delete unneccessary arpa/nameser.h includes
ok millert

snprintf/vsnprintf return < 0 on error, rather than -1.

Remove NULL-checks before free().

Replace <sys/param.h> with <limits.h> and other less dirty headers where
possible. Annotate <sys/param.h> lines with their current reasons. Switch
to PATH_MAX, NGROUPS_MAX, HOST_NAME_MAX+1, LOGIN_NAME_MAX, etc. Change
MIN() and MAX() to local definitions of MINIMUM() and MAXIMUM() where
sensible to avoid pulling in the pollution. These are the files confirmed
through binary verification.
ok guenther, millert, doug (helped with the verification protocol)

In lookup_host(), return to caller a properly calculated message length
instead of using an uninitialized variable. While here, simplify the
code that calculates the message length, and use snprintf/strlcat more
carefully. ok millert@

opt_map CIRCLEQ -> TAILQ.

ok jsing@ (with nit), jca@

rcsid[] and sccsid[] and copyright[] are essentially unmaintained (and
unmaintainable). these days, people use source. these id's do not provide
any benefit, and do hurt the small install media
(the 33,000 line diff is essentially mechanical)
ok with the idea millert, ok dms

Remove clause 3 and 4. Ok with Chuck. -moj

use macros from sys/queue.h
ok otto@

spreading the lint love (very minimal things..)

Userland programs should include <errno.h> not <sys/errno.h>
OK deraadt@

Use queue macros instead of directly accessing fields. ok miod@ pat@ millert@

protos and repairs

-Wall cleanup: trim unused vars, better err() message in stdhosts,
and constify rcsid[]

deraadt@ ok

simplify a loop, convert it to strlcat; tedu@ ok

4344 lines of KNF diff

use a private svc_run() function that checks for a signal handler flag;
now signal safe. Also, massive KNF.

indent

Fix some problems with xfr of secure maps. Problem reported by
Eric Rungi <rungus@openface.org> -moj

Increased performance in ypserv by two different actions:

1) ypdb_open_db now check error condition when something goes wrong.
Instead of always as before. by request of theo.

2) turned on -DOPTDB in Makefile. This will cause ypserv to keep databases
opened. If not defined ypserv will open and close the map for every
request. I think it was turned off before when yppush didn't work.

If a map changes yppush will signal to ypserv that the map has changed
and ypserv will reopen the map.

ypserv will start closing maps when it run out of file descriptors. Then
it will close the map least resent used.

-moj

Umm, fix signedness error in my last commit, be explicit
about the fact that ordernumbers are 32 bit

careful about bad domain/mapnames

Now with support for "secure" maps a'la FreeBSD. Support exists in makedbm
and ypxfrd since dawn of time. Support is missing in YP-makefile and in
libc. I have the changes for YP-makefile to support master.passwd but since
the support is missing in libc I will not commit them yet. -moj

bail nice

plug db descriptor leak

Fix hostname lookups (improper use of strncpy).

for DNS hosts.byname queries, supply all addresses

Same copyright all over, removed my name from ypdb.[ch], small bugfix
in ypserv_db.c /moj

$OpenBSD$ tag and few spelling errors

force gethostby{name,addr} to use bind (really ugly & trivial hack)

major cleanup by Chuck Cranor
- use queue.h rather than do by hand
- fix up logging to use new yplog
- fix a few off by one errors
- don't double malloc the DBM structure
- avoid possible sprintf buffer overflows
- random code cleanup

+ full set of ypserv tools that compile

snprintf/vsnprintf return < 0 on error, rather than -1.

Remove NULL-checks before free().

Replace <sys/param.h> with <limits.h> and other less dirty headers where
possible. Annotate <sys/param.h> lines with their current reasons. Switch
to PATH_MAX, NGROUPS_MAX, HOST_NAME_MAX+1, LOGIN_NAME_MAX, etc. Change
MIN() and MAX() to local definitions of MINIMUM() and MAXIMUM() where
sensible to avoid pulling in the pollution. These are the files confirmed
through binary verification.
ok guenther, millert, doug (helped with the verification protocol)

In lookup_host(), return to caller a properly calculated message length
instead of using an uninitialized variable. While here, simplify the
code that calculates the message length, and use snprintf/strlcat more
carefully. ok millert@

opt_map CIRCLEQ -> TAILQ.

ok jsing@ (with nit), jca@

rcsid[] and sccsid[] and copyright[] are essentially unmaintained (and
unmaintainable). these days, people use source. these id's do not provide
any benefit, and do hurt the small install media
(the 33,000 line diff is essentially mechanical)
ok with the idea millert, ok dms

Remove clause 3 and 4. Ok with Chuck. -moj

use macros from sys/queue.h
ok otto@

spreading the lint love (very minimal things..)

