[SCM] Samba Shared Repository - branch master updated

[idra at samba.org]

On Sun, May 08, 2011 at 08:04:03PM +0200, Andrew Bartlett wrote:
> The branch, master has been updated
>        via  5def7a1 s4-selftest Add tests for proxy_samba4_dc
>        via  f4d9b40 selftest: Make the ncalrpc dir common between Samba4 and Samba3.
>        via  aae9353 ncalrpc: Force ncalrpc dir to be mode 755 in all users
>        via  eae0612 selftest: Add plugin_s4_dc environment
>        via  f52c6ee s3-auth Add auth_samba4 module
>        via  a47c7f6 s3-passdb: added pdb_samba4
>        via  25cfa29 s3-passdb Redirect domain GUID and SID queries to the passdb stack
>        via  5d2ce40 s4-dsdb Add transactions to dsdb modify helpers
>        via  8882dab s4-samr Remove incorrect transaction_cancel() in error path
>        via  ac82ac4 s4-param cope with doulbe-parsing of -foo and +foo lists
>        via  d1df1cb build: Allow the C code to know if this is a waf build
>       from  c596d85 s4-interfaces: keep interfaces in the order they were declared
> 
> http://gitweb.samba.org/?p=samba.git;a=shortlog;h=master
> 
> 
> - Log -----------------------------------------------------------------
> commit 5def7a1e5ef98d215d1e618b4e8fd9c5839c6e1e
> Author: Andrew Bartlett <abartlet at samba.org>
> Date:   Sun May 8 15:26:28 2011 +0200
> 
>     s4-selftest Add tests for proxy_samba4_dc
>     
>     Autobuild-User: Andrew Bartlett <abartlet at samba.org>
>     Autobuild-Date: Sun May  8 20:03:45 CEST 2011 on sn-devel-104
> 
> commit f4d9b40d7ff8f9e79972294c9dce2a52a67df24f
> Author: Andrew Bartlett <abartlet at samba.org>
> Date:   Sun May 8 10:32:01 2011 +0200
> 
>     selftest: Make the ncalrpc dir common between Samba4 and Samba3.
>     
>     This also avoids creating the directory, as the startup routines will
>     create it with the correct permissions.
>     
>     Andrew Bartlett
> 
> commit aae9353ecf56323b63da66aa84d8a0a4f219d94d
> Author: Andrew Bartlett <abartlet at samba.org>
> Date:   Sun May 8 10:29:27 2011 +0200
> 
>     ncalrpc: Force ncalrpc dir to be mode 755 in all users
>     
>     This allows this directory to be shared between Samba3 and Samba4 in a
>     Franky-style setup easily.
>     
>     Andrew Bartlett

This one has interesting security implications.
The dir was 700 to control who had access to it.
Have you checked that the s3 code using ncalrpc is ok with widening access to this directory ?

 
> commit eae061293bad97e6e70526b4e755ca69c02d0ae9
> Author: Andrew Bartlett <abartlet at samba.org>
> Date:   Sun May 8 06:55:33 2011 +0200
> 
>     selftest: Add plugin_s4_dc environment
>     
>     This environment uses pdb_samba4 and auth_samba4 to plug these
>     critical subsystems into a mixed Samba3/Samba4 DC, in a similar way to
>     the 'Franky' proposal.
>     
>     Andrew Barltett
> 
> commit f52c6eeacc1131fca6656d07eca035668423b008
> Author: Andrew Bartlett <abartlet at samba.org>
> Date:   Sun May 8 14:48:11 2011 +0200
> 
>     s3-auth Add auth_samba4 module
>     
>     This module makes a direct call into the Samba4 auth stack to
>     authenticate Samba4 uses in a Samba3 file server.  The direct call
>     avoids the need to obtain schannel credentials.
>     
>     Andrew Bartlett
>     
>     Pair-Programmed-With: Andrew Tridgell <tridge at samba.org>
> 
> commit a47c7f6b4cfc8ce271cfd5addbf8f901d9ab3c6e
> Author: Andrew Bartlett <abartlet at samba.org>
> Date:   Sat May 7 00:50:04 2011 +1000
> 
>     s3-passdb: added pdb_samba4
>     
>     This uses direct LDB operations and calls to the dsdb library to allow
>     passdb operations (such as pdbedit and smbpasswd) offline, and uses
>     transactions internally for database consistency.
>     
>     Pair-Programmed-With: Andrew Tridgell <tridge at samba.org>
> 
> commit 25cfa29e29bdbb6c84bd85ea02ec542228ae585f
> Author: Andrew Bartlett <abartlet at samba.org>
> Date:   Sun May 8 12:52:06 2011 +0200
> 
>     s3-passdb Redirect domain GUID and SID queries to the passdb stack
>     
>     This is done if the passdb module supports PDB_ADS, and ensures that a
>     random SID is never made up locally for these directories.
>     
>     This is only enabled when in the waf build, due to dependency issues.
>     
>     Andrew Bartlett

The ifdefs here are quite ugly, any reason why you can't simply synchronize
the secrets contents from samba4 instead of adding those ? It shouldn't be
difficult to add the right GUID and SID during ptrovesion so this file is
consistent.

In any case the define name is bad, WAF is the build system not an indication
of functionality, we will soon get a mess of stuff that artificially depend
on the "WAF" build. Can you please create a specific define named against the
funcitonality you want to represent here ? Otherwise 2 years down the road
people will have to guess why this stuff depends on the buildsystem and will be
horribly confused.

> commit 5d2ce400d2324e58861b696a02582d2858d1b25b
> Author: Andrew Bartlett <abartlet at samba.org>
> Date:   Sun May 8 15:56:58 2011 +0200
> 
>     s4-dsdb Add transactions to dsdb modify helpers
> 
> commit 8882dab93e65bb20618ed1dbbd2279f1e3bdc9cb
> Author: Andrew Bartlett <abartlet at samba.org>
> Date:   Sun May 8 15:24:57 2011 +0200
> 
>     s4-samr Remove incorrect transaction_cancel() in error path
>     
>     The transactions are now handled entirely within dsdb_add_user()
>     
>     Andrew Bartlett
> 
> commit ac82ac4b83d3910734943bbc4caba5ef1efefe06
> Author: Andrew Bartlett <abartlet at samba.org>
> Date:   Sun May 8 06:58:19 2011 +0200
> 
>     s4-param cope with doulbe-parsing of -foo and +foo lists
>     
>     For some reason these lists are parsed twice, and so any -foo was
>     failing as it was already removed the first time.
>     
>     Andrew Bartlett
> 
> commit d1df1cb714289b90dc173a72e488773960952402
> Author: Andrew Bartlett <abartlet at samba.org>
> Date:   Sun May 8 17:35:05 2011 +0200
> 
>     build: Allow the C code to know if this is a waf build
>     
>     This allows addition of functionality that can't be handled (for
>     example, due to dependencies) in the autoconf build.
>     
>     Andrew Bartlett

As noted above, bad idea IMHO.

Simo.