Further reduce the no-op build times (by 35% compared the original)

Revert "Revert "Reduce the no-op build times by 30%""

This reverts commit cbb30d993e3aab6a294f90d4f64714d988423363.

Revert "Reduce the no-op build times by 35%"

This reverts commit 81580e6bc7d17d1b31fca873924c49117ed7db80.

Reduce the no-op build times by 35%

Revert "Reduce the no-op build times by 30%"

This reverts commit 1c5ef289a872cf0dc78b2e9dd35b20fa99db86fb.

TODO: script: add test-loop.sh: loop over make test TESTS=<pat> until failure occurs

call:

TESTS="pattern" ./test-loop.sh

TODO: test_delayed_write_update3

s4:torture:basic: use torture_assert macros in some helper macros

No change in behaviour.

Signed-off-by: Michael Adam <obnox@samba.org>

s4:torture:basic: use torture_assert macros in finfo_after_write test

Signed-off-by: Michael Adam <obnox@samba.org>

TODO(SQ?) s4:torture:basic: use assert in delayed_write_update2

Signed-off-by: Michael Adam <obnox@samba.org>

s4:torture:basic: use torture_assert macros in delayed_write_update2

This does not introduce a behaviour change except for
early returns.
There is no point in continuing when something already
failed.

Signed-off-by: Michael Adam <obnox@samba.org>

s4:torture:basic: whitespace cleanup in delayed_write_update2

No diff seen with git diff -w

Signed-off-by: Michael Adam <obnox@samba.org>

TODO: also run the delaywrite against the s3dc environment again

Signed-off-by: Michael Adam <obnox@samba.org>

TMP: start proper timers for torture tests...

s4:torture:base: be even more generous in the defer_open test

After 1600 consecutive runs, the test failed for me again on
a busy server with elapsed time 1.82s. setting the absolute
generosity to 2 gives us room up to 2.15. That should hopefully
suffice.

Signed-off-by: Michael Adam <obnox@samba.org>

windbindd: Make cm_connect_lsa_tcp static

It is only used in winbindd_cm.c

Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Sat Oct  4 02:34:49 CEST 2014 on sn-devel-104

smbd: Always use uint8_t for inbuf pointer

This avoids some casts from char pointers.

Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Pass only cmd to construct_reply_common instead of complete req

construct_reply_common only needs the commands code and the inbuf field,
not the complete request.

Signed-off-by: Christof Schmitt <cs@samba.org>

smbd: Use MIN macro in fake_sendfile

Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

s3-winbindd: Make wcache_sid_to_name static

It is only used in winbindd_cache.c

Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

s3-winbindd: Remove extern declaration for cache_methods from winbindd_dual.c

cache_methods is not used in winbindd_dual.c

Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

lib/util: Use charset_compat.h if SAMBA_UTIL_CORE_ONLY

When doing a CTDB standalone build we don't want to use dynconfig,
since this introduces a lot of unwanted complexity.  To avoid this,
either:

* charset needs to be nobbled to avoid loading the case tables, since
  this depends on dynconfig; or

* charset needs to be avoid completely, so some functions need to be
  replaced with their ASCII counterparts.

The 2nd options seems more honest and less error-prone.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Jeremy Allison <jra@samba.org>

lib/util: Factor out subsystem samba-util-core from samba-util

samba-util depends on Samba-specific code.  Exclude this code from
samba-util-core.  When told to, via SAMBA_UTIL_CORE_ONLY, only build
samba-util-core and dependencies.  When SAMBA_UTIL_CORE_ONLY is not
defined then the behaviour should be unchanged.

Standalone builds of CTDB will be done against samba-util-core to
avoid pulling in Samba-specific code.

