libc: modernize lint stubs

Use prototype function definitions instead of identifier lists, as C23
does not support the latter.

Remove redundant /*ARGSUSED*/ comments and parentheses.

branches: 1.4.112;
sync argument/return type of [hn]to[nh][ls] to XNET 5.2 (uint{16,32}_t).
as discussed on tech-net.

branches: 1.3.4;
fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well. No functional changes.

branches: 1.2.6;
Make these work again, by undef'ing the macro that could cause syntax errors.
Also add NOSTRICT in the {h,n}to{n,h}s() cases where the prototype would get
widened...

branches: 1.1.2;
lint stubs for functions which ports use assembly versions. When using
an assembly version of a function, add the right lint stub to the Makefile.inc
which specifies the source to build the function (see the
arch/i386/*/Makefile.inc). Can't just lint the normal C versions of these
functions, because some of them don't _have_ normal C versions and if
that were done dependencies would get messed up.

libc: modernize lint stubs

Use prototype function definitions instead of identifier lists, as C23
does not support the latter.

Remove redundant /*ARGSUSED*/ comments and parentheses.

branches: 1.4.112;
sync argument/return type of [hn]to[nh][ls] to XNET 5.2 (uint{16,32}_t).
as discussed on tech-net.

branches: 1.3.4;
fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well. No functional changes.

branches: 1.2.6;
Make these work again, by undef'ing the macro that could cause syntax errors.
Also add NOSTRICT in the {h,n}to{n,h}s() cases where the prototype would get
widened...

branches: 1.1.2;
lint stubs for functions which ports use assembly versions. When using
an assembly version of a function, add the right lint stub to the Makefile.inc
which specifies the source to build the function (see the
arch/i386/*/Makefile.inc). Can't just lint the normal C versions of these
functions, because some of them don't _have_ normal C versions and if
that were done dependencies would get messed up.

libc: modernize lint stubs

Use prototype function definitions instead of identifier lists, as C23
does not support the latter.

Remove redundant /*ARGSUSED*/ comments and parentheses.

branches: 1.4.112;
sync argument/return type of [hn]to[nh][ls] to XNET 5.2 (uint{16,32}_t).
as discussed on tech-net.

branches: 1.3.4;
fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well. No functional changes.

branches: 1.2.6;
Make these work again, by undef'ing the macro that could cause syntax errors.
Also add NOSTRICT in the {h,n}to{n,h}s() cases where the prototype would get
widened...

branches: 1.1.2;
lint stubs for functions which ports use assembly versions. When using
an assembly version of a function, add the right lint stub to the Makefile.inc
which specifies the source to build the function (see the
arch/i386/*/Makefile.inc). Can't just lint the normal C versions of these
functions, because some of them don't _have_ normal C versions and if
that were done dependencies would get messed up.

libc: modernize lint stubs

Use prototype function definitions instead of identifier lists, as C23
does not support the latter.

Remove redundant /*ARGSUSED*/ comments and parentheses.

branches: 1.4.112;
sync argument/return type of [hn]to[nh][ls] to XNET 5.2 (uint{16,32}_t).
as discussed on tech-net.

branches: 1.3.4;
fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well. No functional changes.

branches: 1.2.6;
Make these work again, by undef'ing the macro that could cause syntax errors.
Also add NOSTRICT in the {h,n}to{n,h}s() cases where the prototype would get
widened...

branches: 1.1.2;
lint stubs for functions which ports use assembly versions. When using
an assembly version of a function, add the right lint stub to the Makefile.inc
which specifies the source to build the function (see the
arch/i386/*/Makefile.inc). Can't just lint the normal C versions of these
functions, because some of them don't _have_ normal C versions and if
that were done dependencies would get messed up.

Revert CC_WNO_USE_AFTER_FREE from Makefile's (thanks uwe@)

Sprinkle CC_WNO_USE_AFTER_FREE for GCC 12

All of them are blamed for idiom equivalent to:
newbuf = realloc(buf, size);
p = newbuf + (p - buf);

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.

bsd.own.mk: rename to CC_WNO_ADDRESS_OF_PACKED_MEMBER

Provide a single variable
CC_WNO_ADDRESS_OF_PACKED_MEMBER
with options for both clang and gcc, to replace
CLANG_NO_ADDR_OF_PACKED_MEMBER
CC_NO_ADDR_OF_PACKED_MEMBER
GCC_NO_ADDR_OF_PACKED_MEMBER

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

new GCC_NO_* uses for warning issues. most of the users of the new
GCC_NO_RETURN_LOCAL_ADDR are bugs in GCC itself, not the code.

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."

More cases to disable gcc-8 warnings only when we're using gcc>=8

Add -Wno-error for gcc-8 warnings

Add userland support for SCTP and manpages.

branches: 1.87.4; 1.87.6;
Include IPv6 global variable in USE_INET6=no libc

This ensures a binary built with USE_INET6=yes libc can still link at
runtime with a USE_INET6=no libc. Of course IPv6 functionnality is not
available, but dynamic linking is not killed by missing symbols such
as in6addr_any.

branches: 1.86.8;
remove various HAVE_GCC=45 fragments.

branches: 1.85.4;
document the scopeid functions

add inet6_scopeid

Retire OSI network stack. OK core@

Avoid gcc 4.5.3 compiler bug on ia64 by compiling hesiod.c with -O1 only.

branches: 1.81.6;
Change CMSG_SPACE and CMSG_LEN to provide Integer Constant Expressions
again. This was changed in sys/socket.h r1.51 to work around fallout
from the IPv6 aux data migration. It broke the historic ABI on some
platforms. This commit restores compatibility for netbsd32 code on such
platforms and provides a template for future changes to the CMSG_*
alignment. Revert PCC/Clang workarounds in postfix and tmux.

Apply fix for previous changes per request from joerg@:
http://mail-index.NetBSD.org/source-changes-d/2011/12/25/msg004454.html

XXX1: More Makefiles should be fixed otherwise they will be
copy-and-pasted blindly again and again.
XXX2: Probably we need more explicit and simple macro to check active CC.

Specify "-fno-tree-ter" to getaddrinfo.c and gethnamaddr.c on arm and gcc45
for workaround of a possible optimazation bug.

On my W-ZERO3 and Zaurus, hostname lookup against CNAME by some commands
(ping(8), ftp(1) etc.) fails even though nslookup(8) returns a proper name
against the same CNAME, after NetBSD/arm ports has been switched to
gcc 4.5 since 201110311420Z (i.e 201110310000Z binaries worked but
201111010000Z ones not).

Building getaddrinfo.c and gethnamaddr.c in libc with "-O2 -fno-tree-ter"
(or using objects built by old gcc 4.1) seems to fix this issue.

Accroding to nonaka@, the following gcc bugzilla test case also fails
with -O2 but works with -O2 -fno-tree-ter on NetBSD/zaurus 5.99.57:
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48863#c4
but CNAME lookup didn't fail on his environment even without this workaround.

branches: 1.78.6;
Introduce allocaddrinfo(3) to make writing getaddrinfo() nss plugins easier.
Discussed on tech-net.

another missing \

remove duplicated line

missing \

sort, add missing, fix section

new resolver functions.

Add a new socket option for unix domain sockets: LOCAL_PEEREID, to make
it possible to get the pid, euid and egid of the process at the remote
end at the time it did bind() or connect().

Add a new libc function, getpeereid() to easily get at the euid and egid.
As a consequence, bump libc's minor number.

Document the LOCAL_PEEREID socket option in unix(4).

Based on contribution by Arne H. Juul, minor modifications by myself.

Remove the XNS man pages.

Move ns_addr and ns_ntoa to compat libc. Add a stripped version of
<netns/ns.h> which only include struct ns_addr and the ns_addr and ns_ntoa
function prototypes.

Add support for RFC 3542 Adv. Socket API for IPv6 (which obsoletes 2292).
* RFC 3542 isn't binary compatible with RFC 2292.
* RFC 2292 support is on by default but can be disabled.
* update ping6, telnet and traceroute6 to the new API.

From the KAME project (www.kame.net).
Reviewed by core.

replace with written-from-scratch manpage. (fixes ISOC copyright issue)

Only compile in IPv6 support if ${USE_INET6} != "no"

MKINET6 is for providing IPv6 infrastructure.
USE_INET6 is for compiling IPv6 support into the programs (needs MKINET6).

revert previous change while discussion is in progress.

RFC cannot be used as manpage regarding to the ISOC copyright boilerplate.
remove it until someone writes it up from scratch/freely-redistributable
text is found.

Add MLINK for gethostent as suggested by Matthias Drochner.

Finish bind9 resolver merge.

branches: 1.62.2;
getservent_r/getprotoent_r

Make building IPv6 support controllable by MKINET6

MAKEVERBOSE tweaks

Split the notion of building Hesiod, Kerberos, S/key, and YP
infrastructure and using that infrastructure in programs.

