IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
TDB2 returns a negative error number on failure. This is compatible
if we always check for != 0 instead of == -1.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We change all the headers and wscript files to use tdb_compat; this
means we have one place to decide whether to use TDB1 or TDB2.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
TDB2's API is slightly different from TDB1. In particular, all functions
return 0 (TDB_SUCCESS) or a negative error number, rather than -1 or tdb_null
and storing the error in tdb_error() (though TDB2 does that as well).
The simplest fix is to replace all the different functions with a wrapper,
and that is done here.
Compatibility functions:
tdb_null: not used as an error return, so not defined by tdb2.
tdb_fetch_compat: TDB1-style data-returning tdb_fetch.
tdb_firstkey_compat: TDB1-style data-returning tdb_firstkey
tdb_nextkey_compat: TDB1-style data-returning tdb_nextkey, with
TDB2-style free of old key.
tdb_errorstr_compat: TDB1-style tdb_errorstr() which takes TDB instead of ecode.
TDB_CONTEXT: TDB1-style typedef for struct tdb_context.
tdb_open_compat: Simplified open routine which takes log function, sets
TDB_ALLOW_NESTING as Samba expects, and adds TDB_CLEAR_IF_FIRST support.
Things defined away in TDB2 wrappers:
tdb_traverse_read: TDB2's tdb_traverse only uses read-locks anyway.
tdb_reopen/tdb_reopen_all: TDB2 detects this error itself.
TDB_INCOMPATIBLE_HASH: TDB2 uses the Jenkins hash already.
TDB_VOLATILE: TDB2 shouldn't have freelist scaling issues.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The two error tables need to be combined, but for now seperate the names.
(As the common parts of the tree now use the _common function,
errmap_unix.c must be included in the s3 autoconf build).
Andrew Bartlett
Autobuild-User: Andrew Bartlett <abartlet@samba.org>
Autobuild-Date: Mon Jun 20 08:12:03 CEST 2011 on sn-devel-104
This can be used to tell if a talloc stackframe is currently
available. Callers can use this to decide if they will use
talloc_tos() or instead use an alternative strategy. This gives us a
way to safely have calls to talloc_tos() in common code that may end
up in external libraries, as long as all talloc_tos() calls in these
pieces of common code check first that a stackframe is available.
This strange parameter is apparently very rarely used, and it seems to
me that on modern networks, if clients don't have correct clocks and
DST offsets, that many other things (Kerberos) start to fail pretty
quickly, and time and DST tables tend to be internet delivered anyway.
Autobuild-User: Andrew Bartlett <abartlet@samba.org>
Autobuild-Date: Sat Jun 11 03:54:45 CEST 2011 on sn-devel-104
This helps ensure the string cannot be ambiguous, while also ensuring
that it remains simple in the non-cluster case.
The asymmetry of reading get_my_vnn() but writing based on
NONCLUSTER_VNN is acceptable because in the non-clustered case, they
are equal, and in the clustered case we will print the full string.
Andrew Bartlett
Andrew Bartlett complained that valgrind needs --partial-loads-ok=yes otherwise
the Jenkins hash makes it complain.
My benchmarking here revealed that at least with modern gcc (4.5) and CPU
(Intel i5 32 bit) there's no measurable performance penalty for the
"correct" code, so rip out the optimized one.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Autobuild-User: Rusty Russell <rusty@rustcorp.com.au>
Autobuild-Date: Wed Jun 8 11:05:47 CEST 2011 on sn-devel-104
Only waiting for writability doesn't get fd errors back with poll.
So always begin by selecting for readability, and if we get it then
see if bytes were available to read or it really is an error condition.
If bytes were available, remove the select on read as we know we
will retrieve the error when we've finished writing and start
reading the reply (or the write will timeout or fail).
Metze and Volker please check.
Autobuild-User: Jeremy Allison <jra@samba.org>
Autobuild-Date: Mon Jun 6 21:53:16 CEST 2011 on sn-devel-104
This brings these helpful utility functions in common, as they are not
based on either loadparm system.
(The 'modules dir' parameter from Samba4 will shortly be removed, so
there is no loss in functionality)
Andrew Bartlett
Don't trigger the write handler and remove the POLLOUT flag for this fd. Report errors on TEVENT_FD_READ requests only.
Metze please check !
