hook gzip to the build so that we can test it in other OS's. Not used yet.

tools/elftoolchain: Handle dependencies properly.

This should properly resolve the problem christos was trying to
address by a bogus make includes hack and #ifdefs in elftoolchain.

tools/Makefile: clean up TOOLDIR validation

At the point where tools/Makefile is loaded, TOOLDIR doesn't have to
exist yet, it will be created later. Remove the redundant quotes from
the .error messages.

build isl tool for GCC 12.

tools/Makefile: Correct inverted comment. NFC

Disconnect 'tools/lib{elf,dwarf}' from the build.

These libraries are now built by 'tools/elftoolchain/lib{elf,dwarf}'
respectively.

When building tools use a dedicated subdirectory for compiling
Elftoolchain code.

Build and install <sys/elfdefinitions.h> in ${TOOLDIR}/include.

This change prepares the ground for upgrading libelf and libdwarf to
their current Elftoolchain versions.

Make dtc on riscv

If we are building date earlier, we need to use an uninstalled copy
of host-mkdep and libnbcompat otherwise the build process looks for
tools that might not be there yet (e.g. fresh build).
Reviewed by martin.

Add a date tool

branches: 1.209.2;
PR port-mvme68k/56146

Build wrtvid as tools.

Make mips-elf2ecoff available for all mips

tools: allow MKCROSSGDB to work when not building gcc

Build mkubootimage for evbsh3 too.

Update LLVM to 10.0.0git (01f3a59fb3e2542fce74c768718f594d0debd0da)

Exclude more subdirs if TOOLS_BUILDRUMP=yes.

Exclude subdirs added in the last 3 years but not need for rumpkernel builds.

Fix build, "pax" must be built before "libprop".

Ok: Matthew Green mrg@

Add support for libprop as a host tool library.

Add MKLLVMRT to build a subset of the LLVM stack for JIT purposes.
Hook up AMDGPU backend for clang.

Revert MKLLVMCMDS changes.

split MKLLVM into MKLLVM and MKLLVMCMDS, like MKGCC and MKGCCCMDS:

- only build clang or clang-tblgen if MKLLVMCMDS != "no"
- move HAVE_MESA_VER earlier so we can detect mesa 18 and enable
MKLLVM on platforms that need it
- move the clang commands and headers into llvmcmds group
- only build the clang specific LLVM libraries if MKLLVMCMDS != "no"

much of this work came from maya@'s github commit
f90685c11d4460d3098fa35f48b58d1893e974e0, reworked for HAVE_MESA_VER
and llvmcmds vs MKCLANG.


build tested on several platforms and on amd64 and i386 with several
different build options (clang build, gcc build, gcc build with llvm
libs).

use MACHINE_CPU for elf2aout

Add an arm-specific elf2aout tool since objdump does not support a.out anymore

Update build glue for xz-5.2.4. Hook up tools version.

hook in the tools version of grep

branches: 1.194.2;
For EXTERNAL_TOOLCHAIN, MKLLVM=yes needs to build only tablegen.

branches: 1.193.2;
Fix .endif comment for TOOLS_BUILDRUMP

Better EXTERNAL_TOOLCHAIN support from Zachary McGrew
- Allow mdsetimage and dbsym to be built with EXTERNAL_TOOLCHAIN
- Allow BFDDIR and IBERTYDIR to be overwritten for mdsetimage and dbsym

branches: 1.191.2;
Hook in dtc/libfdt now that it builds for me

Disable dtc/libfdt for now

No need to set dtc twice.

Add dtc

branches: 1.187.6;
Sync tools build glue with LLVM pre-4.0.0.

branches: 1.186.2;
move cvslatest last

Tool version

move sortinfo after pax

Add sortinfo

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.)

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}

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.

build genassym for TOOLS_BUILDRUMP

add gpt

Changes to existing files to enable building AARCH64 userland.
evbarm64-el
This is clang only. While gcc4.8 supports aarch64, no netbsd support has
been written for aarch64 with gcc4.8.

descend to vgrind

Add cat, rpcgen and stat to TOOLS_BUILDRUMP. They're useful for
installing headers for the app stack cases that need it done, and small
enough to not matter for the kernel-only cases that don't.

remove remaining makefile support for GCC < 45 that i found.

branches: 1.173.2;
ctfconvert/ctfmerge need libdwarf &c., so build them if MKCTF=yes.

Add MKCTF

This variable separates CTF stuffs from MKDTRACE; we can build DTrace
solely without building and using them. This allows us to use DTrace
even if CTF stuffs have problems (actually they have now).

This variable would be merged into MKDTRACE eventually, once CTF stuffs
work correctly again.

Rename NetBSD/hp700 to NetBSD/hppa.

Unfortunately our VCS isn't very helpful here.

Update LLVM/Clang snapshot to r189662. This includes many improvements
to the new SLP vectorizer, fast codegen support for PPC and using
terminfo for the color sequences. In addition, add (optional) build
infrastructure for lld to allow easier testing. This is controlled by
MKLLD.

Even if we aren't making a crossgdb still make the object dir for it.

branches: 1.168.4;
Add infrastructure for building MCLinker.

There are other little endian arm machine_arches other than "arm"
Use MACHINE_CPU and TARGET_ENDIANNESS instead of MACHINE_ARCH

Use TOOLS_BUILDRUMP=yes to restrict building of tools to those required
for building rump kernels. This cuts down the NetBSD source tree required
by buildrump.sh to a fraction and addresses a size/download time concern
from a person packaging rump kernels for Linux.

No change to the normal case (tested with several -j32/-j64 builds).

reviewed by apb

Make tools/Makefile build host-mkdep, compat, and binstall in that order,
before installing any of them. It installs them in any order relative
to each other. The .WAIT points that used to be between these three
tools are replaced by .ORDER and dependencies.

This is needed in preparation for making tools/compat install
some files into TOOLDIR.

TOOLDIR must be valid, unless MKTOOLS=no.
Add tests and .error directives to enforce this.

If tools are not wanted, then set SUBDIR= empty instead of
defining all the subdir-related targets to do nothing. This way,
the targets are not invoked at all, instead of being invoked but
doing nothing.

remove trailing white space

According to the i.MX23 Reference Manual section 3.1, Page 3-3:
"The i.MX23 always operates in litle-endian mode."
So build elftosb and bootloader for this processor only for evbarm(-el).

Build gmake iff we're going to need it.

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.

Add elftosb tools to evbarm userland and toolchain.
Elftosb is used to create a digitaly signed "secure boot" file.
This sb file can be booted by the first stage boot loader found in
Freescale i.MX23 and i.MX28 application processors.