* MKHESIOD, MKKERBEROS, MKSKEY, and MKYP control building
of the infratsructure (libraries, support programs, etc.)

* USE_HESIOD, USE_KERBEROS, USE_SKEY, and USE_YP control
building of support for using the corresponding API
in various libraries/programs that can use it.

As discussed on tech-toolchain.

Add a sockatmark(3), the 1003.1g wrapper around SIOCATMARK.

remove obsolete MAN lines for getipnodeby* (has been comented out)

branches: 1.56.2;
it seems that gethostent(3) is no longer supplied in libc.
couple of wording corrections.
From: Hubert Feyrer <hubert.feyrer@informatik.fh-regensburg.de>

to get (f)lex output to go to stdout, use -t (a flag supported by
flex) rather than -o/dev/stdout which uses a flex feature (-o isn't
in vendor lex, at least on solaris 7) and also requires that
"/dev/stdout" function in the expected manner.

use BSDI-origin if_{index,name}* functions. they use getifaddrs(3) as
backend and are more robust against SIOCGIFCONF alignment issue.
now getifaddrs always takes care of the issue. (sync with kame tree)
use weak symbol just in case.

Listing generated files nslexer.c nsparser.c in SRCS breaks make depend
in a clean dir. Put the actual src files nslexer.l nsparser.y there.

make CMSG_ALIGN always synchronize with kernel's idea of ALIGNBYTES.
ancillary data alignment will be ALIGNBYTES, not sizeof(long) - 1, from now.

CMSG_xx will NOT resolve into constant. if you use CMSG_xx to allocate
arrays, you'll lose.

bump shlib minor for libc.

NOTE: if you are on top of arch with ALIGNBYTES != sizeof(long) - 1,
you need to recompile IPv6-related binaries. there is no way to guarantee
backward compat in this aspect. sorry for this. this should be the last
backward compat breakage for IPv6-related ancillary data manipulation.
(we still have PR 9516 for unix-domain sockets...)

add freeifaddrs.3 as MLINKS.

add getifaddrs(3) from bsdi4, as mentioned in tech-net (in SIOCGIFCONF
related thread).
this will save future apps from complexity of SIOCGIFCONF.

getifaddrs(3) does not use SIOCGIFCONF internally, it uses sysctl.

add missing manpage, gethostbyname2(3).

add IPv6 support to rcmd() and friends.
rcmd() and iruserok(): unchanged
ruserok(): became address family independent
rcmd_af(), iruesrok_sa(): address family independent

move -DINET6 upwards, into lib/libc/Makefile.
- we need it for lib/libc/rpc soon
- this is not clear to add CPPFLAGS in subdir/Makefile.inc (as it affect other
subdirs as well)

add manpage for inet6_{option,rthdr}_*.

add freeaddrinfo(3) and gai_strerror(3) as links to getaddrinfo(3)

branches: 1.44.4;
Add indirect reference stubs for inet_[ap]ton() for environments not supporting
weak aliases; addresses PR lib/7435 but leaves the original file intact.

add IPv6 support functions.

get{addr,name} info are implemented to have as little impact to existing
resolver code as possible, so they are NOT the optimal implementation.
They are at this moment not very thread safe (as they call
gethostby{name,addr}).

(shlib minor version)++, as new interfaces are added.

TODO: getipnodeby{name,addr} - which needs total reimplementation of
gethostby{name,addr}.
upgrade rcmd.c for multiple af support (needed for IPv6-ready rsh/rlogin)

Do a lame hack to reduce the size of the lexer state buffer, which reduces
memory consumption substantially. The buffer is only used to print line
numbers...

branches: 1.41.2;
use ${ARCHDIR} instead of ${.CURDIR}/arch/${MACHINE_ARCH} where appropriate
libc builds on pmax again

add hesiod(3) et al

pull nsswitch up to main branch
convert getnetby*() to nsswitch (from resolv.conf lookup)

merge nsswitch into main branch

install links for inet_ntop(3) and inet_pton(3)

Add sorting of resolved addresses.

add missing .Nm entries. from spz@serpens.swb.de.

branches: 1.34.2;
Make linkntoa.3 a link to linkaddr.3 again; fixes PR lib/4017 from maximum
entropy <entropy@tardis.bernstein.com>.

merge bind 4.9.6 resolver. still need to fix <resolv.h>

inet_pton4() did not handle
- incompletely specified ip addresses
- ip addresses specified in hex
like inet_aton() and inet_addr(). This broke parsing of /etc/hosts
(127.1 localhost stopped working before the resolver is started).
The fix is to use the inet_aton() code to implement inet_pton4(), and
then use the new inet_pton4() code to implement inet_aton() and inet_addr().
At that point inet_addr.c is empty and has been removed.

* add man page for inet_net_ntop(3) and inet_net_pton(3)
* xref them in inet(3)
* use Va instead of Fn for 'errno' in inet(3)

install a man page (link) for orcmd(3).

branches: 1.29.2;
create indirect references (or function stubs) to dn_comp(), res_close()
and res_send(), as required.

merge in the bind 4.9.5-P1 resolver. we are now _mostly_ the
same as this, with our changes also included (resolv.conf and
other minor differences).

Add iso_addr.3 to MAN. Fixes misc/2942.

branches: 1.26.4;
Install linkaddr.3 as link_addr.3 and link_ntoa.3 too.

Added explicit return types to all functions in SYNOPSIS section.
Added documentation to hstrerror() function.
Added hstrerror MLINK.

merge with the resolver in bind 4.9.3-P1.

include proper links to ethers.3

mention the functions that need to be provided, instead of providing dflts.

clean up import. also convert everything to new Id format.

list the files needed on the alpha.

specify man pages the new way.

make sure link for inet_aton is in alphabetical order in MLINKS.

Include files for pmax

link inet.3 for inet_aton()

fix bug 535, reported by Brian Moore.

ethers(3) man page by Roland McGrath <roland@frob.com>

ethers(3) sunos-like support from Roland McGrath <roland@frob.com>
YP code fixed by myself (looked in wrong maps, didn't free things :-)
used by rarpd & friends.

Added some Sparc sources.

Changed occurance of MACHINE_ARCH == "hp300" to MACHINE_ARCH == "m68k"

get pc532 moved into ns32k; that's is MACHINE_ARCH

split up the {ntoh,hton}[ls] functions

update for .S and MACHINE_ARCH

make an arch directory in libc, and move all machine dirs into it

pc532 specifics.

Provide C language versions of the netorder functions.
These may be "good enough" for big-endian systems that do not have assembly
language versions of their own. A compiler should be able to do a fairly
good job optimising them, it probably won't be smart enough to omit the
stack frame, but then again, these functions won't be called unless the
macro versions are undef'd.
On the other hand, they are only intended for bootstrap purposes on little-
endian systems. They should be replaced with assembly language versions as
soon as possible.

Add RCS identifiers.

Needed for netstat to compile, add iso_addr &c
(from Havard Eidnes <Havard.Eidnes@runit.sintef.no>)

Compile {iso,ns}_{addr,ntoa}.

branches: 1.1.1;
Initial revision

Use reach-over sources from common/lib/libc

A little libc namespace housekeeping exercise:
* Make vfprintf_unlocked() an internal function, c.f. __svfscanf_unlocked().
* Add internal names for arc4random(), endnetpath(), fhstatvfs(),
fstatvfs(), mkstemp(), shquote(), statvfs(), taddr2uaddr(), uaddr2taddr(),
uuid_create_nil(), uuid_is_nil(), and wcwidth().
* Include namespace.h where supposed to.

whitespace KNF

LP64 warnings

add __RCSID().

change copyright notice from 3-clause TNF -> 4-clause TNF.
not sure if which we prefer, but the point is to use exactly the same
copyright. From: matthew green <mrg@eterna.com.au>

make CMSG_ALIGN always synchronize with kernel's idea of ALIGNBYTES.
ancillary data alignment will be ALIGNBYTES, not sizeof(long) - 1, from now.

CMSG_xx will NOT resolve into constant. if you use CMSG_xx to allocate
arrays, you'll lose.

bump shlib minor for libc.

NOTE: if you are on top of arch with ALIGNBYTES != sizeof(long) - 1,
you need to recompile IPv6-related binaries. there is no way to guarantee
backward compat in this aspect. sorry for this. this should be the last
backward compat breakage for IPv6-related ancillary data manipulation.
(we still have PR 9516 for unix-domain sockets...)

Finish bind9 resolver merge.

