bsd.own.mk: rename GCC_NO_* to CC_WNO_*

Rename compiler-warning-disable variables from
GCC_NO_warning
to
CC_WNO_warning
where warning is the full warning name as used by the compiler.

GCC_NO_IMPLICIT_FALLTHRU is CC_WNO_IMPLICIT_FALLTHROUGH

Using the convention CC_compilerflag, where compilerflag
is based on the full compiler flag name.

introduce some common variables for use in GCC warning disables:

GCC_NO_FORMAT_TRUNCATION -Wno-format-truncation (GCC 7/8)
GCC_NO_STRINGOP_TRUNCATION -Wno-stringop-truncation (GCC 8)
GCC_NO_STRINGOP_OVERFLOW -Wno-stringop-overflow (GCC 8)
GCC_NO_CAST_FUNCTION_TYPE -Wno-cast-function-type (GCC 8)

use these to turn off warnings for most GCC-8 complaints. many
of these are false positives, most of the real bugs are already
commited, or are yet to come.


we plan to introduce versions of (some?) of these that use the
"-Wno-error=" form, which still displays the warnings but does
not make it an error, and all of the above will be re-considered
as either being "fix me" (warning still displayed) or "warning
is wrong."

branches: 1.15.82;
We only need libutil in native mode.

More programs using efun.

Consistently check ${TOOLCHAIN_MISSING} against "no".

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.

support MAKEVERBOSE and use ${HOST_SH}

Now that <bsd.prog.mk> DTRT if HOSTPROG is defined (i.e, it is a no-op),
there's no need to special-case .include-ing it.

* 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.

Move check of USE_NEW_TOOLCHAIN so it'll actually be *defined* for a native
crunchgen binary where appropriate.

Use .PATH to find mkskel.sh; formatting cleanup.

When adding something to CPPFLAGS, use +=, not =

If USE_NEW_TOOLCHAIN is defined use objcopy instead of crunchide as it works now.

use CPPFLAGS instead of CFLAGS

branches: 1.3.2;
1) RCSid police
2) Add __RCSIDs where apropriate.
3) WARNS=1, and clean up sources for WARNS=1 (including replacement of
a mktemp with a mkstemp even though it was probably safe...)
4) Some other small cosmetic changes

local changes; install into /usr/bin, deal with sh warnings, deal with obj.MACHINE, etc.

branches: 1.1.1;
Initial revision

Make this produce proper ansi c and knf.

need <stdlib.h> for exit(3).

Add envp argument support to crunchgen.

Even though envp is not the POSIX way of crawling through the environment,
this will hopefully save someone from having to scratch their head
really hard with a crunched program using envp.

from henric@aimnet.com, via OpenBSD

1) RCSid police
2) Add __RCSIDs where apropriate.
3) WARNS=1, and clean up sources for WARNS=1 (including replacement of
a mktemp with a mkstemp even though it was probably safe...)
4) Some other small cosmetic changes

branches: 1.1.1;
Initial revision

fix syntax error and typo in overwritten.

New sentence, new line. Fix typo.

Provide a migration guide from the old flags to -V.

1. Remove all the special handling of variables (-d -p -P -s -S) that
were dealing with DBG (-d) LDSTATIC/NOPIE (-p), and the rest with
disabling/enabling sanitizers.
2. Use emalloc/estrdup for all the allocators instead of only some cases.
3. Add -V varspec which passes variables on the command line (as DBG
and LDSTATIC used to be passed before) instead of appending them
to the on-the-fly Makefile using -v varspec.
4. Change the distrib and rescue Makefiles to use -V instead of the removed
flags.

The motivation of this is to make variable handling consistent, less magical,
and remove the need for changing crunchgen each time we want to add disabling
an option by default.

(as proposed in tech-toolchain)

Also disable ssp and fortify by default.

Add -S to SYNOPSIS. Sort option descriptions.

Introduce a new option -S in crunchgen(1)

The -S argument enables sanitization with a sanitizer in libc.

branches: 1.35.2;
Add new option -s to crunchgen(1) -- enable sanitization

As of today typical sanitizers require dynamic executables, while
crunchgen(1) programs are produced with static properties.