An alternative would be to remove the Samba-dependent code from
samba-util.  However, some of it is used by OpenChange.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Jeremy Allison <jra@samba.org>

lib/util: Clean up includes for util.c

Allows standalone compiles without external includes.h.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Jeremy Allison <jra@samba.org>

lib/util: Clean up includes for fault.c

Add fault.h.  Allows standalone compiles without external includes.h.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Jeremy Allison <jra@samba.org>

lib/util: Clean up includes for substitute.c

Add substitute.h.  Allows standalone compiles without external
includes.h.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Jeremy Allison <jra@samba.org>

lib/util: Replace an SMB_ASSERT()

Avoid a cyclic dependency.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Jeremy Allison <jra@samba.org>

lib/util: Clean up includes for signal.c

Add signal.h.  Allows standalone compiles without external includes.h.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Jeremy Allison <jra@samba.org>

lib/util: Clean up includes for time.[ch]

Allows standalone compile without external includes.h.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Jeremy Allison <jra@samba.org>

lib/util: Clean up includes for data_blob.[ch]

Allows standalone compile without external includes.h.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Jeremy Allison <jra@samba.org>

lib/util: Clean up includes for xfile.[ch]

The include files added to xfile.h are already included
unconditionally elsewhere (replace.h, samba_util.h, ...) so Samba
can't be built without them.

To minimise dependencies, only include specific util headers instead
of samba_util.h.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Jeremy Allison <jra@samba.org>

lib/util: Clean up includes for blocking.c

Add blocking.h.  Allows standalone compile without external
includes.h.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Jeremy Allison <jra@samba.org>

profiling: Make WITH_PROFILE span more in smbprofile.h

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Fri Oct  3 22:17:46 CEST 2014 on sn-devel-104

profiling: Remove a big DEBUG statement

I would like to have the freedom to play with the profiling implementation.
This is kindof in the way.

This code is from pre-SVN days:

> commit 7914e9351abb5271ebb4990c3b1fe495d15a4eda
> Author:     Jeremy Allison <jra@samba.org>
> AuthorDate: Thu Oct 5 18:50:18 2000 +0000
> Commit:     Jeremy Allison <jra@samba.org>
> CommitDate: Thu Oct 5 18:50:18 2000 +0000
>
>     Herb's fixes for profiling & compiler warnings.
>     Jeremy.

Herb, please speak up quickly if you still need this :-)

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

profiling: Remove some #ifdefs

The DO_PROFILE_INC thingies already #define to nothing without
WITH_PROFILE, and any sane compiler will just not compile the if-condition
if there is no body to be executed.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

profiling: Only compile profile/profile.c if profiling is enabled

This conditional compile avoids some #ifdef WITH_PROFILE, which makes the code
more readable

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

profiling: Only compile utils/status_profile.c if profiling is enabled

This conditional compile avoids some #ifdef WITH_PROFILE, which makes the code
more readable

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

profiling: Fix a typo

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

profiling: Move some #defines to profile.c

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

profiling: Fix a typo

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

profiling: Make "struct profile_header" static

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

winbind3: Fix pwent variable substitution

Commit 0ce46318 (winbind3: Simplify fillup_pw_field) broke variable
substitution by copying from the wrong (unsubstituted) buffer.  Fix it.

Signed-off-by: Justin Maggard <jmaggard10@gmail.com>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Richard Sharpe <rsharpe@samba.org>

Reduce the no-op build times by 30%

Change-Id: Ie68436c1e7c75c1786e9ed6b6a54d2b55abbbcea
Reviewed-by: Matthieu Patou <mat@matws.net>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User(master): Michael Adam <obnox@samba.org>
Autobuild-Date(master): Fri Oct  3 13:25:18 CEST 2014 on sn-devel-104

waf: add k5crypto to the list of system libraries if they are reported by the krb5config but not found as requied by the different tests

Otherwise on some systems we have error message like
"Implied dependency k5crypto in winbind_krb5_locator is of type EMPTY"
Even though krb5config indicated that this library is needed and the
library is present

Signed-off-by: Matthieu Patou <mat@matws.net>
Reviewed-by: Michael Adam <obnox@samba.org>

librpc: simplify ndr_pull_dnsp_string

total_len is 1 and never varies.