Copyright (c) 2004-2010 Freescale Semiconductor, Inc.

fix EXTERNAL_TOOLCHAIN build

branches: 1.156.2;
Decouple binutils build from HAVE_GCC. Merge common fragments of
HAVE_GCC and HAVE_PCC.

Revert .WAIT removal and add comment why it is needed between *tblgen.

Do not build tic more than once

move tic later, after all the tools it needs are built.

tweak dependencies for tic, which needs nbperf, awk, and sed.

branches: 1.151.2;
Remove an unnecessary .WAIT between llvm-tblgen and clang-tblgen.

Hook up nbperf into the normal tool build.

Update LLVM/Clang snapshot to r141658. This primarily fixes a regression
preventing Qt4 from building.

Build nbm68k-elf2aout as a host tool on m68k targets
for an ancient PROM bootloader.
Mostly taken from nbmips-elf2ecoff.

install elf2ecoff for all mips64 systems. it's needed for stand.

always build gmake if HAVE_GCC.
build gmp, mpfr and then mpc if HAVE_GCC >= 45.

branches: 1.145.2;
Upgrade Heimdal to 1.5pre1 by switching the build from crypto/dist/heimdal
to crypto/external/bsd/heimdal. The latter was just imported as the head
of the Heimdal tree as of a few days ago.

.WAIT for makefs, because in most builds, it's the first target
which includes Makefile.nbincludes, whose beforedepend: has
some race condition I don't fully understand, and which has been
causing many (but not all!) -j16 builds on the autobuild cluster
to fail in tools since sometime between 201103061600 and 201103070900.

With the proper voodoo, I was able to get 8 builds in a row to fail
without this change, and 16 in a row to succeed with it.

XXX a proper fix likely involves reworking the beforedepend: target
of Makefile.nbincludes, but this should get the build cluster back up
to full functionality in the meanwhile.

mkubootimage is certainly useful for sandpoint

Add LLVM and clang build system. It is enabled by MKLLVM=yes.
Upstream sources can be fetched by running "make checkout" in
src/external/bsd/llvm, they will be properly imported once the
integration and missing features are sorted out.

Make mkubootimage for evbmips and evbppc too.

branches: 1.140.2;
add new option to build/use tools/pigz for compressing sets. for now
USE_PIGZGZIP defaults to "no". (it depends upon local <pthread.h> and -lz.)

you can set it to "yes" on most modern platforms just fine.

hook mkubootimage into the build

Bye bye atf-compile, you won't be missed.

DTrace: add libelf, libdwarf, libctf, and CTF tools to the toolchain,
built when MKDTRACE=yes.

Build mandoc unconditional if man pages are desirable. All
compatibility issues are supposed to be sorted out.
Build groff conditionally on MKGROFF. It is currently the only component
that requires C++.

Store our generated files in CVS and remove nbperf from the toolchain.
This allows us to build consistent libs without an ever changing hash.

Userland now builds and uses terminfo instead of termcap.

OK: core@, jdc@

To unbreak the build, temporarily remove the build of libelf here,
since no other part of the build depends on this yet.

Suggested as a temporary workaround by thorpej@

rework the main SUBDIR list here as suggested by enami. put .WAIT on
the same line as the target that needs to wait, but before the target,
rather than on the previous line, which led to confusion and an error
being introduced in the previous change.

move the build of pax before libelf. fixes my build of tools/libelf,
though i didn't look to see why libelf needs pax.

Add support for building libelf in a host-tool environment.

XXX Makefile.disklabel is really mis-named, but I'm not going to tackle
that problem right now.

Give m4 its own .WAIT. Fixes build issue on NetBSD-5.

Add MKMANDOC option to trigger building cat pages and HTML pages with
mdocml. Add a lintmanpages target for developers to check man pages
with the strictest settings of mandoc.

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

Make grep a host tool.

Convert more MACHINE tests to MACHINE_ARCH

branches: 1.124.4; 1.124.8; 1.124.12;
Build sed earlier, just after mktemp. Note that sed and mktemp do not
depend on each other, but the .WAIT is needed to get them both installed
before the remainder of the tools build.

Remove the local handling of the NOSUBDIR variable. We can rely
on bsd.subdir.mk's handling of the NOSUBDIR variable.

Also add some comment to .if/.else/.endif and .for/.endfor lines.

Add awk as a host tool.

Per uwe@'s suggestion, build nbgroff under wider circumstances.
I choose to build it if MKDOC != no or if MKHTML != no. It used
to be built only when MKMAN != no.

Add a .WAIT to ensure that sed is installed early, so that TOOL_SED can
be used during the build of other tools.

Build strfile both as a host tool and as an installed program:
* in games/fortune/strfile/Makefile, build strfile as a
regular program instead of as a host tool;
* add tools/strfile directory to build strfile as a host tool;
* in tools/Makefile, add strfile to SUBDIR list;
* in BSD.*.mk, define TOOL_STRFILE variable;
* in games/fortune/datfiles/Makefile, use TOOL_STRFILE when creating
databases at build time;
* in distrib/sets/lists/games/mi, mention usr/games/strfile.

Build pcc as host tool if requested.

Remove references to bebox-elf2pef and bebox-mkbootimage, now that they
have been integrated in common powerpc tools and subsequently removed.

branches: 1.116.2;
Build powerpc-mkbootimage for rs6000 and bebox too.

Convert prep to use the new unified mkbootimage rather than it's own. This
mkbootimage can be shared by prep, rs6000, and bebox.

TODO: Convert bebox to use this. Needs someone to test.

branches: 1.114.6;
Add the atf host tools

This adds reachover Makefiles to build the atf tools required during a
cross-build.

branches: 1.113.4;
Build join(1) as a host tool.

A commit too far - don't try to build veriexecgen for just now

Build veriexecgen as a host tool

Make sys/exec_elf.h and machine/elf_machdep.h available to host
tools, and allow paxctl to use them when compiled as a host
tool. This fixes the build on MacOS X (and perhaps other platforms)

Add some notes about the bogosity of some of the .WAIT stuff.
Remove a 'never had a sane effect' .WAIT from a .ORDER line.
Don't add a (probably incorrect) dependency for 'make install' since
it contradicts a .ORDER line elsewhere when .ORDER applies recursively
to child nodes.
I also think that the stuff that tries to add dependencies between
the stuff that bsd.subdir.mk generates does not DTRT at all.
I suspect that build.sh builds things in an order that avoids issues here.

Add paxctl as a tool.

If PREVIOUSTOOLDIR was empty, then don't complain that it has moved.

