fix tools build

PR/57891: Ricardo Branco: add symbolic flags printing (from FreeBSD)

adapt to ${CC_WNO_IMPLICIT_FALLTHROUGH}

Use ${CC_WNO_IMPLICIT_FALLTHROUGH} instead of
the older style more complex expressions.

Divide stat.1 into stat.1 (now only includes stat(1)) and readlink.1

Apologies to cvs commit purists, but making this division required
line by line reading of the man pages, and I simply could not resist
also correcting some errors, addressing some omissions, improving some
wording ... all at the same time.

convert HAVE_GCC == 7 to HAVE_GCC >= 7.

rework the previous to only apply if !HOSTPROG

apply -Wno-error=implicit-fallthrough on files that have fallthru
comments but are missed by the current checker.

branches: 1.8.42;
document non-literal format strings

More properly host-tool-ify stat

pass in -DHOSTPROG if a HOSTPROG

Fix a tyop - we want the readlink binary and manual pages when HOSTPROG
is not defined.

first attempt at making this a HOSTPROG

NetBSD id police

A new command line option, -q, suppresses *stat(2) failure messages.
Output is now done via stdio, instead of with my stupid homegrown
buffering (I don't even know why I did it that way in the first
place). Also, from Johan Karlsson <johan@freebsd.org>, eliminate a
spurious newline if no output was generated, and link stat(1) to
readlink(1) for similarity to OpenBSD.

Add a stat(1) program to /usr/bin. It provides a simple formatting
interface, with reasonable defaults, to the information provided by
stat(2) so that scripts can use the data therein much more easily.

Make readlink(1) compliant with the POSIX-2004 (Issue 8).

POSIX 2004 requires that if a given file name to readlink(1)
doesn't refer to a symlink, an error be printed.

That conflicts with historic usage on NetBSD where readlink is
silent in that case (except for the exit status) -- rather than
altering our default behaviour, look for POSIXLY_CORRECT in the
environment, and issue an error only in that case (just means
switching between -q and -v as the default noise option).

While here, generate usage if no file args are given - POSIX
requires at least 1 (it actually specifies exactly 1) and while
we can extend that without issues to allow more than one, allowing
0 args (just because we do that for stat(1) and readlink(1) is
the same code) makes no sense at all for readlink - stdin cannot
possibly be a symbolic link.

Also, in the case of stat(1) if we do have no args, and the
fstat(STDIN) fails for some reason, don't claim that it is
lstat (or stat, depending upon -L) which failed:
stat: (stdin): lstat: Bad file descriptor
makes no sense at all. This can occur with something like
stat <&-
which is kind of obscure, but it is easy to DTRT.

No pullups planned.

branches: 1.6.4;
readlink: sort SEE ALSO

Add an Xr to realpath(1) (which does much the same as readlink -f).
Try not to forget the damn comma!

Remove dot at end of SEE ALSO.

readlink(1): Add italic correction.

readlink(1): Add missing .El, use non-random .Bl -width

Divide stat.1 into stat.1 (now only includes stat(1)) and readlink.1

Apologies to cvs commit purists, but making this division required
line by line reading of the man pages, and I simply could not resist
also correcting some errors, addressing some omissions, improving some
wording ... all at the same time.

PR bin/47333

Note that -L will fall back to use lstat() if the stat() requested
returns ENOENT (if the following lstat() fails, the ENOENT from
the lstat() is returned - that generally indicates that the original
ENOENT came from the filename passed to stat() rather than the
value of the symlink. (The man page doesn't say all of that.)

If "stat -L name" returns data from a symlink (eg: if -f %T is @)
then name must refer to a symlink which points to nothing.
(The man page does say that.)

branches: 1.47.2;
PR/57891: Ricardo Branco: add symbolic flags printing (from FreeBSD)

Remove the notge (I added a day or so ago) about IFS interfering with -s
output (potentially) - it is almost certainly wrong.

stat(1): Fix two more -width.

stat(1): Misc fixes.

Use \(em, fix macro usage

Some corrections to stat to make it more reasonable, and comply with
the man page, in particular, in one of the EXAMPLES it is claimed:

This example produces output very similar to that from find ... -ls
(except that find(1) displays the time in a different format, and find(1)
sometimes adds one or more spaces after the comma in “major,minor” for
device nodes):

and then uses %9Z in the format, but conveniently does not use a
device file to demonstrate.

If it had, the result would have been:

8033 0 brw-r----- 1 root operator 4, 51 Jun 8 02:44:45 2022 /dev/sd3

Note the spaces which were claimed to exist (sometimes) in find(1) output
but not with stat. Oops.