Lack of specified -s will:
- generate a Makefile file with NOSANITIZER=
- build programs that are dependencies with NOSANITIZER=

In future there is an option to handle sanitization in statically linked
programs.

An idea with -s LGTM by <christos>

branches: 1.34.2;
New sentence, new line. Remove superfluous Pps.

Handle static PIE

Merge riastradh-drm2 to HEAD.

Use Mt for email addresses.

branches: 1.30.2;
Document section stripping.

branches: 1.29.2;
Remove unnecessary Bk/Ek pairs from SYNOPSIS.
No effective change except where I used the opportunity to sort options
and/or option descriptions.

branches: 1.28.6;
Use paragraph breaks before notes, not just line breaks.

branches: 1.27.26;
Bump date for previous.

Treat multiple

SPECIAL prog {objs,objpaths,keepsymbols} itemX

lines (where "prog" is fixed) in crunchgen configuration the same as

SPECIAL prog {objs,objpaths,keepsymbols} item1 ... itemN

OK by David Laight

Sort options in SYNOPSIS and in description; use more mdoc macros.

Change the way crunchgen works when doing 'reachover' builds.
Instead of trying to identify all the .o files that make each target rely
on the targets own Makefile having a program.ro target that will compile
all the objects and link them into a single relocatable.
Rename each program's 'main' to '_crunched_<prog>_stub' instead of compiling
a small piece of C.
The old behaviour can be forced by specifying -O, and is also done if the
config file specifies 'objs' or 'objpaths'.
As well as simplifying the logic, this means you only get a single 'prog.ro
is up to date' message for each program during the build.

The conffile argument is not optional; also, move it behind all
the options. Bump date.

Add -v VARSPEC, so that we can pass variable assignments to make(1)

branches: 1.21.4;
progams->programs (Igor Sobrado, PR misc/19700) + homegrown->home-grown

.Nm does not need a dummy argument ("") before punctuation or
for correct formatting of the SYNOPSIS any longer.

Two typos. From Adrian Mrva. eg -> e.g.

add -o to synopsis

Note -d flag in SYNOPSIS.

Generate <>& symbolically. I'm avoiding .../dist/... directories for now.

Add special command "keepsymbols", allows the specified external symbols
not to be hidden. From Jens Nilsson, jens@rockstorm.se.

Use .Bd -literal.

Whitespace, punctuation, section header fixes, sort sections.

Add a -d option to set the DBG flag to whatever flags the builder wants. If
it's not set, default to -Os (so it's not hard coded into the binary.)

Generate a makefile by default that does all object file builds in local
subdirs by binary (i.e. bad144 gets built in $OBJDIR/bad144) via a reach over
make. (Sets .PATH and .CURDIR and does a make for the objects needed in that
dir).

Add a -o option that does the old behavior of searching for obj files by
querying the various object environment vars and searching. Without this
option none of the objdir special options, environment vars, etc will have
effect.

Document all the changes in the man page (and specificly note that without -o
none of the objdir functions take effect).

While I'm here.. document the rest of the environment variables I found in
this program.

Document the MAKE environment variable.

More and more .Os cleanups. .Os is defined in the tmac.doc-common file,
so we shouldn't override it with versions in the manpages. Many more to
come.

Clean up SYNOPSIS formatting.

add another caveat entry.

macroize BSD, NetBSD, FreeBSD and misc cleanup

Fix type: srt-root -> src-root

1) RCSid police
2) Add __RCSIDs where apropriate.
3) WARNS=1, and clean up sources for WARNS=1 (including replacement of
a mktemp with a mkstemp even though it was probably safe...)
4) Some other small cosmetic changes

clean up formatting glitches (bogus newlines in source)

add two flags:
-D to specify a "root" from which relative source dir paths
are assumed to begin, and
-L to specify the library path for included libs.

branches: 1.1.1;
Initial revision

crunchgen(1): Implement -L <lib-dir> option for real.

Pass it through via LDADD in the generated .mk output, before any -l
options.