Patch from Jukka Salmi in PR 33095.

* Fix parallel build race condition between real{all,depend} and
<bsd.subdir.mk> targets.
* No need to pass TOOLDIR to recursive cleandir in real{all,depend}.

With help from he@.
OK by tron@.

Another attempt to fix the race in parallel builds.

Invoking "make cleandir" with "PREVIOUSTOOLDIR" set to the previous
location will not stop recursive "make cleandir" calls. Pass the new
location instead.

gettext should .WAIT for autoconf.
(The former still doesn't seem to work on read-only source; separate issue)

Avoid getting into a loop of "make cleandir" instances if the location
tool directory has been changed or wasn't recorded before.

Remove ".WAIT" again because it doesn't fix the problem.

Add an extra ".WAIT" to the "realall realdepend" target to make sure
it finished cleaning the tools directory in case of tool directory
mismatch before we try to rebuild the tools.

only build lint and lint2 if MKLINT != "no".

Make TOOL_SED, to allow using our own sed in cases where the host's version
is busted in some way.

The TOOL_SED doesn't use TIOCGWINSIZE, and has some changes to include
nbtool_config.h in a cross build environment.

Combined with some other changes (not yet committed), this allows cross
compiling m68k code on Solaris 10/amd64.

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

check HAVE_GCC == 4 not USE_GCC4.

Don't barf if USE_GCC4 is not defined.

Turn on gmake builds if GCC4 is enabled (and force it's install before continuing

The realall and realdepend bit that regenerates tools needs to be marked
.MAKE otherwise the 'make' it runs has its own list of jobs instead of
using the jobserver of the enclosing make.

Make the tools using Makefile.disklabel wait so they don't stomp each other.
Also for now add a dependency to Makefile.disklabel so this doesn't run
3 times. (XXX this really should just be done differently)

.include Makefile.disklabel in installboot, so that the
nbinclude/ header files will be installed. Build installboot after
pax, since Makefile.disklabel uses pax to install the headers.

Make disklabel(8) and fdisk(8) into "host tools " last step: build
and install ${TOOLDIR}/bin/${MACHINE_GNU_PLATFORM}-disklabel,
${TOOLDIR}/bin/${MACHINE_GNU_PLATFORM}-fdisk by "reaching over" to
the sources in ${NETBSDSRCDIR}/sbin/{disklabel fdisk}/.

To avoid clashes with a build-host's header files, especially on
*BSD, the host-tools versions of fdisk and disklabel search for
#includes such as disklabel.h, disklabel_acorn.h, disklabel_gpt.h,
and bootinfo.h in a new #includes namespace, nbinclude/. That is,
they #include <nbinclude/sys/disklabel.h>, <nbinclude/machine/disklabel.h>,
<nbinclude/sparc64/disklabel.h>, instead of <sys/disklabel.h> and
such. I have also updated the system headers to #include from
nbinclude/-space when HAVE_NBTOOL_CONFIG_H is #defined.

Build genassym as a host tool. Part of PR toolchain/30350

We don't need aout2bb any longer.

Add some quotes to prevent a syntax error.

Convert mkboot into a full cross tool. Last hurdle for hp300 to building on
non NetBSD/bsd hosts.

Convert remaining host tools for amiga to correct cross tools. Now allows
complete amiga builds on non netbsd/bsd hosts

Convert bebox's remaining host tools over to cross tools scheme. Builds to sets
w/o error now.

Build hp700's mkboot as a host tool called nbhp700-mkboot. mkboot also
gets built and installed in a hp700 distribution.

TODO

- merge with hp300

- pick a preferred method for dealing with the elf headers.
hp700-mkboot and prep-mkbootimage (bintuils) vs mips-elf2ecoff and
tools/installboot

Add sparkcrc (used to generate checksums for acorn32 bootloader sparkive)
as a host tool.

branches: 1.81.2;
Consistently check ${TOOLCHAIN_MISSING} against "no".

Fix typo.

Add macppc-fixcoff tool.

Use (cd ${.CURDIR} && ${MAKE} cleandir) instead of ${MAKE} cleandir
Suggested by Ross Harvey.

always build tools/binutils now.

Build sgivol as a host tool when compiling for sgimips.

add mknod as a host tool

add gdb53 as a host tool for use as a crossgdb
replaces gdb50 from toolchain on archs which use gcc3

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.

Add host tool version of ncdcs

MAKEVERBOSE support

- build "binutils" and "gcc" separately if USE_TOOLS_TOOLCHAIN is not set
to "yes".
- adjust dbsym/mdsetimage to look in the correct place for libbfd.
- new mknative-gcc based in toolchain/mknative

Add missing "cleandir" rule.

add stat(1) as a host tool

also do the MKTOOLS and PREVIOUSTOOLDIR checks for "realdepend"

* If a `PREVIOUSTOOLDIR' file doesn't exist, or its contents are not
the same as "$TOOLDIR", run "make cleandir" here first.
This should prevent various problems where $TOOLDIR is encoded into the
build process for various tools (lint, groff, texinfo, toolchain),
$TOOLDIR changes, and various programs don't get rebuilt correctly.
Should fix PR [toolchain/21988].

* Clean up how the `${MKTOOLS} == no' check is performed.

* Remove "cleantools"; it made a big assumption about the location of
TOOLDIR that won't be correct in many cases.

Citrus iconv support(3)
add host tools : mkcsmapper and mkesdb
mkcsmapper - coded character set mapping table compiler
mkesdb - encoding scheme database compiler

Add db(1) as a host tool

build file(1) as a host tool (for compiling magic files)

Make hexdump a host tool.

enable sunlabel...

Change the TOOLCHAIN_MISSING from == "no" to != "yes", to be consistent
with the other tests of this variable.

Since TOOLCHAIN_MISSING is defined in bsd.own.mk, it can never be undefined.
Instead, make sure the value of it is "no"

* 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.

We need binstall to be installed, before attempting to use it to install mktemp.

Descend into mktemp.

The last commit wasn't tested. Add a missing \

Don't build toolchain bits for hppa, for now.

Add cat(1) as a host tool.

Don't try to build toolchain bits for ns32k.

Add fgen as a host tool, used for sys/arch/sparc/bootblk.

branches: 1.50.2;
gratuitious comment addition

remove now-unnecessary sparc-installboot

On prep, descend into prep-mkbootimage.

vax-installboot has been deleted, so don't recurse into it.

enable installboot

Nuke NO_DBSYM; dbsym now uses BFD for everything.