__warn_references and __indr_reference semicolon fixups. From
Chris Demetriou <cgd@pa.dec.com>.

branches: 1.3.2;
Do the non-__indr_reference verions of thse like the cat*(3) functions,
modulo a slight kludge to get around the namespace trickery played
in <resolv.h>.

pull in <netinet/in.h>, s/#Include/#include, and fix up __dn_comp() prototype,
so that this compiles.

create indirect references (or function stubs) to dn_comp(), res_close()
and res_send(), as required.

Finish bind9 resolver merge.

This file was incorrect. The indirect name is res_close(),
not _res_close()!

__warn_references and __indr_reference semicolon fixups. From
Chris Demetriou <cgd@pa.dec.com>.

branches: 1.3.2;
Do the non-__indr_reference verions of thse like the cat*(3) functions,
modulo a slight kludge to get around the namespace trickery played
in <resolv.h>.

pull in <netinet/in.h> so that these compile

create indirect references (or function stubs) to dn_comp(), res_close()
and res_send(), as required.

Finish bind9 resolver merge.

__warn_references and __indr_reference semicolon fixups. From
Chris Demetriou <cgd@pa.dec.com>.

branches: 1.3.2;
Do the non-__indr_reference verions of thse like the cat*(3) functions,
modulo a slight kludge to get around the namespace trickery played
in <resolv.h>.

pull in <netinet/in.h> so that these compile

create indirect references (or function stubs) to dn_comp(), res_close()
and res_send(), as required.

Finish bind9 resolver merge.

Add indirect reference stubs for inet_[ap]ton() for environments not supporting
weak aliases; addresses PR lib/7435 but leaves the original file intact.

Finish bind9 resolver merge.

Add indirect reference stubs for inet_[ap]ton() for environments not supporting
weak aliases; addresses PR lib/7435 but leaves the original file intact.

Catch up with all the lint warnings since exit on warning was disabled.
Disable 'missing header declaration' and 'nested extern' warnings for now.

knf, no functional change.

Don't read past the end when the data is exactly the right size. Reported
by tedu @ openbsd in tech-userlevel. Thanks!

branches: 1.14.10;
Update old-style definitions to ANSI, remove a couple of register
definitions along the way. Fixed gcc 4.1 build (thank you vax)

PR/45989: Martin Husemann: lint invocation does include -w only on i386

- turn lint -w for all the platforms after fixing the lint warnings.
- add _DIAGASSERTS() for casts that would assign values to types that
would not fit.
- change types, add casts
- change into ansii prototypes
- turn on _DIAGNOSTIC for libc (during current, to be eliminated for release
builds)

approved by core@

branches: 1.12.6;
merge libbind and bump.

branches: 1.11.4; 1.11.6; 1.11.8; 1.11.10; 1.11.14;
updates from bind-9-5-0

branches: 1.10.12;
Resolve conflicts.

merge in resolver portion of bind-9.4.0-rc2

branches: 1.8.16; 1.8.18;
Fix signed/unsigned comparison warnings.

branches: 1.7.2;
Passing "char" values to ctype(3) functions is problematic.
If an argument of a ctype function is outside "unsigned char"
and if it is not EOF, the behavior is undefined.

The isascii(3) is the sole exception of above and it was used to
be used to check a value was valid for other ctype functions in
ancient systems. On modern systems, the ctype functions take
all values of "unsigned char", and this check is obsolete and
even wrong for non-ASCII systems. However, we leave the isascii()
untouched for now, so as not to change the current behavior.

back out the #ifdef _DIAGNOSTIC argument checks; too many people complained.
_DIAGASSERT() is still retained.

* use _DIAGASSERT() to check pointer arguments against NULL and file
descriptors against -1 (as appropriate).
* add actual checks which to detect stuff that would trigger_DIAGASSERT(),
and attempt to return a sane error condition.
* knf some code
* remove some `register' decls.

the first two items result in the addition of code similar to the
following in various functions:

_DIAGASSERT(path != NULL)
#ifdef _DIAGNOSTIC
if (path == NULL) {
errno = EFAULT;
return (-1);
}
#endif

delint

Fix RCSID's
Fix gcc warnings
Add missing prototypes
Use "namespace.h"

merge in the bind 4.9.5-P1 resolver. we are now _mostly_ the
same as this, with our changes also included (resolv.conf and
other minor differences).

branches: 1.1.1;
Initial revision

Xref bswap(3). Simplify STANDARDS.

Use different wording as "null macro" is too ambiguous.

typo - s/intger/integer/

Update for use of uint{16,32}_t.

branches: 1.11.2;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22280, verified by myself.

Use
.In header.h
instead of
.Fd #include \*[Lt]header.h\*[Gt]
Much easier to read and write, and supported by groff for ages.
Okayed by ross.

Generate <>& symbolically.

Delete the old BUGS section entry:
> On the VAX bytes are handled backwards from most everyone else in
> the world. This is not expected to be fixed in the near future.

Multiple levels of irony there...

Boring whitespace fixes.

branches: 1.6.12;
Last of the .Os cleanups. .Os is defined in the tmac.doc-common file,
so we shouldn't override it with versions in the manpages. Wheee!

add LIBRARY section to man page

in_addr_t and in_port_t changes

clean up import. also convert everything to new Id format.

Add RCS indentifiers.

branches: 1.1.1;
Initial revision

Fix Bl argument.

Remove comma after last Nm entry.
New sentence, new line

Use
.In header.h
instead of
.Fd #include \*[Lt]header.h\*[Gt]
Much easier to read and write, and supported by groff for ages.
Okayed by ross.

Use "const" for all input parameters of ethers(3) functions.

Make SYNOPSIS accurate about required headers.
Fixes PR lib/11021 from wiz.

Generate <>& symbolically.

Boring whitespace fixes.

branches: 1.8.12;
Last of the .Os cleanups. .Os is defined in the tmac.doc-common file,
so we shouldn't override it with versions in the manpages. Wheee!

macroize BSD, NetBSD, FreeBSD and misc cleanup

add LIBRARY section to man page

add missing .Nm entries. from spz@serpens.swb.de.

* fix ether_line to limit size of parsed hostname to MAXHOSTNAMELEN
* modify ether_aton, ether_hostton, and ether_line to take 'const char *'
arguments as appropriate
* document the above in ethers(3)

branches: 1.3.2;
merge in the bind 4.9.5-P1 resolver. we are now _mostly_ the
same as this, with our changes also included (resolv.conf and
other minor differences).

New ARP system, supports IPv4 over any hardware link.

Some of the stuff (e.g., rarpd, bootpd, dhcpd etc., libsa) still will
only support Ethernet. Tcpdump itself should be ok, but libpcap needs
lot of work.

For the detailed change history, look at the commit log entries for
the is-newarp branch.

branches: 1.1.14;
ethers(3) man page by Roland McGrath <roland@frob.com>

make more descriptors that we open as close-on-exec

use fparseln() instead of fgetln() for consistency (this was the last fgetln()
using parser in net/.
cVS: ----------------------------------------------------------------------

branches: 1.23.2; 1.23.8;
Use C89 Prototypes.
Remove use of __P

PR/45989: Martin Husemann: lint invocation does include -w only on i386

- turn lint -w for all the platforms after fixing the lint warnings.
- add _DIAGASSERTS() for casts that would assign values to types that
would not fit.
- change types, add casts
- change into ansii prototypes
- turn on _DIAGNOSTIC for libc (during current, to be eliminated for release
builds)

approved by core@

branches: 1.21.42;
PR lib/34801 from David A. Holland: "static" comes first in declaration.

Use "const" for all input parameters of ethers(3) functions.

branches: 1.19.2;
Avoid run-time construction of format string we can build at compile time

branches: 1.18.4;
add __RCSID().

Delint.
Remove trailing ; from uses of __weak_alias(). The macro inserts this if
needed.

back out the #ifdef _DIAGNOSTIC argument checks; too many people complained.
_DIAGASSERT() is still retained.

* use _DIAGASSERT() to check pointer arguments against NULL and file
descriptors against -1 (as appropriate).
* add actual checks which to detect stuff that would trigger_DIAGASSERT(),
and attempt to return a sane error condition.
* knf some code
* remove some `register' decls.

the first two items result in the addition of code similar to the
following in various functions:

_DIAGASSERT(path != NULL)
#ifdef _DIAGNOSTIC
if (path == NULL) {
errno = EFAULT;
return (-1);
}
#endif

branches: 1.14.2;
delint

Move an output buffer into the BSS.

Fix compile problem.

* fix ether_line to limit size of parsed hostname to MAXHOSTNAMELEN
* modify ether_aton, ether_hostton, and ether_line to take 'const char *'
arguments as appropriate
* document the above in ethers(3)

branches: 1.10.2;
If port provides __weak_alias(), provide an Standard C and POSIX pure
identifier namespace by renaming non standard functions and variables
such that they have a leading underscore. The library will use those
names internally. Weak aliases are used to provide the original names
to the API.