Jeremy.
Autobuild-User: Jeremy Allison <jra@samba.org>
Autobuild-Date: Fri Jun 3 22:53:52 CEST 2011 on sn-devel-104
TDB2 can break this assumption.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Autobuild-User: Rusty Russell <rusty@rustcorp.com.au>
Autobuild-Date: Thu Jun 2 12:07:40 CEST 2011 on sn-devel-104
The usecs arguments are (of course) microseconds, not milliseconds.
This was added by Andreas Schneider in 6c1bcdc2 (tevent: Document the
tevent helper functions.).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Autobuild-User: Rusty Russell <rusty@rustcorp.com.au>
Autobuild-Date: Wed Jun 1 11:47:38 CEST 2011 on sn-devel-104
Several places want "microseconds from current time", and several were
simply handing "usecs" values which could be over a million.
Using a helper to do this is safer and more readable.
I didn't replace any obviously correct callers (ie. constants).
I also renamed wait_nsec in source3/lib/util_sock.c; it's actually
microseconds not nanoseconds (introduced with this code in Volker's
19b783cc Async wrapper for open_socket_out_send/recv).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Several places want "milliseconds from current time", and several were
simply doing "msec * 1000" which can (and does in one place) result in
a usec value over 1 a million.
Using a helper to do this is safer and more readable.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
as discussed on samba-techincal we currelty don't rely on it and we don't want
to flood this warning message during compile.
Autobuild-User: Björn Jacke <bj@sernet.de>
Autobuild-Date: Tue May 31 18:36:53 CEST 2011 on sn-devel-104
Patch from Samuel Thibault <sthibault@debian.org> to fix Debian Bug 610678
resp. BSO #7998. IOV_MAX and UIO_MAXIOV are not defined on GNU Hurd.
Autobuild-User: Björn Jacke <bj@sernet.de>
Autobuild-Date: Mon May 30 00:53:59 CEST 2011 on sn-devel-104
With the recent consolidation of code between s3 and s4, a number of new
dependencies have been implicitly introduced. For example, previous s3
code gained an implicit dependency on talloc after the charset related
consolidation (lib/util/charset/charset.h now includes talloc.h). When
building against the embedded version of talloc this isn't a problem
since the paths are automatically added to the search path, but when
building against the external libraries build failures will occur for
all components that don't directly or indirectly include talloc as
a dependency.
Since charset.h is included from util.h, which in turn is included from
includes.h, this means most of the codebase (s3 and s4) has such an
undeclared dependency.
Therefore, samba-util-common and samba-util have been added as
dependencies to the s3 and s4 code respectively, for all cases where
the source would otherwise fail to build. Additionally, a few other
dependencies are added in specific wscript_build files to address
similar dependency-related problems.
https://bugzilla.samba.org/show_bug.cgi?id=8128
Signed-off-by: Sean Finney <seanius@seanius.net>
Signed-off-by: Matthias Dieter Wallnöfer <mdw@samba.org>
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User: Matthias Dieter Wallnöfer <mdw@samba.org>
Autobuild-Date: Wed May 25 19:22:13 CEST 2011 on sn-devel-104
When building on IPv6-only, doing:
hints.ai_family = AF_INET;
getaddrinfo("0.0.0.0", NULL, &hints, &ppres)
fails as AF_INET is unavailable on an IPv6-only system. This
causes us to fallback to our replacement getaddrinfo code
which is IPv4-only.
As we're only trying to detect a specific AIX bug here,
broaden the tests to find that bug, and also test for
working getaddrinfo in an IPv6-only safe way.
Autobuild-User: Jeremy Allison <jra@samba.org>
Autobuild-Date: Thu May 19 02:21:54 CEST 2011 on sn-devel-104
*talloc_pool_objectcount(pool_tc) == 2 doesn't mean the one of the objects
is the pool itself! So we better check for == 1 and calculate the chunk count.
metze
The optimization of the object_count == 1 case should only happen
for when we're not destroying the pool itself. And it should only
happen if the pool itself is still valid.
If the pool isn't valid (it has TALLOC_FLAG_FREE),
object_count == 1 does not mean that the pool is the last object,
which can happen if you use talloc_steal/move() on memory
from the pool and then free the pool itself.
