PR/54768: elo: Missing cksum(1) convenience links (and docs) for new sha2
functions.

branches: 1.16.46;
Enable WARNS=4 by default for usr.bin, except for:
awk bdes checknr compile_et error gss hxtool kgetcred kinit
klist ldd less lex locale login m4 man menuc mk_cmds
mklocale msgc openssl rpcgen rpcinfo sdiff spell ssh
string2key telnet tn3270 verify_krb5_conf xlint

branches: 1.15.4;
WARNS=4

Add support for SHA2 in cksum.

branches: 1.13.2;
WARNS=3

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.

Add the necessary bits to make cksum buildable as a host tool.

Add MD2, MD4, SHA1, and RMD160 support to cksum. Adding new hashes to
cksum(1) should now be easy.

don' define WARNS=1 here

WARNSify

Add support for generating MD5 checksums, partially derived from the
FreeBSD md5(1) program. MD5 checksums may be generated by using the
"-m" flag to cksum, or by invoking this program as "md5".

Sync with 4.4lite2

merged with 4.4Lite

Add sum(1) alias, per Matt Green. Various other cleanup. Someone should
finish the addition to the HISTORY section.

update to the latest cksum utility from uunet. this one is posix
compliant (i think), and fixes a bogon in the old CRC calculation.
yes, that's right, the CRC's it generates are different.

Add RCS indentifiers.

branches: 1.1.1;
Initial revision

bump date

PR/54768: elo: Missing cksum(1) convenience links (and docs) for new sha2
functions.

branches: 1.48.6;
Remove workaround for ancient HTML generation code.

New sentence, new line.

PR/49134: Kamil Rytarowski: Add -q flag.

branches: 1.45.6;
Remove unknown \*(tN and \*(sP macros.

ok wiz.

branches: 1.44.2;
Omit outdated cryptographic claims as noted in PR 44160.

Interested readers can follow the references or read Wikipedia; this
is the wrong place to explain cryptographic hash functions and give
security advice.

branches: 1.43.6; 1.43.8;
Use plain \* not \\*. Fix trailing whitespace

Don't use macros for the synopsis. Improve markup to break argument
groups together by not splitting opening and closing macro over lines.

branches: 1.41.2;
Clean up the usage display to match reality.
Allow 'sum' to take '-a algorithm', so that its usage matches 'cksum'.
Per discussion with Igor Sobrado resulting from PR 39658.

branches: 1.40.2; 1.40.12; 1.40.24; 1.40.26; 1.40.30;
Bump date for previous.

mention we support sha2, pointed out by dieter roelants

okay hubertf

remove -1, -2, -4, -5, -6, -m.

Document more explicitly that -c takes an optional file argument.

Allow reading the checksum list from stdin if no file is given with -c.
This allows verifying the checksums e.g. after downloading a NetBSD
release:
cat BSDSUM CKSUM MD5 SHA512 SYSVSUM | cksum -c

A fix for HTML output; punctuation nits.

Document -w, too.

Add a switch "-c file" to verify a list of checksums generated by the
cksum program and stored in file against files on disk. E.g. first
run "md5 *.tgz >MD5" to generate a list of MD5 checksums in MD5, then
use the following command to verify them use "md5 -c MD5"

Inspired by Linux' md5sum,
called for by Jukka Salmi <j+nbsd@2006.salmi.ch>on netbsd-help@,
reviewed by rui@.

Use correct macro. From YOMURA Masanori in private mail.

Punctuation nits.

Add support for SHA2 in cksum.

Put back in information about the deprecated flags in the man-page. Be
explicit about them being removed, and suggest using the ``-a'' flag.

Note about deprecated flags.

Add comments about intentionally not documenting the deprecated -1, -2, -4,
-5, -6, and -m flags so they are not mistakenly get documented again in the
future.