The original logic -- which was disabled in 2001, apparently by
accident -- had a default libdir of /usr/lib. But it didn't do
anything with that string, so removing it as a default doesn't change
anything. But the linker doesn't need us to specify that path anyway
(and it would really have to be -L=/usr/lib rather than -L/usr/lib,
whereas we want -L/path/to/obj/lib/libfoo without the `=').

PR toolchain/59841: crunchgen(1) ignores -L argument

Will be needed for:

PR lib/58648: private shared libraries should go in /usr/lib/private,
not /usr/lib

fix various typos in comments.

crunchgen(1): Clear PaX flags instead of removing its ELF note section.

The latter results in zero-filled hole in ELF note segment for EARM,
where PaX section is not located the bottom of that segment (see
src/lib/csu/sysident.S). Fortunately, this hole does not cause real
harms for our in-kernel ELF note parser, except for noisy warnings on
DIAGNOSTIC kernels.

Bump CRUNCH_VERSION.

PR toolchain/52675

branches: 1.94.8;
1. Remove all the special handling of variables (-d -p -P -s -S) that
were dealing with DBG (-d) LDSTATIC/NOPIE (-p), and the rest with
disabling/enabling sanitizers.
2. Use emalloc/estrdup for all the allocators instead of only some cases.
3. Add -V varspec which passes variables on the command line (as DBG
and LDSTATIC used to be passed before) instead of appending them
to the on-the-fly Makefile using -v varspec.
4. Change the distrib and rescue Makefiles to use -V instead of the removed
flags.

The motivation of this is to make variable handling consistent, less magical,
and remove the need for changing crunchgen each time we want to add disabling
an option by default.

(as proposed in tech-toolchain)

Also disable ssp and fortify by default.

remove the hack to remove .eh_frame -- gcc7 is fixed it seems.

while we're still figuring out the gcc7 vs .eh_frame issue, apply
the don't remove eh_frame hack to mips as well. hpcmips testbed
is also failing currently:

[ 3.1238738] panic: init died (signal 6, exit 12)

hack alert time:

on sparc and sparc64, don't remove .eh_frame section. it leads
to failure as something is referenced, and objcopy ends up
emitting a broken binary that can't be run -- it attempts to
load at va=0, beyond having missing referenced data.

also, on sparc64 also don't remove .note.netbsd.mcmodel.


the former should be revised when we can avoid it.

Sync usage with man page.

Introduce a new option -S in crunchgen(1)

The -S argument enables sanitization with a sanitizer in libc.

branches: 1.87.2;
Add new option -s to crunchgen(1) -- enable sanitization

As of today typical sanitizers require dynamic executables, while
crunchgen(1) programs are produced with static properties.

Lack of specified -s will:
- generate a Makefile file with NOSANITIZER=
- build programs that are dependencies with NOSANITIZER=

In future there is an option to handle sanitization in statically linked
programs.

An idea with -s LGTM by <christos>

in dir_search(), don't assume a directory existing is useful, instead
confirm that there is a non zero makefile in there. (this assumes
the makefile is called "Makefile", which is assumed in other places
in crunchgen.c already, so this doesn't make it worse.)

this fixes build issues when an empty subdir exists because some files
were moved subdir at some stage (ktrace, rcorder), and a non-prune
update may look in the wrong dir.

bump version (lots of updates between now and the previous update.)

branches: 1.85.2;
Handle static PIE

Don't make crunch binaries PIE

Use :Q instead of " to quote make vars DBG and LDSTATIC

Remove .note.netbsd.mcmodel and .note.netbsd.pax from the chrunched binaries.
We don't know wether the former would make sense (as currently used, it never
makes sense for static binaries) and we can not conclude safe PAX flags from
combined binaries.

Strip .eh_frame and .eh_frame_hdr from crunched binaries.

branches: 1.80.2; 1.80.8;
static + __dead

Make this produce proper ansi c and knf.

pass LDSTATIC to recursive makes.

move LDSTATIC before the includes because the rules need it early for
conditionals.

Fix a bug.
always mkdir for 'objs' and 'objdir' when NEW_TOOLCHAIN.

Fix WARNS=4 issues (-Wshadow -Wcast-qual)

branches: 1.74.4;
In shell scripts invoked during a build, and in crunchgen, use ${AWK}
instead of plain "awk". The Makefiles that invoke these scripts
or programs will pass AWK=${HOST_AWK:Q}.