Omitting those spaces seems like a good idea, so does using the same
field width for the result in both cases, rather than producing 9 chars
in the ordinary file case, and 19 (2 * 9 + 1) in the device case (for %9Z).

Now we will get:

8033 0 brw-r----- 1 root operator 4,51 Jun 8 02:44:45 2022 /dev/sd3

While here, also make the file number (%@) when reading stdin be 0, it is
not the 1st file in the arg list.

Some (very minor) KNF applied at the same time.

In stat.1 make these changes explicit.

Divide stat.1 into stat.1 (now only includes stat(1)) and readlink.1

Apologies to cvs commit purists, but making this division required
line by line reading of the man pages, and I simply could not resist
also correcting some errors, addressing some omissions, improving some
wording ... all at the same time.

whitespace fixes, use Xr.

- if st_dev/st_rdev == NODEV print -1 instead of unsigned
- update to the latest linux stat(1) format for -x (kind of)
- add the ability to display nanoseconds, like linux does
- change u_intX_t -> uintX_t

Remove workaround for ancient HTML generation code.

Change some standard exit status text to ".Ex -std"
From Eitan Adler in PR bin/47790.

branches: 1.36.6;
Bump date for previous. Remove trailing whitespace.

Honor the -t option to set the time format when running stat -s.
Document the format used by -s on the man page.

branches: 1.34.2; 1.34.8;
Use vis(3) encoding when a string format is modified by the '#' flag.
Also bump copyright years.

More corrections to format compatibility between stat(1), find(1) -ls,
and ls(1) -lT.

Remove unnecessary Ns.

Anothe difference in output format: stat -f "%Z" does not add any
space in "major,minor" for device nodes, whereas both "ls -l" and
"find -ls" add at least one space.

Add examples of formats for stat(1) to simulate "find -ls" and "ls -lTd".

Add verbose mode for readlink.

branches: 1.28.2;
\\ -> \e

Since we have st_birthtime in struct stat, it is in default display.

Drop trailing space.

Be consistent: document the birthtime field of struct stat for
the "B" field specifier.

branches: 1.24.2;
Fix three variable names.
From Todd T. Fries via Jason McIntyre.

Convert TNF licenses to new 2 clause variant

branches: 1.22.6;
Put back the NetBSD tag

Document the flags displayed by the default format, and mention their short
names. From espie@openbsd via jmc@openbsd.

Make sentence easier to parse. From jsing@openbsd via jmc@openbsd.

branches: 1.19.8;
PR/34662: martijnb at atlas dot ipv6 dot stack dot nl: readlink doesn't
grok -f, and there's no alternative (+fix)

Patch applied with minor tweak (%y -> %R, as it was already taken) plus
some nits from myself. Thanks!

Sort SEE ALSO. Drop trailing space.

Bend the 'L', 'M', and 'H' modifiers to work on the size, whereby it
is rounded to the nearest kilobyte, megabyte, or gigabyte.
Implemented at lukem's request since some things can't deal with
overly large numbers when files are really large.

Have to do something like humanize_number(3), but that interface isn't
really what I'm looking for. I think. More examination required.

Some fixes from jmc@openbsd.

correct examples.

Add "dirname"- and "basename"-like functionality to stat(1) by way of
the "high" and "low" selectors on the filename.

Sort options. Use more mdoc macros. Some nit fixes. Bump date.

document default format.

Bump date for last.

Add %B to print the st_birthtime of an inode (adjusting the "raw"
format and "shell" format appropriately), and rename the st_*timespec
fields in the shell-style output to be st_*time, since that's what
they really are.

Fix the sample output in one of the examples

The correct capitalisation of 'NetBSD.org' is (you
guessed it) 'NetBSD.org'.

some mdoc fixes.

New sentence, new line.

Style sync with FreeBSD. Changes via Doug Barton <DougB@FreeBSD.org>.

A new command line option, -q, suppresses *stat(2) failure messages.
Output is now done via stdio, instead of with my stupid homegrown
buffering (I don't even know why I did it that way in the first
place). Also, from Johan Karlsson <johan@freebsd.org>, eliminate a
spurious newline if no output was generated, and link stat(1) to
readlink(1) for similarity to OpenBSD.

Add a simple format that translates to the current file number, fix an
off by one error, and #define a bunch more things to make it clearer.

Formatting, grammar, spelling, and other improvements.

That should be "ls -lTd", not "ls -lT"

Add a stat(1) program to /usr/bin. It provides a simple formatting
interface, with reasonable defaults, to the information provided by
stat(2) so that scripts can use the data therein much more easily.

stat(1): Eliminate a compiler warning

This code at the moment is a bit magical, as it uses pointer arithemtic
on a string and boolean type to return either "lstat" or "stat".