Some changes to cksum:

1. Deprecate the -1, -2, -4, -5, -6, and -m flags. For now, simply remove
them from the documentation.

2. Add and document ``-a algorithm''.

branches: 1.25.2;
-p, -s, -t, and -x are mutually exclusive -- make it obvious in
the SYNOPSIS.

s/the the/the/ (only in sources that aren't regularly imported from
elsewhere).

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22365, verified by myself.

End sentence with a dot.

Fix typo: extra 'a' word in plural sentence. Approved by wiz.

Bump date for last.

Add a -n flag for the hash output forms, so that the output is
"normal", ie, "hash filename" (like all the simple checksum output
forms) as opposed to "hashname (filename) = hash". This output form
is, imho, somewhat more useful because you can pass it directly to
sort to find identical files. For example:

md5 * | sort | uniq -c | grep -v ' 1 '

New sentence, new line.

branches: 1.17.2;
Generate <>& symbolically. I'm avoiding .../dist/... directories for now.

Add sections to xrefs, and remove some unnecessary whitespace while I'm
here.

Add MD2, MD4, SHA1, and RMD160 support to cksum. Adding new hashes to
cksum(1) should now be easy.

- Don't overstrike the |.
- Use the default .Ar argument.

branches: 1.13.10;
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.

add missing .Nm sections.

branches: 1.11.2;
WARNSify

Fix typo, per Dave Huang <khym@bga.com>, PR #3345.

Add support for generating MD5 checksums, partially derived from the
FreeBSD md5(1) program. MD5 checksums may be generated by using the
"-m" flag to cksum, or by invoking this program as "md5".

Sync with 4.4lite2

merged with 4.4Lite

Add sum(1) alias, per Matt Green. Various other cleanup. Someone should
finish the addition to the HISTORY section.

Fix formatting bug: a "1" or "2" argument to -o is manditory.

Use ".St -p1003.2-92" macro instead of spelling it out.

update to the latest cksum utility from uunet. this one is posix
compliant (i think), and fixes a bogon in the old CRC calculation.
yes, that's right, the CRC's it generates are different.

Add RCS indentifiers.

branches: 1.1.1;
Initial revision

cksum.c: don't fault if the hash algorithm is fed invalid data

Addresses PR bin/56897 from RVP, who provided the patch.

cksum: remove trailing whitespace

cksum: fix lint warning about wrong call to isspace

cksum: use POSIX type uint32_t instead of u_int32_t

fix some error handling.

PR/49134: Kamil Rytarowski: Add -q flag.

branches: 1.46.4;
fix unused variable warnings

branches: 1.45.2; 1.45.8;
static + __dead

Clean up the usage display to match reality.
Allow 'sum' to take '-a algorithm', so that its usage matches 'cksum'.
Per discussion with Igor Sobrado resulting from PR 39658.

use a const struct hash pointer

Remove the \n and tabs from the __COPYRIGHT() strings.
Tweak to use a consistent format.

branches: 1.41.12;
Remove duplicate #includes
From Slava Semushin <slava.semushin@gmail.com>, via private mail

kill crypto/rmd160.h and crypto/sha2.h, and instead make symlinks to
/usr/include from /usr/include/sys. This makes all the one way hash
header handling identical.

use c99 initializer.

PR/34584: Geoff C. Wing: cksum not closing filescriptors after finishing
processing files.
Applied patch, thanks!

Be more liberal about whitespaces when verifying checksums.
This allows checking md5/sha1 shecksums generates on Linux, which
add more whitespace.

Based on patch by Jukka Salmi (j+nbsd at 2006 dot salmi dot ch)

remove -1, -2, -4, -5, -6, -m.

Allow reading the checksum list from stdin if no file is given with -c.
This allows verifying the checksums e.g. after downloading a NetBSD
release:
cat BSDSUM CKSUM MD5 SHA512 SYSVSUM | cksum -c

Try guessing hash types. Prefix output with hash type,
the output format was inspired by Linux & OpenBSD.