Signed-off-by: Matthieu Patou <mat@matws.net>
Reviewed-by: Michael Adam <obnox@samba.org>

Fix typo in samba-tool manpage

Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Fri Oct  3 11:00:58 CEST 2014 on sn-devel-104

s4:torture:basic: eliminate potential for false failures in delayed_write_update1c

We want to test that the write did update the write time immediately.
We check this by getting the file info in a loop for a few seconds.
There are several result cases:

- the server updated the write time immediately - success
- the server updated the write time, but not immediately - failure
- the server did not update the write time - failure

The loop is only there to be able to discern between the two
failure cases. The check for success is whether the first
getinfo has reportet the updated write time.

The potential for false failures was the additional timing check.
So if the first fileinfo call just took too long (e.g. due to a
busy system), this was reported as failure.

This patch should eliminate interemittent autobuild failures.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: use assert in the second loop in delayed_write_update1b

We can hence replace the assert after the loop by a success torture_comment.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: add check for size after initial write in delayed_write_update1c

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: add update into past as error condition in delayed_write_update1c

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: make use of torture_assert macros in delayed_write_update1c

This does not change the logic except for adding early
returns in failure cases.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: whitespace cleanup in delayed_write_update1c

no diff with git diff -w

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: eliminate potential for false failures in delayed_write_update1b

We want to test that the write did update the write time immediately.
We check this by getting the file info in a loop for a few seconds.
There are several result cases:

- the server updated the write time immediately - success
- the server updated the write time, but not immediately - failure
- the server did not update the write time - failure

The loop is only there to be able to discern between the two
failure cases. The check for success is whether the first
getinfo has reportet the updated write time.

The potential for false failures was the additional timing check.
So if the first fileinfo call just took too long (e.g. due to a
busy system), this was reported as failure.

This patch should eliminate interemittent autobuild failures.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: use assert in the second loop in delayed_write_update1b

We can hence replace the assert after the loop by a success torture_comment.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: add check for size after initial write in delayed_write_update1b

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: add update into past as error condition in delayed_write_update1b

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: make use of torture_assert macros in delayed_write_update1b

This does not change the logic except for adding early
returns in failure cases.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: whitespace cleanup in delayed_write_update1b

no diff with git diff -w

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: eliminate potential for false failures in delayed_write_update1a

We want to test that the write did update the write time immediately.
We check this by getting the file info in a loop for a few seconds.
There are several result cases:

- the server updated the write time immediately - success
- the server updated the write time, but not immediately - failure
- the server did not update the write time - failure

The loop is only there to be able to discern between the two
failure cases. The check for success is whether the first
getinfo has reportet the updated write time.

The potential for false failures was the additional timing check.
So if the first fileinfo call just took too long (e.g. due to a
busy system), this was reported as failure.

This patch should eliminate interemittent autobuild failures.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: use assert in the second loop in delayed_write_update1a

We can hence replace the assert after the loop by a success torture_comment.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: add check for size after initial write in delayed_write_update1a

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: add update into past as error condition in delayed_write_update1a

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: make use of torture_assert macros in delayed_write_update1a

This does not change the logic except for adding early
returns in failure cases.
But it makes the code more compact and obvious.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: fix spacing in the delayed_write_update1a test.

No diff visible with git diff -w

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: eliminate potential for false failures in delayed_write_update1

We want to test that the write did update the write time immediately.
We check this by getting the file info in a loop for a few seconds.
There are several result cases:

- the server updated the write time immediately - success
- the server updated the write time, but not immediately - failure
- the server did not update the write time - failure

The loop is only there to be able to discern between the two
failure cases. The check for success is whether the first
getinfo has reportet the updated write time.

The potential for false failures was the additional timing check.
So if the first fileinfo call just took too long (e.g. due to a
busy system), this was reported as failure.

