that we can build the compat libraries for
ATF. Also add compat/external/lib/Makefile following the same
example of external/lib/Makefile, which descends into the
external libraries and builds them.

Build kernel modules the same time as the rest of the system libraries.

Invoke postinstall with the environment variable MAKE set up to the relevant
value, so that postinstall picks up correct values.

- retire MKXORG from public use
- introduce X11FLAVOUR to choose src/x11 vs src/external/mit/xorg
for the X11 to build if MKX11=yes is set. it takes the values
of either Xorg or XFree86.
- default to Xorg on alpha, i386, macppc, shark, sparc and sparc64
- remove MKXORG_WITH_XSRC_XSERVER, unused and never useful

branches: 1.262.2;
- enable 32-bit library builds on amd64 and sparc64

- add 3 new top-level targets:
do-compat-lib-csu
do-compat-libgcc
do-compat-lib-libc
and run them before "do-build" if ${MKCOMPAT} != "no"

Add a do-top-obj target, which builds the top-level obj directory.
This uses the new NOSUBDIR support in bsd.subdir.mk.

Insert do-top-obj near the beginning of BUILDTARGETS. The first few
items are now:
cleandir: cleans the tree.
do-top-obj: creates the top level object directory.
do-tools-obj: creates object directories for the host toolchain.
do-tools: builds host toolchain.
obj: creates object directories.

Also improve a comment and split a long line where SUBDIR is set from
_SUBDIR.

Defer the "make obj" step until after tools are built and installed.

There may be commands embedded in "!=" assignments that depend on tools
being available, and previously these would have printed error messages
during "make obj". They may still print error messages during "make
cleandir", which (if it is performed at all) is performed before tools
are built.

add do-sys-rump-net-lib

PR/39687 - Henning Petersen -- comment typo

Improve support for MKXORG!=no to enable builds into external/mit/xorg.
This will be mutually exclusive to MKX11!=no.

Build pcc and pcc-based crtstuff if using pcc compiler.

Install rump libraries and utilities to the base system and remove the
private non-installed build infrastructure from sys/rump.

breakdown of commit:
* install relevant headers into /usr/include/rump
* build sys/rump/librump/rumpuser and sys/rump/librump/rumpkern
from src/lib and install as librumpuser and librump, respectively
+ this retains the ability to test a librump build with just the
kernel sources at hand
* move sys/rump/fs/lib/libukfs and sys/rump/fs/lib/libp2k to src/lib
for general consumption, they are not kernel-space dwellers anyway
* build and install sys/rump/fs/lib/lib$fs as librumpfs_$fs
* add chapter 3 manual pages for rump, rumpuser, ukfs and p2k
* build and install userspace kernel file system daemons if MKPUFFS=yes
is spexified
* retire fsconsole for now, it will make a comeback with an actually
implemented version shortly

if MKXORG != no, use external/mit/xorg subdir instead of x11 subdir for
the mk-x11 target.

Descend into external.
Add 'do-external-lib' target to build external/lib "early"

branches: 1.252.2; 1.252.4;
Consistently document RELEASEDIR/RELEASEMACHINEDIR.

branches: 1.251.4;
Add the tests.tgz set

This adds a new tests.tgz set to releases which includes all the tests
for the system. It is important to note that this set does not rely on
comp.tgz: a user of the system can run the tests without having the
development tools installed, which can be useful in a production machine.

Do not abort on exit code 1 from postinstall check, this is expected.
Reported, fix suggested and tested by David Holland.

(Why does "make build" invoke postinstall check on destdir is another
question, it does not seem to have much sense.)

Provide 'do-libgcc' to build gnu/lib/crtstuff${LIBGCC_EXT} (if necessary)
and then gnu/lib/libgcc${LIBGCC_EXT}.
Makes it easier to cut&paste from the list of targets built with 'make build'.

Add an iso-image-source option which does the same thing as iso-image but
includes source sets as well. The infrastructure for this was already
there, although it needed a bug fix. Will look at adding arbitrary
directories next.

branches: 1.247.4;
Add an iso_image target (iso-image as a target name doesn't play well with
bsd.subdirs.mk) in distrib/makefile, which builds an iso image for $MACHINE
with binary sets, stored in ${RELEASEDIR}/iso. The image is bootable for:
alpha, amd64, cats, i386, pmax, sgimips, sparc, sparc64, sun3, vax.
mac68k/macppc no there yet because of missing feature in makefs.

call iso_image in distrib/ for iso-image in the top Makefile.