This now allows doing something like:

miyu% ( md4 *.c ; md5 *.c ; sha1 *.c ) > /tmp/x
miyu% md5 -c /tmp/x
miyu% cksum -c /tmp/x
miyu% echo $?
0
miyu% date >test.c
miyu% cksum -c /tmp/x
(MD4) test.c: FAILED
(MD5) test.c: FAILED
(SHA1) test.c: FAILED
miyu% echo $?
1

Add a switch "-c file" to verify a list of checksums generated by the
cksum program and stored in file against files on disk. E.g. first
run "md5 *.tgz >MD5" to generate a list of MD5 checksums in MD5, then
use the following command to verify them use "md5 -c MD5"

Inspired by Linux' md5sum,
called for by Jukka Salmi <j+nbsd@2006.salmi.ch>on netbsd-help@,
reviewed by rui@.

sync usage().

Use crypto/rmd160.h.

Avoid defining max. hash length ugliness simply by passing NULL for the
buffer to the File routines.

Noted by pooka@.

Forgot +1 for NUL termination.

Use MAXHASHLEN to indicate largest possible hash.
Fixes buffer overflow found by xtraeme@.

Add support for SHA2 in cksum.

Add comments about intentionally not documenting the deprecated -1, -2, -4,
-5, -6, and -m flags so they are not mistakenly get documented again in the
future.

Some changes to cksum:

1. Deprecate the -1, -2, -4, -5, -6, and -m flags. For now, simply remove
them from the documentation.

2. Add and document ``-a algorithm''.

branches: 1.24.2;
De-register.

whitespace, remove main() prototype.

* Kill __P()
* ANSIfy
* No parenthesis are needed around the return value

Add missing -n, -1, -2, -4, -5, and -6 to cksum usage.
From Kouichirou Hiratsuka in PR 26195.

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

branches: 1.19.2;
Update to keep file lengths in appropriate data types (off_t, as opposed
to u_int32_t).

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22365, verified by myself.

Add a -n flag for the hash output forms, so that the output is
"normal", ie, "hash filename" (like all the simple checksum output
forms) as opposed to "hashname (filename) = hash". This output form
is, imho, somewhat more useful because you can pass it directly to
sort to find identical files. For example:

md5 * | sort | uniq -c | grep -v ' 1 '

branches: 1.16.2;
Protect __RCSID and __COPYRIGHT from being invoked if not defined.

Apparently changing crc() to ccrc() was bad, since mtree does a reachover.
Revert ccrc() to crc() and change local variables called crc to thecrc
instead.

Rename the crc() function to ccrc() since that removes a compiler warning
when WARNS=2.

Add MD2, MD4, SHA1, and RMD160 support to cksum. Adding new hashes to
cksum(1) should now be easy.

convert to use getprogname()

fix nested extern

WARNSify

POSIX.2 utilities must call setlocale(LC_ALL, "");

Add support for generating MD5 checksums, partially derived from the
FreeBSD md5(1) program. MD5 checksums may be generated by using the
"-m" flag to cksum, or by invoking this program as "md5".

Sync with 4.4lite2

merged with 4.4Lite

Add sum(1) alias, per Matt Green. Various other cleanup. Someone should
finish the addition to the HISTORY section.

cksum uses 32-bit quantites, internally

update to the latest cksum utility from uunet. this one is posix
compliant (i think), and fixes a bogon in the old CRC calculation.
yes, that's right, the CRC's it generates are different.

Add RCS identifiers.

branches: 1.1.1;
Initial revision

cksum: use POSIX type uint32_t instead of u_int32_t

Explicitly cast the number of bytes passed to crc_buf() to size_t, as
we made sure it is positive already.
Fixes the build.

Speed up cksum word-by-word with a 15 KB table generated dynamically.

(Really we could generate the main table dynamically too.)

From anonymous.

Fix build.

branches: 1.18.60;
Separate out the CRC functions in a manner that actually makes them useful
for other program binaries.
Fixes build of 'config' where I added code (commented out) to use the CRC
function - but didn't comment out the #include :-(

De-register.