Treat multiple

SPECIAL prog {objs,objpaths,keepsymbols} itemX

lines (where "prog" is fixed) in crunchgen configuration the same as

SPECIAL prog {objs,objpaths,keepsymbols} item1 ... itemN

OK by David Laight

More programs using efun.

detect overflow and exit with an error.

PR/33698: Kevin Massey: use strlcpy/strlcat to avoid stack overwrites.

Fix problems with program names with a "-" in them, where crunchgen was
outputting the original program name for the stub name instead of the
mangled name.

From Mikolaj Golub on tech-embed.

branches: 1.68.2;
Don't use data in freed stack.

branches: 1.67.2;
Make dependency checking for the pull-over make's actually work.
The make depend needs to be it's own separate make so it's done before the
main make runs. This also fixes issues w. -j breaking on 'depend <target'.

In addition, force the sub cross-over make to always run or otherwise it'll
never check dependencies once a valid foo.cro has been created.

Bump to 20050208

Revert part of previous [remove cd to objdir] and bump the crunchdate.
Make has been fixed. The cd was in the wrong place anyway, it should
immediately follow ${MAKE}, so it did not work as expected.

Explicitly cd to the .OBJDIR in the ${PROG}.strip target.
This fixes the "make -j N dependall" build issues people were
seeing in src/rescue, that appears to have been introduced
in rev 1.62.

Use the MAKEVERBOSE macros (if present) to print the pretty message.

Set the CRUNCH_VERSION to 20050130.

Add (unsigned char) cast to ctype functions

Sync usage with man page (sort, add -O).

Change the way crunchgen works when doing 'reachover' builds.
Instead of trying to identify all the .o files that make each target rely
on the targets own Makefile having a program.ro target that will compile
all the objects and link them into a single relocatable.
Rename each program's 'main' to '_crunched_<prog>_stub' instead of compiling
a small piece of C.
The old behaviour can be forced by specifying -O, and is also done if the
config file specifies 'objs' or 'objpaths'.
As well as simplifying the logic, this means you only get a single 'prog.ro
is up to date' message for each program during the build.

Use a single awk command (instead of awk and a lot of greps) to generate
the list of symbols to be renamed.
Rename symbols to <symbol>$$from%%<prog> (not _$$hide$$<prog>.cro$$<symbol>)
so that gdb reports the symbol name instead of "_" (the $$ is treated as a
C++ name mangle and nothing after it is output by default).

Whitespace police:
if, for and while are not functions,
put controlled statements of if and while onto separate lines
put function names into column 1

Remove the always defined RENAME_SYMS and the code that would be
compiled were it defined.
That code is out of date, doesn't work on some archs, and is in the way
of another commit.

Sync usage with man page.

Add -v VARSPEC, so that we can pass variable assignments to make(1)

remove copyright and comment sections. Saves ~40K on the crunched
ramdisk binary.

branches: 1.55.4;
Fix PR#4980 and support common symbols correctly in crunched programs.
Exclude them from renaming and expect the linker to do the right thing in
resolving them all to the library copy (i.e. optarg) or locally as the
case may be. Testing multiple programs crunched which reference optarg shows
correct behavior now.

Support crunch configs which specify relative srcdir's but don't use -D. For
these cases, prepend getwd() onto the front of them so a full correct path
is passed to the reach over make command (as it's running from at least
a subdir of the current directory for each object)

Use grep in favor of egrep in a few places. The pattern being searched
for is just a string and some egrep's (notably the Solaris one) don't like
the -w option.

Change the make(1) command lines generated by crunchgen from:

"make -a -b VAR=VALUE -c -d target" to
"make -a -b -c -d VAR=VALUE target"

to avoid depending on make's undocumented and unportably-implemented
handling of the former case.

Makes crunchgen and nbmake work together on Cygwin.

Overhaul how `build.sh tools' are used:

* Rename "config.h" to "nbtool_config.h" and
HAVE_CONFIG_H to HAVE_NBTOOL_CONFIG_H.
This makes in more obvious in the source when we're using
tools/compat/config.h versus "standard autoconf" config.h

* Consistently move the inclusion of nbtool_config.h to before
<sys/cdefs.h> so that the former can provide __RCSID() (et al),
and there's no need to protect those macros any more.