This is only the first part of this change. It is most of the functions
which are implemented in C for all NetBSD ports. Subsequent changes are
to add the same support to the remaining C files, to assembly files, and
to the automagically generated assembly source used for system calls.
When all of the above is done, ports with weak alias support should add
a definition for __weak_alias to <sys/cdefs.h>.

Fix RCSID's
Fix gcc warnings
Add missing prototypes
Use "namespace.h"

New ARP system, supports IPv4 over any hardware link.

Some of the stuff (e.g., rarpd, bootpd, dhcpd etc., libsa) still will
only support Ethernet. Tcpdump itself should be ok, but libpcap needs
lot of work.

For the detailed change history, look at the commit log entries for
the is-newarp branch.

branches: 1.7.2;
- convert unsafe strcpy(), strcat() and sprintf() to the `n' versions.
- some KNF.

explicitly type ether_{hostton,ntohost,line}(). all return int.

branches: 1.5.4;
clean up import. also convert everything to new Id format.

"struct ether_addr" is in netinet/if_ether.h (as in SunOS)

#include <stdlib.h> & <string.h> to bring function prototypes into scope.

small bug fixed

ethers(3) sunos-like support from Roland McGrath <roland@frob.com>
YP code fixed by myself (looked in wrong maps, didn't free things :-)
used by rarpd & friends.

Use .In for header files instead of .Ar Pa and variations.

We have 2006.

Add EAI_OVERFLOW per RFC3493, POSIX and XPG6.

Use .In for including header files, remove some unnecessary quotes.

written-from-scratch manpage (fixes ISOC copyright violation)

Document that freeaddrinfo(NULL) dumps core.

A source code comment already said this, but it's easier to
find in a man page.

branches: 1.60.6;
getaddrinfo.3: order of the struct members is wrong, correct it

Remove workaround for ancient HTML generation code.

mention ip6addrctl

New sentence, new line.

document AI_SRV

branches: 1.55.6; 1.55.12;
Join URL.

Use the common template for the USENIX paper. Add URL.

Use .In for header files instead of .Ar Pa and variations.

Bump date for previous.

getaddrinfo is defined to use the AF_* constants, even though perhaps its
interface ought to have been defined in terms of PF_*. Add a note about
how the distinction has been lost anyway. Fixes PR lib/42384.

Revert a -> an changes: Adding an 'n' follows pronunciation, not writing.

Bump date for new allocaddrinfo.

fix grammar: a -> an

Introduce allocaddrinfo(3) to make writing getaddrinfo() nss plugins easier.
Discussed on tech-net.

Add missing article.

Document AI_NUMERICSERV.

(When manpage was rewritten some docs were lost.)

Remove comments that do not reflect reality anymore.

Only name <netdb.h> in the synopsis, like the old manual page used to.

Add restrict qualifiers to getaddrinfo() and getnameinfo() arguments.

branches: 1.41.2;
Use In instead of Fd; use \*[Am]\*[Lt]\*[Gt] for html output. Uppercase FREENIX.

title was backwards. djm@openbsd

replace with written-from-scratch manpage. (fixes ISOC copyright issue)

revert previous change while discussion is in progress.

RFC cannot be used as manpage regarding to the ISOC copyright boilerplate.
remove it until someone writes it up from scratch/freely-redistributable
text is found.

Add missing "-".

remove unneeded refs to rfc2553. jmc

document AI_NUMERICSERV

use "%s" for err(), do not pass pointer directly. Pedro Martelletto

Consistently use 'RFC 1234' instead of 'RFC1234' or 'RFC-1234'.
From jmc@openbsd.

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

Patches provided by Joel Baker in PR 22280, verified by myself.

english grammer. by Jason McIntyre via kame

Use
.In header.h
instead of
.Fd #include \*[Lt]header.h\*[Gt]
Much easier to read and write, and supported by groff for ages.
Okayed by ross.

refer nsswitch.conf(5). From: Jim Bernard <jbernard@mines.edu>

ai_addrlen: size_t -> socklen_t.

Rip out <sys/types.h> and <sys/socket.h> inclusions: there are no
pre-requisites for <netdb.h>.

In the examples, initialize struct addrinfo explicitly to either 0 or a
NULL pointer; from der Mouse in PR lib/16634.

generate & symbolically

Generate <>& symbolically.

Sort sections, sort SEE ALSO, and misc. fixes.

Boring whitespace fixes.

there's no host.conf in netbsd. pointed out by smb

branches: 1.19.2;
fix listener side example. from deraadt, sync with kame

- scope separater is percentage, not atmark.
- typo in example. s/err1/errx/.
- some wording improvements.
sync with kame.

english fixup

mention that current get{addr,name}info implementation are not thread-safe.
(sync with kame)
From: Greg Thompson

typo. s/Fa getaddrinfo/Fn getaddrinfo/. sync with kame.

move pitfall example to downwards. wording fixes. sync with kame.

give pitfall examples.
in response to couple of PRs including 10666. sync with kame.

typo in RFC number. KAME PR 263

add cmetz paper, correct example

branches: 1.10.4;
$NetBSD$

update internet draft #. correct samples.
sync getnameinfo function prototype with /usr/include/netdb.h (not
standard conformant yet, needs update - will do soon)

getaddrinfo(3) now obeys nsswitch.conf correctly.

in BUGS section, document that it can behave differently from nsswitch.conf(5)
suggests.
NetBSD PR: 9413

cosmetic (last dot in SEE ALSO section)

sync with latest KAME tree.
add example section to get{addr,name}info.
s/\.Os KAME/.Os/.

need sys/types.h before sys/socket.h

bring in latest KAME get{addr,name}info(3).
getnameinfo(3) globs ai_socktype and ai_protocol correctly.
KAME scopeid extension is implemented.
(draft-ietf-ipngwg-scopedaddr-format-00.txt)

bump up shlib minor
(may not be necessary, but rather big difference in behavior - especially
ai_flags)

branches: 1.2.4;
Add missing LIBRARY sections.

add IPv6 support functions.

get{addr,name} info are implemented to have as little impact to existing
resolver code as possible, so they are NOT the optimal implementation.
They are at this moment not very thread safe (as they call
gethostby{name,addr}).

(shlib minor version)++, as new interfaces are added.

TODO: getipnodeby{name,addr} - which needs total reimplementation of
gethostby{name,addr}.
upgrade rcmd.c for multiple af support (needed for IPv6-ready rsh/rlogin)

Unwrap a line which wasn't wrapped before it had an __UNCONST() added
to it, and now that that is gone again, doesn't need wrapping any more.

That line is now unaltered from what it was 3 revs ago (bit for bit).

NFC.

Make the first argument of ip6_str2scopeid const, instead of __UNCONST.
Thanks kre@. While here s/u_int32_t/uint32_t/

Catch up with all the lint warnings since exit on warning was disabled.
Disable 'missing header declaration' and 'nested extern' warnings for now.

convert explicit length check before unchecked snprintf() with just a
overflow checked snprintf(). for res_debug.c and res_query.c, convert
from sprintf() to snprintf().

tested scp and rcp fail properly with too-long paths.
tested getaddrinfo fails as expected for too-long domains.
tested dig and ping for similar (res_debug.c/res_query.c).
created a temporary fs with quotas to test edquota with a long EDITOR.
did not test ypserv directly, but it's the same pattern.

avoids GCC 12 snprintf() warnings, and reduces total code size.

lib: remove CONSTCOND comment

Since 2021-01-31, lint doesn't need it anymore for the common pattern of
'do ... while (0)'.

Fix typo in comment: allows, not alows.

I don't know what an alow is, maybe some cousin of the alot.
Awaiting a nature documentary by Allie Brosh about these perhaps more
elusive beasts.

PR misc/56473, from jschauma

getaddrinfo(3): malloc + memset -> calloc

Don't complain about additional DNAME records received when resolving A or
AAAA (like already done for SIG/KEY and CNAME).

branches: 1.119.2;
Adjust English usage in message for EAI_NONAME.

Return EAI_NODATA, not EAI_NONAME, for nonexistent hosts.

This causes e.g. "ssh nosuchname" to print "No address associated with
hostname", which is correct, rather than "hostname nor servname
provided, or not known", which is not.

branches: 1.117.2; 1.117.4;
PR/52837: Michael Kaufmann: getaddrinfo() resolves "127.0.0.1 www.example.com"
to 127.0.0.1. Apply the patch from FreeBSD and explain the rationale.

PR/52578: Benjamin M. Schwartz: sync the internal copy of res_nquery for
getaddrinfo, with the real version so that we handle EDNS fallback.

branches: 1.115.6;
KNF, simplify expressions for readability.

branches: 1.114.2;
Avoid shadowing global.