* Kill __P()
* ANSIfy
* No parenthesis are needed around the return value

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

branches: 1.14.2;
Update to keep file lengths in appropriate data types (off_t, as opposed
to u_int32_t).

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22365, verified by myself.

branches: 1.12.2;
Protect __RCSID and __COPYRIGHT from being invoked if not defined.

Remove calculation of crc_total from here, making the routine approximately
twice as fast.

Apparently changing crc() to ccrc() was bad, since mtree does a reachover.
Revert ccrc() to crc() and change local variables called crc to thecrc
instead.

Rename the crc() function to ccrc() since that removes a compiler warning
when WARNS=2.

WARNSify

use const qualifer for crctab

what's the point of having prototypes if they're not actually _used_
in the appropriate places? include extern.h in these files, and fix
up the bugs that its inclusion uncovered.

merged with 4.4Lite

cksum uses 32-bit quantites, internally

update to the latest cksum utility from uunet. this one is posix
compliant (i think), and fixes a bogon in the old CRC calculation.
yes, that's right, the CRC's it generates are different.

Add RCS identifiers.

branches: 1.1.1;
Initial revision

cksum: use POSIX type uint32_t instead of u_int32_t

Separate out the CRC functions in a manner that actually makes them useful
for other program binaries.
Fixes build of 'config' where I added code (commented out) to use the CRC
function - but didn't comment out the #include :-(

cksum: use POSIX type uint32_t instead of u_int32_t

Separate out the CRC functions in a manner that actually makes them useful
for other program binaries.
Fixes build of 'config' where I added code (commented out) to use the CRC
function - but didn't comment out the #include :-(

Avoid defining max. hash length ugliness simply by passing NULL for the
buffer to the File routines.

Noted by pooka@.

Use MAXHASHLEN to indicate largest possible hash.
Fixes buffer overflow found by xtraeme@.

Add support for SHA2 in cksum.

branches: 1.15.2;
whitespace, remove main() prototype.

* Kill __P()
* ANSIfy
* No parenthesis are needed around the return value

Update to keep file lengths in appropriate data types (off_t, as opposed
to u_int32_t).

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).

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22365, verified by myself.

branches: 1.10.2;
Add the necessary bits to make cksum buildable as a host tool.

Apparently changing crc() to ccrc() was bad, since mtree does a reachover.
Revert ccrc() to crc() and change local variables called crc to thecrc
instead.

Rename the crc() function to ccrc() since that removes a compiler warning
when WARNS=2.

Add MD2, MD4, SHA1, and RMD160 support to cksum. Adding new hashes to
cksum(1) should now be easy.

Add support for generating MD5 checksums, partially derived from the
FreeBSD md5(1) program. MD5 checksums may be generated by using the
"-m" flag to cksum, or by invoking this program as "md5".

merged with 4.4Lite

cksum uses 32-bit quantites, internally

update to the latest cksum utility from uunet. this one is posix
compliant (i think), and fixes a bogon in the old CRC calculation.
yes, that's right, the CRC's it generates are different.

Add RCS identifiers.

branches: 1.1.1;
Initial revision

Add MD2, MD4, SHA1, and RMD160 support to cksum. Adding new hashes to
cksum(1) should now be easy.

Add MD2, MD4, SHA1, and RMD160 support to cksum. Adding new hashes to
cksum(1) should now be easy.

fix 64 bit time_t

Fix some lint warnings.

branches: 1.8.2;
whitespace, remove main() prototype.

ANSIfy another function missed in previous.

* Kill __P()
* ANSIfy
* No parenthesis are needed around the return value

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

branches: 1.4.4;
Protect __RCSID and __COPYRIGHT from being invoked if not defined.

Add MD2, MD4, SHA1, and RMD160 support to cksum. Adding new hashes to
cksum(1) should now be easy.