Use the ternary operator instead to make the code easier to read and make
Clang happier.

No functional change intended(tm).

Make readlink(1) compliant with the POSIX-2004 (Issue 8).

POSIX 2004 requires that if a given file name to readlink(1)
doesn't refer to a symlink, an error be printed.

That conflicts with historic usage on NetBSD where readlink is
silent in that case (except for the exit status) -- rather than
altering our default behaviour, look for POSIXLY_CORRECT in the
environment, and issue an error only in that case (just means
switching between -q and -v as the default noise option).

While here, generate usage if no file args are given - POSIX
requires at least 1 (it actually specifies exactly 1) and while
we can extend that without issues to allow more than one, allowing
0 args (just because we do that for stat(1) and readlink(1) is
the same code) makes no sense at all for readlink - stdin cannot
possibly be a symbolic link.

Also, in the case of stat(1) if we do have no args, and the
fstat(STDIN) fails for some reason, don't claim that it is
lstat (or stat, depending upon -L) which failed:
stat: (stdin): lstat: Bad file descriptor
makes no sense at all. This can occur with something like
stat <&-
which is kind of obscure, but it is easy to DTRT.

No pullups planned.

branches: 1.53.2;
stat: don't allow numeric formatting flags on strings

Calling snprintf with "%+s" invokes undefined behavior.

While the change in 1.51 certainly retained binary compat with
what was in 1.50 (while silencing LINT) - it was clearly not the
correct change to make. The code used !FLAG_POUND where it
clearly meant ~FLAG_POUND ... the former is 0, so &= 0 could
be replaced by =0 changing nothing. But that's not what it
should have been doing, other flags should not have been
removed here, just FLAG_POUND.

This problem seems to have existed since support for %#s
was first added in 2011, which kind of suggests how rarely
that format, particularly with other flags (like %#-s)
has ever been used (with no other flags, the bug would not
be noticed).

stat: fix lint warning about constant argument to '!'

No binary change.

fix tools build

PR/57891: Ricardo Branco: add symbolic flags printing (from FreeBSD)

Some corrections to stat to make it more reasonable, and comply with
the man page, in particular, in one of the EXAMPLES it is claimed:

This example produces output very similar to that from find ... -ls
(except that find(1) displays the time in a different format, and find(1)
sometimes adds one or more spaces after the comma in “major,minor” for
device nodes):

and then uses %9Z in the format, but conveniently does not use a
device file to demonstrate.

If it had, the result would have been:

8033 0 brw-r----- 1 root operator 4, 51 Jun 8 02:44:45 2022 /dev/sd3

Note the spaces which were claimed to exist (sometimes) in find(1) output
but not with stat. Oops.

Omitting those spaces seems like a good idea, so does using the same
field width for the result in both cases, rather than producing 9 chars
in the ordinary file case, and 19 (2 * 9 + 1) in the device case (for %9Z).

Now we will get:

8033 0 brw-r----- 1 root operator 4,51 Jun 8 02:44:45 2022 /dev/sd3

While here, also make the file number (%@) when reading stdin be 0, it is
not the 1st file in the arg list.

Some (very minor) KNF applied at the same time.

In stat.1 make these changes explicit.

stat: fix undefined behavior when calling ctype functions

lint says: warning: argument to 'function from <ctype.h>' must be cast
to 'unsigned char', not to 'unsigned int' [342]

avoid a GCC stringop truncate warning.

adjust fallthru comment.

avoid a potential buffer truncation.

branches: 1.43.4;
Handle the new %f (as an addition to the strftime format - -t arg)
correctly, in particular %%f does not contain the %f format, and like
all strftime conversions, %f can occur more than once (not that it is
likely to happen, but just in case...)

KNF, and since this gets compiled as a tool, avoid C99 declarations.

Set the default time format for linux mode (-x) in the correct place,
so it only applies to -x (in particular, not to -s) and only if another
time format has not been explicitly selected (earlier in the arg list).

With luck, this will make lots of tests, which depended upon "stat -s"
actually generating correct sh assignment statements, go back to working again.

- add the sticky bits to mode
- fix linux Birth for the havenots
code review via email from @@@

- if st_dev/st_rdev == NODEV print -1 instead of unsigned
- update to the latest linux stat(1) format for -x (kind of)
- add the ability to display nanoseconds, like linux does
- change u_intX_t -> uintX_t

Unset HAVE_DEVNAME for tools build.
config only checks for the symbol in libc, we need the correct prototype.
Which is excluded even on NetBSD (_NETBSD_SOURCE isn't defined in stdlib.h).