PR/50552: Kyle Amon: RFC3493 section 6.1 wants us to accept non-dotted-quad
INET4 addresses.
XXX: pullup-7

Implement the address selection policy; from FreeBSD

PR/44915: Havard Eidnes:
- Convert NS_NOTFOUND to EAI_NONAME instead
of EAI_NODATA which more inline with what other OS's do.
- Convert NO_DATA to NS_NOTFOUND since there is no equivalent mapping
for nsswitch (perhaps we could add one so what we could recover the exact
error and return EAI_NODATA?)
- If we end up following a CNAME chain that does not find any data return that
instead of internal error.

XXX: pullup-7

PR/50367: Stefan Schaeckeler: libc resolver library does not resolve host
names with underscores ("_"). According to resolv.conf(5) RES_NOCHECKNAME
is on by default; well, it is and gethostbyname(3) obeys it
(gethnamaddr.c:maybe_ok), but getaddrinfo(3) up till this commit not.
XXX: pullup-7, pullup-6

namespace protect allocaddrinfo

mke allocaddrinfo static.

Include several libc functions in rump.ping6

We need getaddrinfo, getifaddrs, if_indextoname and if_nametoindex
in addition to getnameinfo.

branches: 1.106.4; 1.106.6;
this is supposed to be re-entrant, call don't call __hostalias that uses
a static buffer.

CVE 1020946: Fix res leak.

CVE 1020938: Fix memory leak

PR/32373, PR/25827: Add SRV lookup in getaddrinfo(3)
Per DNS-SD (RFC 2782), but only enabled if AI_SRV is set.

KNF, whitespace police -- no code changes

branches: 1.101.2;
Henning Petersen in PR lib/46561: cosmetic change: avoid self assignement.

PR pkg/46206
re-establish fqdn lookup when AI_ADDRCONFIG is used in hints
AI_ADDRCONFIG led to fqdn lookup being skipped as the systems didn't
configure any PF_UNSPEC addresses - check was too strict here.
Thnaks to Ryo ONODERA for testing.

PR 46206: fix programmed SIGSEGV
more work is needed as tests seem to indicate that name resolution now
does no seem to work (firefox reports Server not found)
thanks to Ryo ONODERA for testing.

Add AI_ADDRCONFIG, which makes getaddrinfo() return only address with families
that are already configured in the system.

PR/45989: Martin Husemann: lint invocation does include -w only on i386

- turn lint -w for all the platforms after fixing the lint warnings.
- add _DIAGASSERTS() for casts that would assign values to types that
would not fit.
- change types, add casts
- change into ansii prototypes
- turn on _DIAGNOSTIC for libc (during current, to be eliminated for release
builds)

approved by core@

branches: 1.96.2; 1.96.4;
close on exec fixes:
- open + fcntl -> open O_CLOEXEC
- configuration database file descriptors that can stay open are now opened
fopen(db, "re")

Revert a -> an changes: Adding an 'n' follows pronunciation, not writing.

fix grammar: a -> an

Introduce allocaddrinfo(3) to make writing getaddrinfo() nss plugins easier.
Discussed on tech-net.

Cause getaddrinfo() without socktype specification to work correctly for
numeric addresses. The documentation appears to say this works, and some
other systems support it -- more importantly, why should it _not_ work? If
it does not, getaddrinfo() cannot be used as a general-purpose textual to
binary address conversion utility function; yet it is the only such function
we have in the system, since inet_pton() requires a priori knowledge of the
address family.

This change also causes getaddrinfo() with NULL hint (expressly documented
as working) to work properly for numeric addresses.

branches: 1.91.6; 1.91.8;
Add a missing const.

branches: 1.90.10;
make this compile without INET6 (Scott Ellis)

Since we are going to be looking for ports many times, use a centralized
servent data structure and pass the stayopen hint.

use syslog_r, from OpenBSD

branches: 1.87.2;
fix incomplete initializers

memset the whole struct addrinfo, instead of trying to figure out if we
have a pad member or not.

Call endservent_r. From Hajimu UMEMOTO

use getservbyname_r.

Remove comments that do not reflect reality anymore.

Make macros more readable. No functional change.

Coverity CID 2363: cur is known not to be NULL.

Coverity CID 2364: cur is known not to be NULL.

Coverity CID 2430: cur is known not to be NULL.

Correct obvious typo

Add EAI_OVERFLOW as specified in RFC3493, POSIX and XPG6.

Fix spelling mistake in comment. From OpenBSD.

constify and remove an unnecessary __UNCONST.

WARNS=4

Rest of fix for PR#31184: getaddrinfo() now honors resolv.conf sortlist
directive for dns answers.

Also, unifdef the RESLVSORT (non-)option sillyness.

Reviewed by christos.

make yp stuff re-entrant.

centralize res_*init() calls in __res_get_state()

Finish bind9 resolver merge.

remove prototype decl for removed function

remove unused code

implement AI_NUMERICSERV (as defined in RFC3493)

bound copies

fill ai_canonname field for numeric hostname, by the given hostname.
folows new recommendation in RFC3493

From: "Jason L. Wright" <jason@thought.net>

The man page for getaddrinfo() and RFC2553 says it should return EAI_NONAME
for the case where hostname != NULL, hostname is not a numeric string and
ai_flags has the AI_NUMERICHOST bit set. Instead, it returns EAI_NODATA
for this case. The diff below makes it return EAI_NONAME for this case.

This is why ntpq and ntpdc from ntp-4.1.74 can't resolve hostnames,
but work fine with numeric strings like ::1 or 127.0.0.1.

allocate 64K recieve buffer for DNS responses.

drop AAAA reply with IPv4 mapped address. sync w/kame
draft-itojun-v6ops-v4mapped-harmful-00.txt

pacify lint. from he@netbsd

lint clean. from martin husemann

make more pedantic check on strtoul. from deraadt, sync w/kame

correct port range check. from deraadt, sync w/kame, diff sent to bind-bugs

correct %d/%u mixup.

correct bad practice in the code - it uses two changing variables
to manage buffer (buf and buflen). we eliminate buflen and use
fixed point (ep) as the ending pointer.

this fix is NOT critical.

branches: 1.55.2;
struct addrinfo.ai_addrlen used to be a size_t, per RFC 2553.

In XNS5.2, and subsequently in POSIX-2001 and draft-ietf-ipngwg-rfc2553bis-02
it was changed to a socklen_t.

To accomodate for this while preserving binary compatibility with the
old interface, prepend or append 32 bits of padding, depending on
the (LP64 data model) architecture's endianness.

This should be deleted the next time the libc major number is
incremented.

use snprintf, not sprintf. sync with kame.

branches: 1.53.2;
make local constant arrays const

make sure "qname" argument to getanswer() is properly given. from lukem

sprinkle _DIAGASSERT() around

remove redundant declarations of h_errno and __hostalias

don't bother declaring in6_addrany[] and in6_loopback #ifndef INET6

on /etc/hosts lookup, set "official host name" (the leftmost hostname)
into ai_canonname. this is to synchronize with practice in gethostby*.
comment from ume@mahoroba.org. sync with kame.

implement EDNS0 support, as EDNS0 support will be made mandatory for
IPv6 transport-ready resolvers/DNS servers. need careful configuration
when enable it. (default config is not affected).
see manpage for details.

XXX visible symbol __res_opt() is added, however, it is not supposed to be
called from outside, libc minor is not bumped.

reject empty scopeid/numeric portname explicitly. sync with kame.

use strtoul() to check for all-digit string, not isdigit(). sync with kame.

Passing "char" values to ctype(3) functions is problematic.
If an argument of a ctype function is outside "unsigned char"
and if it is not EOF, the behavior is undefined.

The isascii(3) is the sole exception of above and it was used to
be used to check a value was valid for other ctype functions in
ancient systems. On modern systems, the ctype functions take
all values of "unsigned char", and this check is obsolete and
even wrong for non-ASCII systems. However, we leave the isascii()
untouched for now, so as not to change the current behavior.

return EAI_NODATA, instead of EAI_NONAME, on name resolution errors.
EAI_NONAME does not make sense in these situations. from kame.
From: enami@netbsd.org

branches: 1.42.4;
remove unused struct def. more sanity check for NULL.

add $KAME$

disable AI_ADDRCONFIG (post RFC2553 specification),
it has bug in PF_UNSPEC caes. From: ume
fix boundary check which was broken by delint in the past.

add __RCSID().

supply weak alias for IPv6-related library additions. sorry to be late.

in yphostent, initialize res0 when getaddrinfo fails.