WARNSify

Add support for generating MD5 checksums, partially derived from the
FreeBSD md5(1) program. MD5 checksums may be generated by using the
"-m" flag to cksum, or by invoking this program as "md5".

cksum: use POSIX type uint32_t instead of u_int32_t

* Kill __P()
* ANSIfy
* No parenthesis are needed around the return value

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

branches: 1.9.2;
Update to keep file lengths in appropriate data types (off_t, as opposed
to u_int32_t).

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22365, verified by myself.

branches: 1.7.2;
Protect __RCSID and __COPYRIGHT from being invoked if not defined.

WARNSify

merged with 4.4Lite

cksum uses 32-bit quantites, internally

update to the latest cksum utility from uunet. this one is posix
compliant (i think), and fixes a bogon in the old CRC calculation.
yes, that's right, the CRC's it generates are different.

Add RCS identifiers.

branches: 1.1.1;
Initial revision

kill crypto/rmd160.h and crypto/sha2.h, and instead make symlinks to
/usr/include from /usr/include/sys. This makes all the one way hash
header handling identical.

Use crypto/rmd160.h.

Add MD2, MD4, SHA1, and RMD160 support to cksum. Adding new hashes to
cksum(1) should now be easy.

Add MD2, MD4, SHA1, and RMD160 support to cksum. Adding new hashes to
cksum(1) should now be easy.

kill crypto/rmd160.h and crypto/sha2.h, and instead make symlinks to
/usr/include from /usr/include/sys. This makes all the one way hash
header handling identical.

branches: 1.2.2;
Fix host tool build which got broken when SHA2 support was added.

Add support for SHA2 in cksum.

kill crypto/rmd160.h and crypto/sha2.h, and instead make symlinks to
/usr/include from /usr/include/sys. This makes all the one way hash
header handling identical.

branches: 1.2.2;
Fix host tool build which got broken when SHA2 support was added.

Add support for SHA2 in cksum.

kill crypto/rmd160.h and crypto/sha2.h, and instead make symlinks to
/usr/include from /usr/include/sys. This makes all the one way hash
header handling identical.

branches: 1.2.2;
Fix host tool build which got broken when SHA2 support was added.

Add support for SHA2 in cksum.

cksum: use POSIX type uint32_t instead of u_int32_t

De-register.

* Kill __P()
* ANSIfy
* No parenthesis are needed around the return value

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

branches: 1.10.2;
Update to keep file lengths in appropriate data types (off_t, as opposed
to u_int32_t).

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22365, verified by myself.

branches: 1.8.2;
Protect __RCSID and __COPYRIGHT from being invoked if not defined.

Apparently changing crc() to ccrc() was bad, since mtree does a reachover.
Revert ccrc() to crc() and change local variables called crc to thecrc
instead.

WARNSify

what's the point of having prototypes if they're not actually _used_
in the appropriate places? include extern.h in these files, and fix
up the bugs that its inclusion uncovered.

merged with 4.4Lite

update to the latest cksum utility from uunet. this one is posix
compliant (i think), and fixes a bogon in the old CRC calculation.
yes, that's right, the CRC's it generates are different.

Add RCS identifiers.

branches: 1.1.1;
Initial revision

cksum: use POSIX type uint32_t instead of u_int32_t

De-register.

* Kill __P()
* ANSIfy
* No parenthesis are needed around the return value

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

branches: 1.10.2;
Update to keep file lengths in appropriate data types (off_t, as opposed
to u_int32_t).

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22365, verified by myself.

branches: 1.8.2;
Protect __RCSID and __COPYRIGHT from being invoked if not defined.

Apparently changing crc() to ccrc() was bad, since mtree does a reachover.
Revert ccrc() to crc() and change local variables called crc to thecrc
instead.

WARNSify

what's the point of having prototypes if they're not actually _used_
in the appropriate places? include extern.h in these files, and fix
up the bugs that its inclusion uncovered.

merged with 4.4Lite

update to the latest cksum utility from uunet. this one is posix
compliant (i think), and fixes a bogon in the old CRC calculation.
yes, that's right, the CRC's it generates are different.

Add RCS identifiers.

branches: 1.1.1;
Initial revision