On second thoughts, don't make mktemp a host tool. It's only used once, for
creating a temporary directory for maketars, and mkdir can do that safely
enough.

Add mktemp as a host tool.

Add cksum as a host tool.

Drop into vax-installboot on VAX machines.

Rework recursion to use the original logic again, in the structure cleaned
up by pk (and fix comments to match):

* A .WAIT here means that only the *immediately preceding* target needs
to be added to the list of "install-foo" targets used to ensure that
dependencies are installed (otherwise much more than is necessary
gets installed during the build process, increasing build time here)

* If building only an individual target such as "dependall-yacc", the
prerequisites may need to be built before running their "install-foo"
targets -- so add a dependency in that case on "dependall-foo".

Don't build groff if MKMAN=no. (groff does not build properly on some
platforms because of its completely insane Make structure; this allows it
to be disabled in a sane way.)

Add gettext to tools.

For now this is a maintainer-tool only; none of the NetBSD build needs to
run msgfmt as of yet. (However, its presence is needed to make the
mknative part of src/tools/toolchain produce the same output as a truly
"native" GNU configury of the toolchain.)

Add elf2ecoff as a build tool for mips ports.

* Add comments on all the directory build/install ordering goo.
* Beef up that goo so that for each ".WAIT group" all of the install targets
in the previous group have completed before starting the `all', `depend'
or `dependall' targets.

Revert part of previous; we need the install targets.

Use .ORDER targets to express the dependencies on subdir runs. This makes
the resulting tree complexity O(N) (N=# of subdirs), and it also makes it
easier to comprehend the Makefile logic.

Make sure to not go off installing in a directory before it's fully made.

Roll in fixes to permit cross-compiling from non-NetBSD hosts. This
round has been tested on Solaris/x86 and Linux hosts.

* Add host tools cap_mkdb, ctags, m4, uudecode.
* Protect __RCSID() and __COPYRIGHT() better.
* Reduce the number of places that need to include "config.h", to keep
sources closer to their "vanilla" versions.
* Add more compat #defines and autoconf-checked functions.

Make almost all tools compile and run properly on non-NetBSD hosts. (In
particular, most tools now run correctly on Solaris 7.)

build pax as a host tool, for -N and to work around borken vendor paxes

Move host-mkdep to its own directory. Create and install it before
everything else. Now it's possible for the target mkdep to build and
install like "normal" host-tools do; should remove the function decl
warnings reported by greywolf@starwolf.com on current-users.

revert previous, and reenable tools/makefs now that it works on
netbsd-current and netbsd-1-5

Disable makefs in src/tools for the moment; it's very dependent on -current
NetBSD headers and sources right now. (Temporary workaround for PR 15175.)

build zic as a host tool.

build makefs, and, sparc-installboot if sparc||sparc64

We can't currently build a toolchain for x86-64 here, so don't
attempt to do so.

Typo fix in error message.

Add the framework for autoconf-based src/tools/compat structure. This will
(after it is more fleshed out) allow the host tools to work more cleanly
on non-NetBSD hosts.

Add pwd_mkdb as a toolchain bit for builds

Build makewhatis as a host tool, since it is used at the end of
the build.

XXX Needs a -lz hostlib.

Oops. typo which can break things for non-new toolchain builds if MKTOOLS isn't
set

Check for USETOOLS!=yes in addition to MKTOOLS==no to generate the empty rule
sets. Otherwise platforms not running the new toolchain will end up with
lossage as tools attempt to get built and installed into /bin (due to
bsd.own.mk properly checking USETOOLS before setting up a default TOOLDIR).
Don't print the warning unless MKTOOLS=no

Re-add MKTOOLS as a "yes"/"no" only variable, which prevents recursion on "no"
into the src/tools subdirs on all/depend/dependall/install. Other recursive
targets are not affected.

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.

Build the cross-dbsym.

* Always use the makewrapper when starting a build phase.

* Set stuff mostly in the environment, not on make(1)'s command line.
This is needed for MACHINE and MACHINE_ARCH (noted by bjh21), and keeps
the mk.conf-overriding-env behavior (suggested by several people).

* Use MAKEFLAGS to make the wrapper script more readable.

* When prepping objdir in tools for the default TOOLDIR value, just make
that objdir; leave the recursive "make obj" for later.

Don't actually delete the default tooldir on "cleandir"; else we'll nuke
the nbmake that is probably currently running.

Just leave the "cleantools" manual target as a shortcut for the user.

* Create a variable HOST_OSTYPE that represents the host build environment,
using `uname -s`-`uname -r`-`uname -p || uname -m`. This can be overridden
in mk.conf or the environment, and is used by host tools to ensure that
build hosts do not clobber each other.

* Now that there's an objdir in src/tools, make a default TOOLDIR by creating
"tools.${HOST_OSTYPE}" in the objdir (if it exists) of src/tools. This
means that TOOLDIR is no longer required to be set in a build, though it
still may be set manually if desired.

* Set MKTOOLS to "yes" by default, since the default TOOLDIR is now located
in a known mutable location. This may still be set manually if desired.

* 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.

Add asn1_compile as a host tool.

Pull in mdsetimage to the tools builds. (XXX - using libbfd here would
probably be preferrable long term).

Add mklocale to cross-tools and turn it back on for the builds.

Add crunchgen, msgc and menuc

Make mtree a host tool and use it when appropriate.

Recurse properly; just add "build" to TARGETS (fixes reported problem where
"make clean" doesn't work at the src/tools level).

Work around "make obj" lossage in tools/. Dunno why this is happening
just yet; will investigate this evening.

Add lorder to TOOLDIR host tools.

Only descend into directories if USE_NEW_TOOLCHAIN is set (allows top level
Makefile to use "-m ${.CURDIR}/share/mk" to get the "correct" <bsd.own.mk>
when descending into src/tools/Makefile).

- Create a real cpp binary.
- Set CC properly in environment for mkdep and lint.
- Add compile_et to TOOLDIR host tools.

Bootstrap toolchain framework, first pass.

hook gzip to the build so that we can test it in other OS's. Not used yet.

tools/elftoolchain: Handle dependencies properly.

This should properly resolve the problem christos was trying to
address by a bogus make includes hack and #ifdefs in elftoolchain.

tools/Makefile: clean up TOOLDIR validation

At the point where tools/Makefile is loaded, TOOLDIR doesn't have to
exist yet, it will be created later. Remove the redundant quotes from
the .error messages.

build isl tool for GCC 12.

tools/Makefile: Correct inverted comment. NFC

Disconnect 'tools/lib{elf,dwarf}' from the build.

These libraries are now built by 'tools/elftoolchain/lib{elf,dwarf}'
respectively.