Delint:
- add notreached comment
- remove unused variables
- cast to void * first to avoid casting warnings
- fix bug in _yphostent where cur and sentinel were not
being initialized.

visit all requests even if we get errors from name server.
this will solve name query failures for IPv4-only node, when
the DNS server returns SERVFAIL or NXDOMAIN (should success with an == 0).
PR: 9621

avoid memory leak on valid FQDN + invalid port name.

query NIS hosts.byname only when necessary (= PF_UNSPEC of PF_INET queries).

to honor the search order defined by nsswitch.conf,
getaddrinfo(3) now calls lower-layer resolver functions directly.

XXX code duplicate with gethnamaddr.c/res_search.c - future code
maintenance issue? Since BIND[89] does not address this issue either at
this moment, switching to BIND[89] does not help.

to synchronize with progress of discussion for post
draft-ietf-ipngwg-scopedaddr-format-00.txt (will be 01),
use "address%interface" notation for extended scoped IPv6 address.

NOTE: the change affects link-local addresses only. I hope it to be
the final change on it.....

preparation to flip extended scoped address syntax.

the specification is still under debate; there is high possibility that
"fe80::1%de0" will be the finalist but i'm not sure yet.

make it possible to compile it without -DINET6
From: "Castor Fu" <castor@geocast.com>

sa_len was not properly set due to autoconf #ifdef.
(mistake in yesterday's upgrade)

revise extended scoped address format support. delimiter and the order
is changed, based on discussion in ipngwg scoped address cabal.
past code: fe80::1@de0
now: de0%fe80::1
this will be in sync with next extended address format proposal
(which should be final - I don't want to make this kind of change again).

backout previous commit - freeaddrinfo(NULL) issue.

allow freeaddrinfo(NULL). nothing is talked in RFC2553 (maybe in X/Open?).
however, since NRL freeaddrinfo permits it, we need to.

fix RFC2553 conformance. AI_CANONNAME does not mean reverse query.

fix comment.
better sync with kame tree.

remove #if 0'ed portion (AF filtering in normal name resolution)
add comment to the fragment which confused mycroft

Doh. I'm a dork.

Partially delint.

do not filter address families that are not supported by kernel.
the behavior was introduced in 1.8 -> 1.9, by itojun, and it was
not a correct behavior.

even if we may be able to connect to the the returned addresses,
filtering them out is not a correct thing to do. for example, even if
you have partial unreachablility in IPv4, gethostbyname(3) will return
IPv4 addresses in the unreachable network nonetheless.

remove #if 0'ed sections.

properly check the error result from if_nametoindex(), to disallow
bogus interface index specification like fe80::1@bogus.
(sync with kame)

more complete solution for gethostby*() buffer overwrite issue.
(we need thread-safe gethostby*...)

fix getaddrinfo(3) and getnameinfo(3).

getaddrinfo(3): result from gethostby* will be broken if we call it again.
deep-copy them. we have sa_len so configure them.
getnameinfo(3): we have sa_len so configure them.

bring in latest KAME get{addr,name}info(3).
getnameinfo(3) globs ai_socktype and ai_protocol correctly.
KAME scopeid extension is implemented.
(draft-ietf-ipngwg-scopedaddr-format-00.txt)

bump up shlib minor
(may not be necessary, but rather big difference in behavior - especially
ai_flags)

branches: 1.13.4;
back out the #ifdef _DIAGNOSTIC argument checks; too many people complained.
_DIAGASSERT() is still retained.

* use _DIAGASSERT() to check pointer arguments against NULL and file
descriptors against -1 (as appropriate).
* add actual checks which to detect stuff that would trigger_DIAGASSERT(),
and attempt to return a sane error condition.
* knf some code
* remove some `register' decls.

the first two items result in the addition of code similar to the
following in various functions:

_DIAGASSERT(path != NULL)
#ifdef _DIAGNOSTIC
if (path == NULL) {
errno = EFAULT;
return (-1);
}
#endif

Need "namespace.h".

pacify too-picky compiler.

From: onoe@sm.sony.co.jp

- do not make query for AFs that are not supported by kernel.
i.e. do not make query for IPv6 addresses, when running on non-IPv6 kernel,
or, do not query for IPv4 address on IPv6-only kernel :-)

This kind of behavior is not very well documented in RFC2553. This
may violate the spec.

- on EAI_AGAIN, only retry 3 times (3 is a magic number). Previous code
made retries forever. This solves situation where name server is wrongly
configured and nameserver:53 returns icmp port unreach.

The only proper fix for all getaddrinfo() related twists would be to
implement getipnodebyname() and get rid of wacky get_addr().
We need to contribute bind8 development for this.

fix name resolution problem when you have "hosts: files dns" on
/etc/nsswitch.conf.

This was because of two reasons:
- /etc/hosts lookup code damaged some of internal
state used by gethostbyname2().
- getaddrinfo() was not persistent enough against errors.

Sorry for the delay, and hope this fix all the following PRs
(I checked in my environment immitating those PRs and it worked for me)
PR: 7901, 7912, 7921

capitalize the first word of most error strings except those words that are
argument names to the function. e.g, if `const char *servname' is an argument
name, keep using "servname invalid [...]" instead of "Servname invalid [...]"

add NetBSD RCSID on top.

fix USE_GETIPNODEBY case.

fix behavior when there's dangling CNAME on specific AF.
TODO: have getipnodeby{name,addr} and let them do the trick.

improve lint friendliness on USE_GETIPNODEBY case.
(no change in behavior)

remove trailing `.' on error messages, since messages from strerror() and
h_strerror() don't return messages with a trailing `.' (and the `.' looks
ugly when using warnx() with the error message).

add IPv6 support functions.

get{addr,name} info are implemented to have as little impact to existing
resolver code as possible, so they are NOT the optimal implementation.
They are at this moment not very thread safe (as they call
gethostby{name,addr}).

(shlib minor version)++, as new interfaces are added.

TODO: getipnodeby{name,addr} - which needs total reimplementation of
gethostby{name,addr}.
upgrade rcmd.c for multiple af support (needed for IPv6-ready rsh/rlogin)

avoid various use-after-free issues.

create a ptrdiff_t offset between the start of an allocation region and
some interesting pointer, so it can be adjusted with this offset after
realloc() returns.

found by GCC 12.

lib: remove CONSTCOND comment

Since 2021-01-31, lint doesn't need it anymore for the common pattern of
'do ... while (0)'.

Don't complain about additional DNAME records received when resolving A or
AAAA (like already done for SIG/KEY and CNAME).

branches: 1.92.8; 1.92.18;
- fix various leaks on error
- don't use the wrong error variable in switch
- always set the error return code
- return consistent errors when the input data cannot be handled.

use fparseln() instead of abusing fgetln(), so that the last line that
might not have a terminating newline is processed.

branches: 1.90.2;
CID 1163170, 1164171, resource leak
CID 1161172 double free

fix memory allocation, and an off-by-one

the addition is handled in the alias.

Move more code into the macro, and fix the bug where realloc would cause
a dangling pointer and memory corruption.

Remove MAXALIASES limit

- don't clobber hp in the RES_USE_INET6 case
- increment naddrs in the yp case
- don't use __hostalias(), it is not thread-safe.

move the host file getbyname and getbyaddr functions to sethostent.c so
that they are in a separate file from the dns ones in order for libhack
to use them.

get rid of bogus + 1

fix gethostbyaddr prototype per TOG.

appease lint.

Add not advertised reentrant functions: {get,set,end}hostent_r,
gethostbyname{,2}_r, gethostbyaddr_r. Make getnameinfo(3) use
gethostbyaddr_r(3) so it is re-entrant (ahem __ypdomain). These
are not being advertised because there is a bunch of different
implementation of them that have a variety of type signatures.

If people want to follow someone's implementation, it is now easy.

branches: 1.79.2;
implement no-check-names

PR/45989: Martin Husemann: lint invocation does include -w only on i386

- turn lint -w for all the platforms after fixing the lint warnings.
- add _DIAGASSERTS() for casts that would assign values to types that
would not fit.
- change types, add casts
- change into ansii prototypes
- turn on _DIAGNOSTIC for libc (during current, to be eliminated for release
builds)

approved by core@

branches: 1.77.2; 1.77.4;
close on exec fixes:
- open + fcntl -> open O_CLOEXEC
- configuration database file descriptors that can stay open are now opened
fopen(db, "re")

dprintf is claimed by posix.

Revert a -> an changes: Adding an 'n' follows pronunciation, not writing.

fix grammar: a -> an

branches: 1.73.18;
Appropriate versions are not supplied by bind anymore.

merge in resolver portion of bind-9.4.0-rc2

branches: 1.71.2; 1.71.4;
fix incomplete initializers

Coverity CID 2279: Plug memory leak.

Coverity CID 2280: Plug memory leak.