These changes should make it easier to "tool-ify" a program by adding:
#if HAVE_NBTOOL_CONFIG_H
#include "nbtool_config.h"
#endif
to the top of the source files (for the general case).

fix typo that prevented ${PROG}.unstripped from being removed in "clean"

Keep unstripped binary in $OBJ, use are timestamp instead of prog.stripped.
Patentially useful as a source of symbol information for debugging.

"objcopy --redefine-syms-file" is now "objcopy --redefine-syms".

Add an extra target to the xxx.mk file that is equivalent to 'all', but which
doesn't pick up baggage from bsd.own.mk and bsd.prog.mk.
This stops parallel makes exploding because they try to build ${PROG} to soon.
Fixes toolchain/21563 (with change to src/distrib/common/Makefile.crunch).
Upped version of crunchgen to 0.3

Add a .WAIT to ensure that all the program objects are built before
trying to build the crunched binary.
Might fix toolchain/21563

Make the prog.mk makefile run make in each of the target directories and
then build the program itself.
This means that changes to any of the program source files cause the
crunched binary to be rebuilt.
Seems to work ok with the sysinst ramdisks.

Back out rev 1.43 for now, it seems to create some new problems (see
the sysinst Makefile).

No need to invoke the make command so many times. It can handle all
the target specified in command line.

The name of the working directory for each prog is converted by "s/-/_/".
The directory list to be removed (i.e. make clean) has to be set with
the converted name.

Use new the --redefine-syms-file option to objcopy to rename the symbols
we want to hide instead of making global symbols local (which doesn't
work on MIPS).

- move LDSTATIC=-static to the bottom, so that "LDSTATIC=" in /etc/mk.conf
doesn't cause lossage
- set NOMAN= at the top (rather than MKMAN=no)

Add -o, -D, and -L to usage() to get it in sync with reality.

revert revision 1.37; it causes problems when running make(1) on a
crunchgen-created makefile with a partiallly complete build
(a situation which is more common than you think...)

solving the "parallel build" problem for crunchgen-created makefiles
needs to be done a different way...

fixes [toolchain/17012] from me.

branches: 1.37.2;
Some tweaks to (hopefully) make the crunchgen-generated makefile
parallel-build-safe.

Invoke the "make crunchgen_objs" submake with -B so we aren't confused
if we're being invoked in the context of a parallel build.

Move tempfname[] to fillin_program_objs() (the only place that calls it),
and assign it before each use of mkstemp(3).
Problem noted by Ben Harris <bjh21@netbsd.org> in [toolchain/16140].

Use "r" instead of "r+" to popen(3), since the latter is not portable
and the I/O stream is only read from in any case. Problem noted and fix
suggested by Ben Harris <bjh21@netbsd.org> in [toolchain/16139].

CRUNCHEDPROG=1 is a bit more useful if it's also defined when
determining prog_OBJPATHS ...

pass CRUNCHEDPROG=1 to make(1) when building progs from their Makefiles.
allows different options for programs that are being crunchgen(3)ed.

Include config.h for host tool; protect __RCSID.

Add special command "keepsymbols", allows the specified external symbols
not to be hidden. From Jens Nilsson, jens@rockstorm.se.

Oops...don't nuke the global symbols with objcopy. Just keep the crunched stub
and revert the others to locals

remove superflouos arguments to fprintf()

make sure that we ramdiskbin is static. By the numbscull! Nobody tested this?
BTW. it seems that only make working floppies with USETOOLS=no. Will
investigate more.

Check for MAKEFLAGS in the env and pass those to the test make's run for
finding obj files. Otherwise -m <new mk files> won't get passed and can
cause problems on some cross builds.

If USE_NEW_TOOLCHAIN is defined use objcopy instead of crunchide as it works now.

Convert makefile to use bsd.prog.mk. Makes linking work since the .mk file
takes care of it and this will handle a DESTDIR build setup then.