When building tools use a dedicated subdirectory for compiling
Elftoolchain code.

Build and install <sys/elfdefinitions.h> in ${TOOLDIR}/include.

This change prepares the ground for upgrading libelf and libdwarf to
their current Elftoolchain versions.

Make dtc on riscv

If we are building date earlier, we need to use an uninstalled copy
of host-mkdep and libnbcompat otherwise the build process looks for
tools that might not be there yet (e.g. fresh build).
Reviewed by martin.

Add a date tool

branches: 1.209.2;
PR port-mvme68k/56146

Build wrtvid as tools.

Make mips-elf2ecoff available for all mips

tools: allow MKCROSSGDB to work when not building gcc

Build mkubootimage for evbsh3 too.

Update LLVM to 10.0.0git (01f3a59fb3e2542fce74c768718f594d0debd0da)

Exclude more subdirs if TOOLS_BUILDRUMP=yes.

Exclude subdirs added in the last 3 years but not need for rumpkernel builds.

Fix build, "pax" must be built before "libprop".

Ok: Matthew Green mrg@

Add support for libprop as a host tool library.

Add MKLLVMRT to build a subset of the LLVM stack for JIT purposes.
Hook up AMDGPU backend for clang.

Revert MKLLVMCMDS changes.

split MKLLVM into MKLLVM and MKLLVMCMDS, like MKGCC and MKGCCCMDS:

- only build clang or clang-tblgen if MKLLVMCMDS != "no"
- move HAVE_MESA_VER earlier so we can detect mesa 18 and enable
MKLLVM on platforms that need it
- move the clang commands and headers into llvmcmds group
- only build the clang specific LLVM libraries if MKLLVMCMDS != "no"

much of this work came from maya@'s github commit
f90685c11d4460d3098fa35f48b58d1893e974e0, reworked for HAVE_MESA_VER
and llvmcmds vs MKCLANG.


build tested on several platforms and on amd64 and i386 with several
different build options (clang build, gcc build, gcc build with llvm
libs).

use MACHINE_CPU for elf2aout

Add an arm-specific elf2aout tool since objdump does not support a.out anymore

Update build glue for xz-5.2.4. Hook up tools version.

hook in the tools version of grep

branches: 1.194.2;
For EXTERNAL_TOOLCHAIN, MKLLVM=yes needs to build only tablegen.

branches: 1.193.2;
Fix .endif comment for TOOLS_BUILDRUMP

Better EXTERNAL_TOOLCHAIN support from Zachary McGrew
- Allow mdsetimage and dbsym to be built with EXTERNAL_TOOLCHAIN
- Allow BFDDIR and IBERTYDIR to be overwritten for mdsetimage and dbsym

branches: 1.191.2;
Hook in dtc/libfdt now that it builds for me

Disable dtc/libfdt for now

No need to set dtc twice.

Add dtc

branches: 1.187.6;
Sync tools build glue with LLVM pre-4.0.0.

branches: 1.186.2;
move cvslatest last

Tool version

move sortinfo after pax

Add sortinfo

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.)

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}

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.

build genassym for TOOLS_BUILDRUMP

add gpt

Changes to existing files to enable building AARCH64 userland.
evbarm64-el
This is clang only. While gcc4.8 supports aarch64, no netbsd support has
been written for aarch64 with gcc4.8.

descend to vgrind

Add cat, rpcgen and stat to TOOLS_BUILDRUMP. They're useful for
installing headers for the app stack cases that need it done, and small
enough to not matter for the kernel-only cases that don't.

remove remaining makefile support for GCC < 45 that i found.

branches: 1.173.2;
ctfconvert/ctfmerge need libdwarf &c., so build them if MKCTF=yes.

Add MKCTF

This variable separates CTF stuffs from MKDTRACE; we can build DTrace
solely without building and using them. This allows us to use DTrace
even if CTF stuffs have problems (actually they have now).

This variable would be merged into MKDTRACE eventually, once CTF stuffs
work correctly again.

Rename NetBSD/hp700 to NetBSD/hppa.

Unfortunately our VCS isn't very helpful here.

Update LLVM/Clang snapshot to r189662. This includes many improvements
to the new SLP vectorizer, fast codegen support for PPC and using
terminfo for the color sequences. In addition, add (optional) build
infrastructure for lld to allow easier testing. This is controlled by
MKLLD.

Even if we aren't making a crossgdb still make the object dir for it.

branches: 1.168.4;
Add infrastructure for building MCLinker.

There are other little endian arm machine_arches other than "arm"
Use MACHINE_CPU and TARGET_ENDIANNESS instead of MACHINE_ARCH

Use TOOLS_BUILDRUMP=yes to restrict building of tools to those required
for building rump kernels. This cuts down the NetBSD source tree required
by buildrump.sh to a fraction and addresses a size/download time concern
from a person packaging rump kernels for Linux.

No change to the normal case (tested with several -j32/-j64 builds).

reviewed by apb

Make tools/Makefile build host-mkdep, compat, and binstall in that order,
before installing any of them. It installs them in any order relative
to each other. The .WAIT points that used to be between these three
tools are replaced by .ORDER and dependencies.

This is needed in preparation for making tools/compat install
some files into TOOLDIR.

TOOLDIR must be valid, unless MKTOOLS=no.
Add tests and .error directives to enforce this.

If tools are not wanted, then set SUBDIR= empty instead of
defining all the subdir-related targets to do nothing. This way,
the targets are not invoked at all, instead of being invoked but
doing nothing.

remove trailing white space

According to the i.MX23 Reference Manual section 3.1, Page 3-3:
"The i.MX23 always operates in litle-endian mode."
So build elftosb and bootloader for this processor only for evbarm(-el).

Build gmake iff we're going to need it.

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.

Add elftosb tools to evbarm userland and toolchain.
Elftosb is used to create a digitaly signed "secure boot" file.
This sb file can be booted by the first stage boot loader found in
Freescale i.MX23 and i.MX28 application processors.

Copyright (c) 2004-2010 Freescale Semiconductor, Inc.

fix EXTERNAL_TOOLCHAIN build

branches: 1.156.2;
Decouple binutils build from HAVE_GCC. Merge common fragments of
HAVE_GCC and HAVE_PCC.

Revert .WAIT removal and add comment why it is needed between *tblgen.

Do not build tic more than once

move tic later, after all the tools it needs are built.

tweak dependencies for tic, which needs nbperf, awk, and sed.

branches: 1.151.2;
Remove an unnecessary .WAIT between llvm-tblgen and clang-tblgen.