PR/32616: Yves-Emmanuel JUTARD: Fix compilation with -DDEBUG

Rest of fix for PR#31184: getaddrinfo() now honors resolv.conf sortlist
directive for dns answers.

Also, unifdef the RESLVSORT (non-)option sillyness.

Reviewed by christos.

Re-enable resolv.conf sortlist support in gethostbyname().

Fixes first part of PR#31184. addrsort() was not exposed to the rest of
libc as it turns out not to be usable for getaddrinfo(), so a different
sorter will be implemented for that.

appease gcc -Wuninitialized

Fix gethostent(3) (which was reenabled in rev 1.63) so that it actually
works by allowing _gethtent() to treat host.h_addrtype==0 && host.h_length==0
as "any address family" when parsing.

Restore gethostent() as a supported interface. Yes, it maybe obsolete
but it is specified by IEEE Std 1003.1, 2004 Edition (POSIX) and
the X/Open standards (Issue 6 and XNS 5.2).

* revert change removing gethostent() from gethostbyname(3) man page
* delete kruft from gethnamaddr.c leaving only gethostent() as a
wrapper around _gethtent().
* revert recent changes to <netdb.h>
+ restore gethostent() prototype
+ restore freehostent() prototype; handle similar to non-shipped
getipnodby*() prototypes
+ use correct XOPEN_SOURCE version (520 not 500) for freeaddrinfo()
prototype; interface specified by XNS5.2 not XNS5

Reviewd by <christos> and <drochner>.

now that e.f.f.3.ip6.arpa is ready, we no longer need to query ip6.int

Change remaining K&R style function definition to ANSI; Reviewd by <christos>.

centralize res_*init() calls in __res_get_state()

Finish bind9 resolver merge.

branches: 1.58.2;
Back out rev. 1.56; `pointer points the byte just after the valid region'
is normal case. If the parser wants something more, there should be
another bounds check for it.

paren around macro arg

off-by-one. from millert@openbsd

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

Patches provided by Joel Baker in PR 22280, verified by myself.

bound copies

Fix signed/unsigned comparison warnings.

allocate 64K recieve buffer for DNS responses.

drop AAAA reply with IPv4 mapped address. sync w/kame
draft-itojun-v6ops-v4mapped-harmful-00.txt

set h_errno properly if internal string manipulation fails.

eliminate sprintf, use snprintf. check strlcat error.

query ip6.arpa then ip6.int for IPv6 reverse lookup. follows RFC3152.

correct DIAGASSERT

avoid remote buffer overrun on hostbuf[]. From: Joost Pol <joost@pine.nl>

correct bad practice in the code - it uses two changing variables
to manage buffer (buf and buflen). we eliminate buflen and use
fixed point (ep) as the ending pointer.

this fix is critical.

do not pass uninitialized variable to nsdispatch().

__STDC__ is always defined on NetBSD.

be picky about string manipulation in reverse lookups

branches: 1.42.2;
style

do not reverse-lookup scoped IPv6 address. it does not make sense as
there's no way to give scope identifier. sync w/kame

Change the `len' arguemtn to gethostbyaddr(3) from int to socklen_t in
accordance with XNS5.2; fixes PR standards/12703.

branches: 1.39.2;
remove redundant declarations.

Varargsify dprintf to allow for format string checking

on gethostby* with NIS backend, make a lookup against ipnodes.{byname,byaddr}
for non-IPv4 address. obeys solaris8 practice.

XXX does not support scoped address extension, as gethostby* are not
scope-aware. always use getaddrinfo/getnameinfo.
XXX it is not very useful at this moment, if you define multiple entries for
single hostname. see PR 10713 for detail.

plug host_addrs overrun. rogue NIS server can transmit tons of lines,
effectively overrun host_addrs buffer.

Passing "char" values to ctype(3) functions is problematic.
If an argument of a ctype function is outside "unsigned char"
and if it is not EOF, the behavior is undefined.

The isascii(3) is the sole exception of above and it was used to
be used to check a value was valid for other ctype functions in
ancient systems. On modern systems, the ctype functions take
all values of "unsigned char", and this check is obsolete and
even wrong for non-ASCII systems. However, we leave the isascii()
untouched for now, so as not to change the current behavior.

add CONSTCOND where appropriate

branches: 1.33.4;
Fix an alignment problem introduced by the patch for PR lib/8032.
Fixes PR lib/10048 by J.T. Conklin.

sync with BIND 4.9.7. strict boundary checking.

Try to silence some of the lint warnings. Unfortunately we cannot fix
them all right now until we fix the ALIGN() macro to return u_long in
all the ports.

Delint.

Delint.
Remove trailing ; from uses of __weak_alias(). The macro inserts this if
needed.

ignore KEY and SIG records when matching responses to the query. we
get those as extras from DNSsec-enabled name servers.

change default nsswitch.conf entry for `hosts' from:
hosts: dns files
to
hosts: files dns

this should speed things up a lot at boot if hostnames that are in
/etc/hosts are used for ifconfig, route, etc.

branches: 1.26.2;
back out the #ifdef _DIAGNOSTIC argument checks; too many people complained.
_DIAGASSERT() is still retained.

* use _DIAGASSERT() to check pointer arguments against NULL and file
descriptors against -1 (as appropriate).
* add actual checks which to detect stuff that would trigger_DIAGASSERT(),
and attempt to return a sane error condition.
* knf some code
* remove some `register' decls.

the first two items result in the addition of code similar to the
following in various functions:

_DIAGASSERT(path != NULL)
#ifdef _DIAGNOSTIC
if (path == NULL) {
errno = EFAULT;
return (-1);
}
#endif

do not make reverse query against IPv6 unspecified, or loopback addresses.

From: ume@mahoroba.org

- make sure that host_addr is suitably aligned
- add appropriate casts
- take out & in from array; it is meaningless

Make sure that "h_addr_list" pointers are longword aligned because some
broken applications depend on it. Fixes PR lib/8032 by David K Brownlee.

fix name resolution problem when you have "hosts: files dns" on
/etc/nsswitch.conf.

This was because of two reasons:
- /etc/hosts lookup code damaged some of internal
state used by gethostbyname2().
- getaddrinfo() was not persistent enough against errors.

Sorry for the delay, and hope this fix all the following PRs
(I checked in my environment immitating those PRs and it worked for me)
PR: 7901, 7912, 7921

check address family on getting entries from /etc/hosts.

fix hardcoded T_A.
this fixes gethostbyname2(host, AF_INET6) to actually work.

As discussed on "tech-net@netbsd.org":
Enhance the file resolver to return all IP addresses of a host on a
call to gethostbyname().

Remove unused 'len' and the relevant gcc kludges.

Don't bother de-linting inside __GNUC__ ifdefs.

specifically include stdarg/varargs.h

branches: 1.14.2;
minor knf; don't exceed 80 columns ;)

add ARGSUSED

constify ns_dtab defs

convert to new nsdispatch(3). lookup defaults back to "dns files"

before each call to nsdispatch(), set h_errno to NETDB_INTERNAL, and only
change to NETDB_SUCCESS if nsdispatch() returns NS_SUCCESS. it is up to the
internal methods to change h_errno from NETDB_INTERNAL to a more appropriate
value when an error is detected.

pull nsswitch up to main branch
convert getnetby*() to nsswitch (from resolv.conf lookup)

delint

branches: 1.7.2;
(bcmp(), bcopy(), bzero()) -> (memcmp(), memcpy(), memset())

Since these files are shared with parts of the BIND-4 userland (outside libc),
do the internal symbol renaming dance only if inside the libc build environment.

Need internal names for gethostby{addr,name}().

$Id: -> Id: (to make hunting down other bad Ids easier).

Bring over changes where were mistakenly committed to gethostnamadr.c:

date: 1997/07/21 14:07:54; author: jtc; state: Exp; lines: +3 -2
If port provides __weak_alias(), provide an Standard C and POSIX pure
identifier namespace by renaming non standard functions and variables
such that they have a leading underscore. The library will use those
names internally. Weak aliases are used to provide the original names
to the API.

This is only the first part of this change. It is most of the functions
which are implemented in C for all NetBSD ports. Subsequent changes are
to add the same support to the remaining C files, to assembly files, and
to the automagically generated assembly source used for system calls.
When all of the above is done, ports with weak alias support should add
a definition for __weak_alias to <sys/cdefs.h>.

merge bind 4.9.6 resolver. still need to fix <resolv.h>

branches: 1.1.1;
Initial revision

Sections 2 and 3 have RETURN VALUES, not DIAGNOSTICS

fix gethostbyaddr prototype per TOG.

branches: 1.28.42; 1.28.48;
Clarify that gethostbyname and gethostbyname2 can deal with
string representations of numeric addresses. PR 28667

Add commas in enumerations; drop trailing whitespace; bump date for previous.