Add a -d option to set the DBG flag to whatever flags the builder wants. If
it's not set, default to -Os (so it's not hard coded into the binary.)

Generate a makefile by default that does all object file builds in local
subdirs by binary (i.e. bad144 gets built in $OBJDIR/bad144) via a reach over
make. (Sets .PATH and .CURDIR and does a make for the objects needed in that
dir).

Add a -o option that does the old behavior of searching for obj files by
querying the various object environment vars and searching. Without this
option none of the objdir special options, environment vars, etc will have
effect.

Document all the changes in the man page (and specificly note that without -o
none of the objdir functions take effect).

Fix some lossage in previous commit: don't remove everything when cleaning
in a source directory, just the object files. This is a _very temporary_
fix; I will untangle the meaning of UPDATE, etc. in this context presently.

Changes to the emitted makefiles that actually build the objects:

1) Always do a make clean before building objects in any directory. This
is wasteful, but there's really no other simple way to cope with the
fact that the compilation settings (e.g. CFLAGS) appropriate for the
non-crunched build of a program may not be appropriate for the crunched
build. If the objdir magic in make didn't rely upon the presence of an
"obj" or "obj.${MACHINE}" symlink, we could abuse it to handle this but
unfortunately, it does.

2) Override $DBG to cause object files to be built with -Os. We can't emit
"DBG?=" into the generated makefile because of order-of-inclusion issues
with the system Makefiles; the result would be that the default setting
(currently -O2) would always be used instead of -Os. If you're crunching,
you almost certainly are doing it to get a smaller executable (!) so -Os
is almost certainly appropriate for you.

fix nested extern

change the direct call to crunchide to ${CRUNCHIDE}

Modify this to obey $MAKE in the environment, instead of allways calling
"make" hardcoded. This should cause my cross-ramdisk-builds to stop
blowing a gasket.

Teach crunchgen.c about MAKEOBJDIRPREFIX.

branches: 1.16.2;
Fix uninitialized structure element.

branches: 1.15.2;
Don't declare 'extern opt*' getopt variables.

Don't try to use /tmp/Makefile...

Make rules for finding object directory match those used by make:
obj.${MACHINE} takes precedence over obj.
the generated makefile should probably just use the print-objdir hack instead.

Use uname(3) to determine MACHINE type.

put temporary file in /tmp, and cd into src dirs and make w/ "Makefile"
(rather than making w/ "srcdir/Makefile") so that relative path specs in
srcdirs work out correctly.

Generate a makefile which is cross-crunching compatabile:

1) Include bsd.sys.mk so we pick up DESTDIR changes to fine .h files.
2) Only set STRIP if it's unset.
3) Generate .cro files rather than .lo files. .lo now is used for "local
objects" - obj's for the host machine not the target machine.

branches: 1.9.2;
Teach crunchgen to play nice in a cross-compiling environment. Make
the machine type a run-time decision a la make. Also, use ${LD}
for the linker, not ld.

Note: you will have to delete the crunchgen cache file for these
changes to take effect.

__AUDIT__ cleanup.

1) RCSid police
2) Add __RCSIDs where apropriate.
3) WARNS=1, and clean up sources for WARNS=1 (including replacement of
a mktemp with a mkstemp even though it was probably safe...)
4) Some other small cosmetic changes

give crunchide a C label name with -k, since crunchgen really doesn't
want to have to care about the object format in use, and C label prefix
is object format dependent.

Apply `topdir' prefix to relative paths in `special srcdir <prog> <path>'
lines.

Allow an override of STRIP=something_else in the generated makefile.

add two flags:
-D to specify a "root" from which relative source dir paths
are assumed to begin, and
-L to specify the library path for included libs.

local changes; install into /usr/bin, deal with sh warnings, deal with obj.MACHINE, etc.

branches: 1.1.1;
Initial revision

Fix WARNS=4 issues (-Wshadow -Wcast-qual)

branches: 1.2.56;
1) RCSid police
2) Add __RCSIDs where apropriate.
3) WARNS=1, and clean up sources for WARNS=1 (including replacement of
a mktemp with a mkstemp even though it was probably safe...)
4) Some other small cosmetic changes

branches: 1.1.1;
Initial revision