Thanks to Volker for noticing this!
metze
It is a private library, and OpenChange has their own which they use, so
it's not for them either.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Autobuild-User: Rusty Russell <rusty@rustcorp.com.au>
Autobuild-Date: Tue May 10 05:21:19 CEST 2011 on sn-devel-104
I'd like Samba to use the native OpenLDAP and MIT Kerberos libs.
Attached are some patches to do that. (relative to git master)
It does not build for me without these.
(OpenIndiana is an off-shoot of OpenSolaris See http://www.openindiana.org)
Autobuild-User: Jeremy Allison <jra@samba.org>
Autobuild-Date: Sat May 7 02:20:14 CEST 2011 on sn-devel-104
This makes it easier to compile this in the top level with s3 and s4
headers.
Andrew Bartlett
Autobuild-User: Andrew Bartlett <abartlet@samba.org>
Autobuild-Date: Fri May 6 08:50:52 CEST 2011 on sn-devel-104
gen_fnmatch was a duplicate symbol in the top level build.
gen_fnmatch() used for simple non-CIFS pattern matching, so selecting
the lib/util implementation should not be a concern.
Andrew Bartlett
This #if _SAMBA_BUILD == 3 is very unfortunate, as it means that in
the top level build, these options are not available for these
databases. However, having two different tdb_wrap lists is a worse
fate, so this will do for now.
Andrew Bartlett
This isn't a very good public header, but util.h includes it, so we
don't have much choice in the short term.
Andrew Bartlett
Autobuild-User: Andrew Bartlett <abartlet@samba.org>
Autobuild-Date: Tue May 3 08:37:22 CEST 2011 on sn-devel-104
strequal() is now implemented in terms of strcasecmp_m() which is
tested in smbtorture and which does not talloc() for ASCII or
non-ASCII comparions, and has an ASCII fast-path.
Andrew Bartlett
The only caller of push_string() (not to be confused with
push_string_check()) in the common code was encode_pw_buffer(), and it
didn't use the alignment or STR_UPPER flags.
The talloc_strupper() and talloc_strlower() functions are tested in
smbtorture, and are next_codepoint() based.
Andrew Bartlett
This function is problematic because a string may expand in size when
changed into upper or lower case. This will then push characters off
the end of the string in the s3 implementation, or panic in the former
s4 implementation.
Andrew Bartlett
If we optimize on top of raw realloc() we need
TC_INVALIDATE_SHRINK_CHUNK together with TC_UNDEFINE_GROW_CHUNK
(with was missing and caused false positive valgrind warnings).
But that is really slow, as we do a lot of talloc_realloc calls in samba.
That's why we only to TC_UNDEFINE_SHRINK_CHUNK() for now.
metze
This breaks the fast path into handling for -1 and handling for
specified lenghts, avoding branch operations on each character.
Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>
Signed-off-by: Andrew Tridgell <tridge@samba.org>
This punts partial UTF16 strings to iconv() to deal with, as it's not
a fast path any longer if it's got an odd length.
Andrew Bartlett
Signed-off-by: Andrew Tridgell <tridge@samba.org>
The caller may wish to inspect the partially converted string.
Andrew Bartlett
Autobuild-User: Andrew Bartlett <abartlet@samba.org>
Autobuild-Date: Wed Apr 20 05:17:48 CEST 2011 on sn-devel-104
This is now API compatible with the existing code in lib/util/charset
lazy_initialize_conv() is no longer called as init_iconv() is called
when the smb.conf is processed, and get_conv_handle() will auto-init
with defaults if required.
load_case_tables_library() is no longer requried as all binaries and
libraries already load these in their entry points, as otherwise all
the other string functions would fail.
Andrew Bartlett
Transactions have the side effect of generating bigger files.
By removing the transaction files get as much as 30% smaller.
Autobuild-User: Simo Sorce <idra@samba.org>
Autobuild-Date: Tue Apr 19 23:34:37 CEST 2011 on sn-devel-104
If it's really the recovery area, we can trust the rec_len field, and
don't have to go groping for bitpatterns.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Autobuild-User: Rusty Russell <rusty@rustcorp.com.au>
Autobuild-Date: Tue Apr 19 14:15:22 CEST 2011 on sn-devel-104
ldb can create huge records when saving indexes.
Limit the tdb expansion to avoid consuming a lot of memory for
no good reason if the record being saved is huge.
