It is infeasible for some ARM ports to use anything but ELF/new-toolchain,
and this allows for the ARM ports that aren't ready for ELF to stay at
MACHINE_ARCH == arm32 until they're ready. The rest of the ARM ports
have a userbase such that if a flag day happens with the compiler (there
are a couple of ABI issues still being discussed), that it won't matter
so much.
Discussed with Ben Harris, Matt Thomas, and Chris Gilbert.
* 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 $(srcdir), and mark them with .MADE: to prevent rebuilding.
* Only build cross tools in src/tools; remove some messy logic in
src/usr.* and src/gnu/usr.* that would do target filename rewriting
(improves consistency and readability).
* Add the ability to build cross gdb at tool build time by setting
MKCROSSGDB (default no) to "yes" in mk.conf.
* Add src/tools/groff and set up paths to work with this cross groff.
is set (and MAKEOBJDIR[PREFIX] is not in use). This permits multiple hosts
to compile hostprogs (not just src/tools) for a given target without
colliding with each other.
Currently restricted only to new toolchain builds, as this requires a new
make(1) in order to function properly.
source directory. This will use BSDOBJDIR of OBJMACHINE is unset, and will
create a directory '.MACHINE' inside BSDOBJDIR if OBJMACHINE is set.
MAKEOBJDIR and MAKEOBJDIRPREFIX based objdirs already cope with the top
level directory just fine.
creation, to exclude conflicting make targets from botching pkgsrc), to
avoid defining INSTALL_FILE. pkgsrc is self-contained except for
<bsd.own.mk>, so this doesn't have an adverse effect on real <bsd.*.mk>
build trees.
descriptive comments. The guts of hostprog, lib, and prog are mostly
unchanged; this will be done in another pass.
Make several things work properly with UNPRIVILEGED set, and make the
install-time "cmp" logic work for MANZ. Also reimplement INCS{DIR,NAME}_foo
(requested by wiz) in a backwards compatible way.
Reviewed by christos.
generates too many false positives that require gross amounts of
workaround, and the other WARNS=2 stuff is quite useful. Per discussion
with simonb & christos.
using `uname -s`-`uname -r`-`uname -p || uname -m`. This can be overridden
in mk.conf or the environment, and is used by host tools to ensure that
build hosts do not clobber each other.
* Now that there's an objdir in src/tools, make a default TOOLDIR by creating
"tools.${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.
running from inside of $BSDSRCDIR (recurse back and look for file/dir
signatures or reach /, whichever comes first).
Do it once and then pass the results to sub-makes.
Finally, if using the new toolchain and USETOOLS=yes check $TOOLDIR
before using it. Print an error message out if it's not set when expected
here.
generic architecture case; this permits the later ELF clause to set
them properly for ELF platforms.
This will cause crtbeginS.o and crtendS.o to be included in shared
libraries built on all ELF platforms, not just alpha, mips, vax, and
sparc.
(and dependall) targets to work properly.
XXX THIS IS BROKEN. It's highly questionalble that we should be applying
our .depend-generating rules to a host system -- the mkdep we currently
use assumes GCC.