This patch should eliminate interemittent autobuild failures.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: use assert in the second loop in delayed_write_update1

We can hence replace the assert after the loop by a success torture_comment.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: add check for size after initial write to delayed_write_update1

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: add update into past as error condition in delayed_write_update1

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: make use of torture_assert macros in delayed_write_update1

No logic change except for adding early returns.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: whitespace fix in delayed_write_update1

No diff with git diff -w

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: remove two unused variables from delayed_write_update3c

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: use torture_assert macros in delayed_write_update

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture:basic: whitespace cleanup in delayed_write_update

git diff -w shows no diff

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

torture: add torture_assert_u64_not_equal()

Signed-off-by: Michael Adam <obnox@samba.org>

torture: add torture_assert_int_not_equal

Signed-off-by: Michael Adam <obnox@samba.org>

idl: declare frstrans_VersionVector public so that we can dump it to blob and vice versa

Change-Id: Ica8dbe899eb257603ce6f7b2ea224c6533bdea46
Reviewed-by: Matthieu Patou <mat@matws.net>
Signed-off-by: Matthieu Patou <mat@matws.net>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User(master): Matthieu Patou <mat@samba.org>
Autobuild-Date(master): Thu Oct  2 14:27:02 CEST 2014 on sn-devel-104

idl: frstrans.idl add frstrans_RdcClose

Change-Id: I32e88f542da7df21e253cfbf8c2b3cc17043c6c7

Signed-off-by: Matthieu Patou <mat@matws.net>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

idl: improve frsrpc by adding missing flag, fixing typo (W -> VV) and making needed structure public

Change-Id: Ia507b5f627464ab78a145e54b276fc8d64ee3c14
Signed-off-by: Matthieu Patou <mat@matws.net>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

idl: Add frs blobs for dumping of frs related blobs with ndrdump

Change-Id: I5f8375fee663e452c26f264d73f637beffb1f209
Signed-off-by: Matthieu Patou <mat@matws.net>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

idl: Add compression_guid

Change-Id: I3cfe3c8f4a405fcf9a54092b03e87f00e7abd2fb
Signed-off-by: Matthieu Patou <mat@matws.net>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

idl: frsrpc depends on bkupblobs.idl and fscc.idl

Change-Id: I6ce1f4b28752765e7bc59c9eb7474ebd68d9b399
Signed-off-by: Matthieu Patou <mat@matws.net>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

idl: add fscc needed by frsrpc

Change-Id: Ib44a57b38a194e727242267a4821c88a15219eab
Signed-off-by: Matthieu Patou <mat@matws.net>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

idl: add nt backup blobs format

Change-Id: I4a43e1fe5fcebf2a4647ea2fbc9985663c0d7af2
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Matthieu Patou <mat@matws.net>

idl: frsrpc we don't need NDR_REMAINING as size of the blob is one the wire

Change-Id: Ia3f1a16668e97ae7b45a8d9ee4d0539c2652eb82
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Matthieu Patou <mat@matws.net>

idl: frsrpc cleaning, opnum 4->10 didn't exists

Change-Id: I6b45ffdead170e9f83deb8d16082cc9bb0366d47
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Matthieu Patou <mat@matws.net>

lib: Use GUID_buf_string in GUID_string

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Matthieu Patou <mat@matws.net>

Create talloc-less functions for formating GUID

Signed-off-by: Matthieu Patou <mat@matws.net>
Reviewed-by: Volker Lendecke <vl@samba.org>
Change-Id: I84adc45d2476dd225315bfc7dbb313f1b36b31e7

swrap: Bump version to 1.1.2.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Matthieu Patou <mat@matws.net>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Thu Oct  2 12:00:14 CEST 2014 on sn-devel-104

swrap: Add support for eventfd with unsigned count variable.

The prototype in glibc 2.20.90 changed.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

swrap: Add a trace message for swrap_socket().

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

swrap: Implement fcntl() to catch F_DUPFD.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

swrap: Include the function name in the debug output.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

swrap: Silence alignment warnings.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

swrap: Fix type punning warnings when loading functions.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

swrap: Fix access to struct members in log messages.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

swrap: Fix whitespace errors.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

swrap: Update copyright notice.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

swrap: Wrap fopen to detect stale file descriptors.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

swrap: Use swrap_address in swrap_accept().

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

swrap: Remove unused sockaddr_dup() function.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>