Honor the -t option to set the time format when running stat -s.
Document the format used by -s on the man page.

branches: 1.36.2; 1.36.8;
Use vis(3) encoding when a string format is modified by the '#' flag.
Also bump copyright years.

static + __dead

Add verbose mode for readlink.

branches: 1.33.2;
Do not crash if a date cannot be represented (localtime returning
NULL), use the Epoch value instead.

Don't pass unused argument. From Doug Barton.

Use strlcpy, not strncpy, when the desired semantics are strlcpy's
rather than strncpy's.

Improve previous with comments.

Don't printf time_t with %d; fixes PR 44128 from yamt. With this change it
successfully prints mtimes after 2038.

Fix WARNS=4 issues (-Wcast-qual -Wsign-compare)

branches: 1.27.6;
The ofmt variable is actually a bit mask (not the character that was
in the format string) so that we can "or" it with the bits in the
formats variable. This fixes the missing " -> " in front of the real
path (when you use %SR).

Also, the ?: needs another space.

branches: 1.26.2;
Remove clause 3 and 4 from TNF licenses

branches: 1.25.4;
Fix a segfault when doing 'stat -f %R' on the stdin file handle, instead
fake the filename '(stdin)' like the %N format.

branches: 1.24.2; 1.24.4; 1.24.8;
PR/34662: martijnb at atlas dot ipv6 dot stack dot nl: readlink doesn't
grok -f, and there's no alternative (+fix)

Patch applied with minor tweak (%y -> %R, as it was already taken) plus
some nits from myself. Thanks!

Bend the 'L', 'M', and 'H' modifiers to work on the size, whereby it
is rounded to the nearest kilobyte, megabyte, or gigabyte.
Implemented at lukem's request since some things can't deal with
overly large numbers when files are really large.

Have to do something like humanize_number(3), but that interface isn't
really what I'm looking for. I think. More examination required.

Fix a trivial truncation case, and eliminate a corner case that might
print a nul character.

Now that tools/compat provides a check, make sure st_birthtimensec exists
before using it

Add "dirname"- and "basename"-like functionality to stat(1) by way of
the "high" and "low" selectors on the filename.

Completely rework how tools/compat is done. Purge all uses/references to
_NETBSD_SOURCE as this makes cross building from older/newer versions of
NetBSD harder, not easier (and also makes the resulting tools 'different')

Wrap all required code with the inclusion of nbtool_config.h, attempt to
only use POSIX code in all places (or when reasonable test w. configure and
provide definitions: ala u_int, etc).

Reviewed by lukem. Tested on FreeBSD 4.9, Redhat Linux ES3, NetBSD 1.6.2 x86
NetBSD current (x86 and amd64) and Solaris 9.

Fixes PR's: PR#17762 PR#25944

If using stat (the -L flag) and it fails, fall back to lstat(). It
may be the case that we're examining a broken symlink, and anything is
better than nothing.

branches: 1.17.2;
Take the time to set gottime so that you don't have to waste time
trying to figure out why the time isn't the time you expected all the
time.

be consistent using #if HAVE_STRUCT_STAT_ST_xxx (versus #ifdef)

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 for systems which don't have st_*timespec. PR 23214.

More properly host-tool-ify stat

* don't compile in support for st_birth #ifdef HOSTPROG
* remove %B (birth) from default format strings, since %B isn't available
#ifdef HOSTPROG, and it's not standard enough to be fussed about anyway

first attempt at making this a HOSTPROG

Add %B to print the st_birthtime of an inode (adjusting the "raw"
format and "shell" format appropriately), and rename the st_*timespec
fields in the shell-style output to be st_*time, since that's what
they really are.

use readlink with bufsize - 1; approved thorpej.

Use limits.h instead of sys/syslimits.h since it's more potrable.

Style sync with FreeBSD. Changes via Doug Barton <DougB@FreeBSD.org>.

Header reorganization makes a shorter list. From Doug Barton <DougB@FreeBSD.org>

The bitwise or-ing together of letters is almost *guaranteed* to do
the wrong thing, as in this case. I should read the preprocessor
output more often.

A new command line option, -q, suppresses *stat(2) failure messages.
Output is now done via stdio, instead of with my stupid homegrown
buffering (I don't even know why I did it that way in the first
place). Also, from Johan Karlsson <johan@freebsd.org>, eliminate a
spurious newline if no output was generated, and link stat(1) to
readlink(1) for similarity to OpenBSD.

When asked for the string representation of a device that isn't
found by devname(3), don't report a "bad format"...just use "???"
instead.

branches: 1.2.2;
Add a simple format that translates to the current file number, fix an
off by one error, and #define a bunch more things to make it clearer.

Add a stat(1) program to /usr/bin. It provides a simple formatting
interface, with reasonable defaults, to the information provided by
stat(2) so that scripts can use the data therein much more easily.