Restore gethostent() as a supported interface. Yes, it maybe obsolete
but it is specified by IEEE Std 1003.1, 2004 Edition (POSIX) and
the X/Open standards (Issue 6 and XNS 5.2).

* revert change removing gethostent() from gethostbyname(3) man page
* delete kruft from gethnamaddr.c leaving only gethostent() as a
wrapper around _gethtent().
* revert recent changes to <netdb.h>
+ restore gethostent() prototype
+ restore freehostent() prototype; handle similar to non-shipped
getipnodby*() prototypes
+ use correct XOPEN_SOURCE version (520 not 500) for freeaddrinfo()
prototype; interface specified by XNS5.2 not XNS5

Reviewd by <christos> and <drochner>.

Sort sections.

branches: 1.24.2;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22280, verified by myself.

Use
.In header.h
instead of
.Fd #include \*[Lt]header.h\*[Gt]
Much easier to read and write, and supported by groff for ages.
Okayed by ross.

function with n.

Generate <>& symbolically.

Sort sections, sort SEE ALSO, and misc. fixes.

Change the `len' arguemtn to gethostbyaddr(3) from int to socklen_t in
accordance with XNS5.2; fixes PR standards/12703.

branches: 1.18.2;
it seems that gethostent(3) is no longer supplied in libc.
couple of wording corrections.
From: Hubert Feyrer <hubert.feyrer@informatik.fh-regensburg.de>

add missing period.

document gethostbyname2().

Last of the .Os cleanups. .Os is defined in the tmac.doc-common file,
so we shouldn't override it with versions in the manpages. Wheee!

highlight nsswitch.conf(5) in the first paragraph

reference nsswitch.conf(5)

fix mdoc references

add LIBRARY section to man page

merge in lite2 rcsids and a couple of man page fixes we had not yet found.

branches: 1.9.2;
alternate -> alternative, per PR 2643

add 'const' qualifiers to various routines (as per the include prototypes
and the implementation)

Added explicit return types to all functions in SYNOPSIS section.
Added documentation to hstrerror() function.
Added hstrerror MLINK.

clean up import. also convert everything to new Id format.

new resolver. changes from 4.9.2 by <paul@vix.com>, as well as a number
of features and optimizations by myself.

h_errno is an int, not a "struct".

Fix spelling errors.

Add RCS indentifiers.

branches: 1.1.1;
Initial revision

This file was deleted with the last resolver merge, and mistakenly
brought back to life with the __weak_alias() commit.

If port provides __weak_alias(), provide an Standard C and POSIX pure
identifier namespace by renaming non standard functions and variables
such that they have a leading underscore. The library will use those
names internally. Weak aliases are used to provide the original names
to the API.

This is only the first part of this change. It is most of the functions
which are implemented in C for all NetBSD ports. Subsequent changes are
to add the same support to the remaining C files, to assembly files, and
to the automagically generated assembly source used for system calls.
When all of the above is done, ports with weak alias support should add
a definition for __weak_alias to <sys/cdefs.h>.

merge bind 4.9.6 resolver. still need to fix <resolv.h>

Fix RCSID's
Fix gcc warnings
Add missing prototypes
Use "namespace.h"

branches: 1.23.2;
handle hex hostnames (from vixie, as reported in PR#3535).

oops; call getanswer correctly.

re-add YP parts.

merge in the bind 4.9.5-P1 resolver. we are now _mostly_ the
same as this, with our changes also included (resolv.conf and
other minor differences).

Prevent buffer overflow (and thus return of bogus data) when a DNS record
has more than MAXADDRS "A" records for a host.

Patch from Paul A. Vixie <paul@vix.com>, slightly modified to fit in with
NetBSD debug framework.

- convert unsafe strcpy(), strcat() and sprintf() to the `n' versions.
- some KNF.

_gethtbyname's arg should be const char *, not char *

branches: 1.16.2;
fix another bogon in last change; some changes derived from openbsd.

fix bogon in last change.

always update length; from vixie.

branches: 1.13.4; 1.13.6;
Use inet_aton(), not inet_addr().

don't forget to take address before deref'ing thru. from fvdl.

clean up import. also convert everything to new Id format.

another type change

one more u_long -> u_int32_t change

u_int_{16,32}_t stuff

set h_errno for non-BIND failed lookups too

new resolver. changes from 4.9.2 by <paul@vix.com>, as well as a number
of features and optimizations by myself.

fix bug where yp gethostby* lookups could have extra stuff in the alias
fields, and leave bogon stuff in the address fields...

support for lookup keyword in resolv.conf

Declare rcsid strings so they are stored in text segment.

Add RCS identifiers.

branches: 1.1.1;
Initial revision

Add freeifaddrs to the NAME section

Remove workaround for ancient HTML generation code.

Add ifa_addrflags to ifaddrs.

branches: 1.16.2;
unsigned -> unsigned int

Update synopses to match u_int/u_long/etc. changes. Do not bump date as
it's not a substantive change.

Remove trailing whitespace.

Add a small example.

branches: 1.12.6; 1.12.12;
Use .In for header files instead of .Ar Pa and variations.

Mark up NULL. New sentence, new line. Bump date for previous.

For non-AF_LINK ifaddrs, ifa_data is NULL. AFAICT, this has always been
so. Say so in the documentation.

Fix preamble

branches: 1.8.38; 1.8.42; 1.8.46;
Use
.In header.h
instead of
.Fd #include \*[Lt]header.h\*[Gt]
Much easier to read and write, and supported by groff for ages.
Okayed by ross.

Generate <>& symbolically.

Sort sections, sort SEE ALSO, and misc. fixes.

Boring whitespace fixes.

Fix date.

branches: 1.3.2;
Use .{B,Bs,F,N,O}x and fix miscellaneous errors.

add freeifaddrs(), which reclaims region allocated by getifaddrs().
in sync with kame and bsdi.

add getifaddrs(3) from bsdi4, as mentioned in tech-net (in SIOCGIFCONF
related thread).
this will save future apps from complexity of SIOCGIFCONF.

getifaddrs(3) does not use SIOCGIFCONF internally, it uses sysctl.

Add ifa_addrflags to ifaddrs.

branches: 1.15.14;
PR/45989: Martin Husemann: lint invocation does include -w only on i386

- turn lint -w for all the platforms after fixing the lint warnings.
- add _DIAGASSERTS() for casts that would assign values to types that
would not fit.
- change types, add casts
- change into ansii prototypes
- turn on _DIAGNOSTIC for libc (during current, to be eliminated for release
builds)

approved by core@

branches: 1.14.4;
Use RT_ROUNDUP instead of a private copy since that's what the kernel uses.
(t_getaddrinfo passes for both i386 (compat_netbsd32) and amd64 with this
change).

branches: 1.13.2;
Need getifaddrs() for the rump client since it executes sysctl(2)

Bring getifaddrs(3) behavior in line with the documentation: the
ifa_data member of every AF_LINK struct ifaddrs points at the
corresponding struct if_data. In ifconfig(8), do not try to suppress
duplicate AF_LINK ifaddrs by checking for a NULL ifa_data.

Don't copy out two AF_LINK struct ifaddrs for each active link-layer
address. getifaddrs(3) used to copy out one ifaddrs for the kernel's
RTM_IFINFO message, and one more for the kernel's RTM_NEWADDR message.
I suppress the first duplicate with a highly conservative change that
wastes a little bit of ifaddrs storage. The storage is not leaked.

branches: 1.11.12; 1.11.16; 1.11.20;
Use __arraycount().

branches: 1.10.22;
lint clean

unifdef for auditability.

cope with sa_len < sizeof(struct sockaddr).
From: Patrik Lindergren <patrik@datacom.nu>
sync with kame

branches: 1.7.2;
sprinkle _DIAGASSERT() around

define HAVE_IFM_ADDRS for non-bsdi. sync with kame

- avoid non portable casts
- cast arguments to avoid promotion warnings

add __RCSID().

supply weak alias for IPv6-related library additions. sorry to be late.

add freeifaddrs(), which reclaims region allocated by getifaddrs().
in sync with kame and bsdi.

add getifaddrs(3) from bsdi4, as mentioned in tech-net (in SIOCGIFCONF
related thread).
this will save future apps from complexity of SIOCGIFCONF.

getifaddrs(3) does not use SIOCGIFCONF internally, it uses sysctl.

libc: new function parameter, new line

branches: 1.44.2;
getnameinfo(3): fix/prettify markup

Document the NI_NUMERICSCOPE flag.

PR lib/57832

branches: 1.42.8; 1.42.16;
Remove workaround for ancient HTML generation code.

branches: 1.41.18;
Update comment.

getnameinfo is now thread-safe so remove bugs section.

add and document AF_LOCAL (and the rest that were not)