Hook up nbperf into the normal tool build.

Update LLVM/Clang snapshot to r141658. This primarily fixes a regression
preventing Qt4 from building.

Build nbm68k-elf2aout as a host tool on m68k targets
for an ancient PROM bootloader.
Mostly taken from nbmips-elf2ecoff.

install elf2ecoff for all mips64 systems. it's needed for stand.

always build gmake if HAVE_GCC.
build gmp, mpfr and then mpc if HAVE_GCC >= 45.

branches: 1.145.2;
Upgrade Heimdal to 1.5pre1 by switching the build from crypto/dist/heimdal
to crypto/external/bsd/heimdal. The latter was just imported as the head
of the Heimdal tree as of a few days ago.

.WAIT for makefs, because in most builds, it's the first target
which includes Makefile.nbincludes, whose beforedepend: has
some race condition I don't fully understand, and which has been
causing many (but not all!) -j16 builds on the autobuild cluster
to fail in tools since sometime between 201103061600 and 201103070900.

With the proper voodoo, I was able to get 8 builds in a row to fail
without this change, and 16 in a row to succeed with it.

XXX a proper fix likely involves reworking the beforedepend: target
of Makefile.nbincludes, but this should get the build cluster back up
to full functionality in the meanwhile.

mkubootimage is certainly useful for sandpoint

Add LLVM and clang build system. It is enabled by MKLLVM=yes.
Upstream sources can be fetched by running "make checkout" in
src/external/bsd/llvm, they will be properly imported once the
integration and missing features are sorted out.

Make mkubootimage for evbmips and evbppc too.

branches: 1.140.2;
add new option to build/use tools/pigz for compressing sets. for now
USE_PIGZGZIP defaults to "no". (it depends upon local <pthread.h> and -lz.)

you can set it to "yes" on most modern platforms just fine.

hook mkubootimage into the build

Bye bye atf-compile, you won't be missed.

DTrace: add libelf, libdwarf, libctf, and CTF tools to the toolchain,
built when MKDTRACE=yes.

Build mandoc unconditional if man pages are desirable. All
compatibility issues are supposed to be sorted out.
Build groff conditionally on MKGROFF. It is currently the only component
that requires C++.

Store our generated files in CVS and remove nbperf from the toolchain.
This allows us to build consistent libs without an ever changing hash.

Userland now builds and uses terminfo instead of termcap.

OK: core@, jdc@

To unbreak the build, temporarily remove the build of libelf here,
since no other part of the build depends on this yet.

Suggested as a temporary workaround by thorpej@

rework the main SUBDIR list here as suggested by enami. put .WAIT on
the same line as the target that needs to wait, but before the target,
rather than on the previous line, which led to confusion and an error
being introduced in the previous change.

move the build of pax before libelf. fixes my build of tools/libelf,
though i didn't look to see why libelf needs pax.

Add support for building libelf in a host-tool environment.

XXX Makefile.disklabel is really mis-named, but I'm not going to tackle
that problem right now.

Give m4 its own .WAIT. Fixes build issue on NetBSD-5.

Add MKMANDOC option to trigger building cat pages and HTML pages with
mdocml. Add a lintmanpages target for developers to check man pages
with the strictest settings of mandoc.

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

Make grep a host tool.

Convert more MACHINE tests to MACHINE_ARCH

branches: 1.124.4; 1.124.8; 1.124.12;
Build sed earlier, just after mktemp. Note that sed and mktemp do not
depend on each other, but the .WAIT is needed to get them both installed
before the remainder of the tools build.

Remove the local handling of the NOSUBDIR variable. We can rely
on bsd.subdir.mk's handling of the NOSUBDIR variable.

Also add some comment to .if/.else/.endif and .for/.endfor lines.

Add awk as a host tool.

Per uwe@'s suggestion, build nbgroff under wider circumstances.
I choose to build it if MKDOC != no or if MKHTML != no. It used
to be built only when MKMAN != no.

Add a .WAIT to ensure that sed is installed early, so that TOOL_SED can
be used during the build of other tools.

Build strfile both as a host tool and as an installed program:
* in games/fortune/strfile/Makefile, build strfile as a
regular program instead of as a host tool;
* add tools/strfile directory to build strfile as a host tool;
* in tools/Makefile, add strfile to SUBDIR list;
* in BSD.*.mk, define TOOL_STRFILE variable;
* in games/fortune/datfiles/Makefile, use TOOL_STRFILE when creating
databases at build time;
* in distrib/sets/lists/games/mi, mention usr/games/strfile.

Build pcc as host tool if requested.

Remove references to bebox-elf2pef and bebox-mkbootimage, now that they
have been integrated in common powerpc tools and subsequently removed.

branches: 1.116.2;
Build powerpc-mkbootimage for rs6000 and bebox too.

Convert prep to use the new unified mkbootimage rather than it's own. This
mkbootimage can be shared by prep, rs6000, and bebox.

TODO: Convert bebox to use this. Needs someone to test.

branches: 1.114.6;
Add the atf host tools

This adds reachover Makefiles to build the atf tools required during a
cross-build.

branches: 1.113.4;
Build join(1) as a host tool.

A commit too far - don't try to build veriexecgen for just now

Build veriexecgen as a host tool

Make sys/exec_elf.h and machine/elf_machdep.h available to host
tools, and allow paxctl to use them when compiled as a host
tool. This fixes the build on MacOS X (and perhaps other platforms)

Add some notes about the bogosity of some of the .WAIT stuff.
Remove a 'never had a sane effect' .WAIT from a .ORDER line.
Don't add a (probably incorrect) dependency for 'make install' since
it contradicts a .ORDER line elsewhere when .ORDER applies recursively
to child nodes.
I also think that the stuff that tries to add dependencies between
the stuff that bsd.subdir.mk generates does not DTRT at all.
I suspect that build.sh builds things in an order that avoids issues here.

Add paxctl as a tool.

If PREVIOUSTOOLDIR was empty, then don't complain that it has moved.

Patch from Jukka Salmi in PR 33095.

* Fix parallel build race condition between real{all,depend} and
<bsd.subdir.mk> targets.
* No need to pass TOOLDIR to recursive cleandir in real{all,depend}.

With help from he@.
OK by tron@.

Another attempt to fix the race in parallel builds.

Invoking "make cleandir" with "PREVIOUSTOOLDIR" set to the previous
location will not stop recursive "make cleandir" calls. Pass the new
location instead.

gettext should .WAIT for autoconf.
(The former still doesn't seem to work on read-only source; separate issue)

Avoid getting into a loop of "make cleandir" instances if the location
tool directory has been changed or wasn't recorded before.