Userland programs should include <errno.h> not <sys/errno.h>
OK deraadt@

Use queue macros instead of directly accessing fields. ok miod@ pat@ millert@

protos and repairs

-Wall cleanup: trim unused vars, better err() message in stdhosts,
and constify rcsid[]

deraadt@ ok

simplify a loop, convert it to strlcat; tedu@ ok

4344 lines of KNF diff

use a private svc_run() function that checks for a signal handler flag;
now signal safe. Also, massive KNF.

indent

Fix some problems with xfr of secure maps. Problem reported by
Eric Rungi <rungus@openface.org> -moj

Increased performance in ypserv by two different actions:

1) ypdb_open_db now check error condition when something goes wrong.
Instead of always as before. by request of theo.

2) turned on -DOPTDB in Makefile. This will cause ypserv to keep databases
opened. If not defined ypserv will open and close the map for every
request. I think it was turned off before when yppush didn't work.

If a map changes yppush will signal to ypserv that the map has changed
and ypserv will reopen the map.

ypserv will start closing maps when it run out of file descriptors. Then
it will close the map least resent used.

-moj

Umm, fix signedness error in my last commit, be explicit
about the fact that ordernumbers are 32 bit

careful about bad domain/mapnames

Now with support for "secure" maps a'la FreeBSD. Support exists in makedbm
and ypxfrd since dawn of time. Support is missing in YP-makefile and in
libc. I have the changes for YP-makefile to support master.passwd but since
the support is missing in libc I will not commit them yet. -moj

bail nice

plug db descriptor leak

Fix hostname lookups (improper use of strncpy).

for DNS hosts.byname queries, supply all addresses

Same copyright all over, removed my name from ypdb.[ch], small bugfix
in ypserv_db.c /moj

$OpenBSD$ tag and few spelling errors

force gethostby{name,addr} to use bind (really ugly & trivial hack)

major cleanup by Chuck Cranor
- use queue.h rather than do by hand
- fix up logging to use new yplog
- fix a few off by one errors
- don't double malloc the DBM structure
- avoid possible sprintf buffer overflows
- random code cleanup

+ full set of ypserv tools that compile

Remove NULL-checks before free().

Replace <sys/param.h> with <limits.h> and other less dirty headers where
possible. Annotate <sys/param.h> lines with their current reasons. Switch
to PATH_MAX, NGROUPS_MAX, HOST_NAME_MAX+1, LOGIN_NAME_MAX, etc. Change
MIN() and MAX() to local definitions of MINIMUM() and MAXIMUM() where
sensible to avoid pulling in the pollution. These are the files confirmed
through binary verification.
ok guenther, millert, doug (helped with the verification protocol)

In lookup_host(), return to caller a properly calculated message length
instead of using an uninitialized variable. While here, simplify the
code that calculates the message length, and use snprintf/strlcat more
carefully. ok millert@

opt_map CIRCLEQ -> TAILQ.

ok jsing@ (with nit), jca@

rcsid[] and sccsid[] and copyright[] are essentially unmaintained (and
unmaintainable). these days, people use source. these id's do not provide
any benefit, and do hurt the small install media
(the 33,000 line diff is essentially mechanical)
ok with the idea millert, ok dms

Remove clause 3 and 4. Ok with Chuck. -moj

use macros from sys/queue.h
ok otto@

spreading the lint love (very minimal things..)

Userland programs should include <errno.h> not <sys/errno.h>
OK deraadt@

Use queue macros instead of directly accessing fields. ok miod@ pat@ millert@

protos and repairs

-Wall cleanup: trim unused vars, better err() message in stdhosts,
and constify rcsid[]

deraadt@ ok

simplify a loop, convert it to strlcat; tedu@ ok

4344 lines of KNF diff

use a private svc_run() function that checks for a signal handler flag;
now signal safe. Also, massive KNF.

indent

Fix some problems with xfr of secure maps. Problem reported by
Eric Rungi <rungus@openface.org> -moj

Increased performance in ypserv by two different actions:

1) ypdb_open_db now check error condition when something goes wrong.
Instead of always as before. by request of theo.

2) turned on -DOPTDB in Makefile. This will cause ypserv to keep databases
opened. If not defined ypserv will open and close the map for every
request. I think it was turned off before when yppush didn't work.

If a map changes yppush will signal to ypserv that the map has changed
and ypserv will reopen the map.

ypserv will start closing maps when it run out of file descriptors. Then
it will close the map least resent used.

-moj

Umm, fix signedness error in my last commit, be explicit
about the fact that ordernumbers are 32 bit

careful about bad domain/mapnames

Now with support for "secure" maps a'la FreeBSD. Support exists in makedbm
and ypxfrd since dawn of time. Support is missing in YP-makefile and in
libc. I have the changes for YP-makefile to support master.passwd but since
the support is missing in libc I will not commit them yet. -moj

bail nice

plug db descriptor leak

Fix hostname lookups (improper use of strncpy).

for DNS hosts.byname queries, supply all addresses

Same copyright all over, removed my name from ypdb.[ch], small bugfix
in ypserv_db.c /moj

$OpenBSD$ tag and few spelling errors

force gethostby{name,addr} to use bind (really ugly & trivial hack)

major cleanup by Chuck Cranor
- use queue.h rather than do by hand
- fix up logging to use new yplog
- fix a few off by one errors
- don't double malloc the DBM structure
- avoid possible sprintf buffer overflows
- random code cleanup

+ full set of ypserv tools that compile