Remove ".WAIT" again because it doesn't fix the problem.

Add an extra ".WAIT" to the "realall realdepend" target to make sure
it finished cleaning the tools directory in case of tool directory
mismatch before we try to rebuild the tools.

only build lint and lint2 if MKLINT != "no".

Make TOOL_SED, to allow using our own sed in cases where the host's version
is busted in some way.

The TOOL_SED doesn't use TIOCGWINSIZE, and has some changes to include
nbtool_config.h in a cross build environment.

Combined with some other changes (not yet committed), this allows cross
compiling m68k code on Solaris 10/amd64.

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

check HAVE_GCC == 4 not USE_GCC4.

Don't barf if USE_GCC4 is not defined.

Turn on gmake builds if GCC4 is enabled (and force it's install before continuing

The realall and realdepend bit that regenerates tools needs to be marked
.MAKE otherwise the 'make' it runs has its own list of jobs instead of
using the jobserver of the enclosing make.

Make the tools using Makefile.disklabel wait so they don't stomp each other.
Also for now add a dependency to Makefile.disklabel so this doesn't run
3 times. (XXX this really should just be done differently)

.include Makefile.disklabel in installboot, so that the
nbinclude/ header files will be installed. Build installboot after
pax, since Makefile.disklabel uses pax to install the headers.

Make disklabel(8) and fdisk(8) into "host tools " last step: build
and install ${TOOLDIR}/bin/${MACHINE_GNU_PLATFORM}-disklabel,
${TOOLDIR}/bin/${MACHINE_GNU_PLATFORM}-fdisk by "reaching over" to
the sources in ${NETBSDSRCDIR}/sbin/{disklabel fdisk}/.

To avoid clashes with a build-host's header files, especially on
*BSD, the host-tools versions of fdisk and disklabel search for
#includes such as disklabel.h, disklabel_acorn.h, disklabel_gpt.h,
and bootinfo.h in a new #includes namespace, nbinclude/. That is,
they #include <nbinclude/sys/disklabel.h>, <nbinclude/machine/disklabel.h>,
<nbinclude/sparc64/disklabel.h>, instead of <sys/disklabel.h> and
such. I have also updated the system headers to #include from
nbinclude/-space when HAVE_NBTOOL_CONFIG_H is #defined.

Build genassym as a host tool. Part of PR toolchain/30350

We don't need aout2bb any longer.

Add some quotes to prevent a syntax error.

Convert mkboot into a full cross tool. Last hurdle for hp300 to building on
non NetBSD/bsd hosts.

Convert remaining host tools for amiga to correct cross tools. Now allows
complete amiga builds on non netbsd/bsd hosts

Convert bebox's remaining host tools over to cross tools scheme. Builds to sets
w/o error now.

Build hp700's mkboot as a host tool called nbhp700-mkboot. mkboot also
gets built and installed in a hp700 distribution.

TODO

- merge with hp300

- pick a preferred method for dealing with the elf headers.
hp700-mkboot and prep-mkbootimage (bintuils) vs mips-elf2ecoff and
tools/installboot

Add sparkcrc (used to generate checksums for acorn32 bootloader sparkive)
as a host tool.

branches: 1.81.2;
Consistently check ${TOOLCHAIN_MISSING} against "no".

Fix typo.

Add macppc-fixcoff tool.

Use (cd ${.CURDIR} && ${MAKE} cleandir) instead of ${MAKE} cleandir
Suggested by Ross Harvey.

always build tools/binutils now.

Build sgivol as a host tool when compiling for sgimips.

add mknod as a host tool

add gdb53 as a host tool for use as a crossgdb
replaces gdb50 from toolchain on archs which use gcc3

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.

Add host tool version of ncdcs

MAKEVERBOSE support

- build "binutils" and "gcc" separately if USE_TOOLS_TOOLCHAIN is not set
to "yes".
- adjust dbsym/mdsetimage to look in the correct place for libbfd.
- new mknative-gcc based in toolchain/mknative

Add missing "cleandir" rule.

add stat(1) as a host tool

also do the MKTOOLS and PREVIOUSTOOLDIR checks for "realdepend"

* If a `PREVIOUSTOOLDIR' file doesn't exist, or its contents are not
the same as "$TOOLDIR", run "make cleandir" here first.
This should prevent various problems where $TOOLDIR is encoded into the
build process for various tools (lint, groff, texinfo, toolchain),
$TOOLDIR changes, and various programs don't get rebuilt correctly.
Should fix PR [toolchain/21988].

* Clean up how the `${MKTOOLS} == no' check is performed.

* Remove "cleantools"; it made a big assumption about the location of
TOOLDIR that won't be correct in many cases.

Citrus iconv support(3)
add host tools : mkcsmapper and mkesdb
mkcsmapper - coded character set mapping table compiler
mkesdb - encoding scheme database compiler

Add db(1) as a host tool

build file(1) as a host tool (for compiling magic files)

Make hexdump a host tool.

enable sunlabel...

Change the TOOLCHAIN_MISSING from == "no" to != "yes", to be consistent
with the other tests of this variable.

Since TOOLCHAIN_MISSING is defined in bsd.own.mk, it can never be undefined.
Instead, make sure the value of it is "no"

* 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.

We need binstall to be installed, before attempting to use it to install mktemp.

Descend into mktemp.

The last commit wasn't tested. Add a missing \

Don't build toolchain bits for hppa, for now.

Add cat(1) as a host tool.

Don't try to build toolchain bits for ns32k.

Add fgen as a host tool, used for sys/arch/sparc/bootblk.

branches: 1.50.2;
gratuitious comment addition

remove now-unnecessary sparc-installboot

On prep, descend into prep-mkbootimage.

vax-installboot has been deleted, so don't recurse into it.

enable installboot

Nuke NO_DBSYM; dbsym now uses BFD for everything.

On second thoughts, don't make mktemp a host tool. It's only used once, for
creating a temporary directory for maketars, and mkdir can do that safely
enough.

Add mktemp as a host tool.

Add cksum as a host tool.

Drop into vax-installboot on VAX machines.

Rework recursion to use the original logic again, in the structure cleaned
up by pk (and fix comments to match):

* A .WAIT here means that only the *immediately preceding* target needs
to be added to the list of "install-foo" targets used to ensure that
dependencies are installed (otherwise much more than is necessary
gets installed during the build process, increasing build time here)

* If building only an individual target such as "dependall-yacc", the
prerequisites may need to be built before running their "install-foo"
targets -- so add a dependency in that case on "dependall-foo".

Don't build groff if MKMAN=no. (groff does not build properly on some
platforms because of its completely insane Make structure; this allows it
to be disabled in a sane way.)

Add gettext to tools.

For now this is a maintainer-tool only; none of the NetBSD build needs to
run msgfmt as of yet. (However, its presence is needed to make the
mknative part of src/tools/toolchain produce the same output as a truly
"native" GNU configury of the toolchain.)

Add elf2ecoff as a build tool for mips ports.

* Add comments on all the directory build/install ordering goo.
* Beef up that goo so that for each ".WAIT group" all of the install targets
in the previous group have completed before starting the `all', `depend'
or `dependall' targets.

Revert part of previous; we need the install targets.

Use .ORDER targets to express the dependencies on subdir runs. This makes
the resulting tree complexity O(N) (N=# of subdirs), and it also makes it
easier to comprehend the Makefile logic.

Make sure to not go off installing in a directory before it's fully made.

Roll in fixes to permit cross-compiling from non-NetBSD hosts. This
round has been tested on Solaris/x86 and Linux hosts.

* Add host tools cap_mkdb, ctags, m4, uudecode.
* Protect __RCSID() and __COPYRIGHT() better.
* Reduce the number of places that need to include "config.h", to keep
sources closer to their "vanilla" versions.
* Add more compat #defines and autoconf-checked functions.

Make almost all tools compile and run properly on non-NetBSD hosts. (In
particular, most tools now run correctly on Solaris 7.)

build pax as a host tool, for -N and to work around borken vendor paxes

Move host-mkdep to its own directory. Create and install it before
everything else. Now it's possible for the target mkdep to build and
install like "normal" host-tools do; should remove the function decl
warnings reported by greywolf@starwolf.com on current-users.

revert previous, and reenable tools/makefs now that it works on
netbsd-current and netbsd-1-5

Disable makefs in src/tools for the moment; it's very dependent on -current
NetBSD headers and sources right now. (Temporary workaround for PR 15175.)

build zic as a host tool.

build makefs, and, sparc-installboot if sparc||sparc64

We can't currently build a toolchain for x86-64 here, so don't
attempt to do so.

Typo fix in error message.

Add the framework for autoconf-based src/tools/compat structure. This will
(after it is more fleshed out) allow the host tools to work more cleanly
on non-NetBSD hosts.

Add pwd_mkdb as a toolchain bit for builds

Build makewhatis as a host tool, since it is used at the end of
the build.

XXX Needs a -lz hostlib.

Oops. typo which can break things for non-new toolchain builds if MKTOOLS isn't
set

Check for USETOOLS!=yes in addition to MKTOOLS==no to generate the empty rule
sets. Otherwise platforms not running the new toolchain will end up with
lossage as tools attempt to get built and installed into /bin (due to
bsd.own.mk properly checking USETOOLS before setting up a default TOOLDIR).
Don't print the warning unless MKTOOLS=no

Re-add MKTOOLS as a "yes"/"no" only variable, which prevents recursion on "no"
into the src/tools subdirs on all/depend/dependall/install. Other recursive
targets are not affected.

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.

Build the cross-dbsym.

* Always use the makewrapper when starting a build phase.

* Set stuff mostly in the environment, not on make(1)'s command line.
This is needed for MACHINE and MACHINE_ARCH (noted by bjh21), and keeps
the mk.conf-overriding-env behavior (suggested by several people).

* Use MAKEFLAGS to make the wrapper script more readable.

* When prepping objdir in tools for the default TOOLDIR value, just make
that objdir; leave the recursive "make obj" for later.

Don't actually delete the default tooldir on "cleandir"; else we'll nuke
the nbmake that is probably currently running.

Just leave the "cleantools" manual target as a shortcut for the user.

* Create a variable HOST_OSTYPE that represents the host build environment,
using `uname -s`-`uname -r`-`uname -p || uname -m`. This can be overridden
in mk.conf or the environment, and is used by host tools to ensure that
build hosts do not clobber each other.

* Now that there's an objdir in src/tools, make a default TOOLDIR by creating
"tools.${HOST_OSTYPE}" in the objdir (if it exists) of src/tools. This
means that TOOLDIR is no longer required to be set in a build, though it
still may be set manually if desired.

* Set MKTOOLS to "yes" by default, since the default TOOLDIR is now located
in a known mutable location. This may still be set manually if desired.

* 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.

Add asn1_compile as a host tool.

Pull in mdsetimage to the tools builds. (XXX - using libbfd here would
probably be preferrable long term).

Add mklocale to cross-tools and turn it back on for the builds.

Add crunchgen, msgc and menuc

Make mtree a host tool and use it when appropriate.

Recurse properly; just add "build" to TARGETS (fixes reported problem where
"make clean" doesn't work at the src/tools level).

Work around "make obj" lossage in tools/. Dunno why this is happening
just yet; will investigate this evening.

Add lorder to TOOLDIR host tools.

Only descend into directories if USE_NEW_TOOLCHAIN is set (allows top level
Makefile to use "-m ${.CURDIR}/share/mk" to get the "correct" <bsd.own.mk>
when descending into src/tools/Makefile).

- Create a real cpp binary.
- Set CC properly in environment for mkdep and lint.
- Add compile_et to TOOLDIR host tools.

Bootstrap toolchain framework, first pass.

tools/Makefile: clean up TOOLDIR validation

At the point where tools/Makefile is loaded, TOOLDIR doesn't have to
exist yet, it will be created later. Remove the redundant quotes from
the .error messages.

build isl tool for GCC 12.

tools/Makefile: Correct inverted comment. NFC

Disconnect 'tools/lib{elf,dwarf}' from the build.

These libraries are now built by 'tools/elftoolchain/lib{elf,dwarf}'
respectively.

When building tools use a dedicated subdirectory for compiling
Elftoolchain code.

Build and install <sys/elfdefinitions.h> in ${TOOLDIR}/include.

This change prepares the ground for upgrading libelf and libdwarf to
their current Elftoolchain versions.

Make dtc on riscv

If we are building date earlier, we need to use an uninstalled copy
of host-mkdep and libnbcompat otherwise the build process looks for
tools that might not be there yet (e.g. fresh build).
Reviewed by martin.

Add a date tool

branches: 1.209.2;
PR port-mvme68k/56146

Build wrtvid as tools.

Make mips-elf2ecoff available for all mips

tools: allow MKCROSSGDB to work when not building gcc

Build mkubootimage for evbsh3 too.

Update LLVM to 10.0.0git (01f3a59fb3e2542fce74c768718f594d0debd0da)