Backport of the fix for BUG#33730 "Full table scan instead selected partitions for query more than 10 partitions"
from 6.0, made in sergefp@mysql.com-20090205190644-q8632sniogedhtsu

(from revision konstantin@mysql.com-20080627154042-923m6lzk7z77lrgj ). This moves the slow part
(10 seconds over 13) into a separate big test.

2. Mark ndb, rpl_ndb test suites experimental.

Revision: 
2597.72.1 revid:sp1r-Reggie@core.-20080403153947-15243
removed instances of __NT__ from code. We now only build "NT" binaries

http://lists.mysql.com/commits/59686

Cleanup pthread_self(), pthread_create(), pthread_join() implementation on Windows.
Prior implementation is was unnecessarily complicated and even differs in embedded
and non-embedded case.
      
Improvements in this patch:
* pthread_t is now the unique thread ID, instead of HANDLE returned by beginthread
      
This simplifies pthread_self() to be just straight GetCurrentThreadId().
prior it was much  art involved in passing the beginthread() handle from the caller
to the TLS structure in the child thread ( did not work for the main thread of
course)
      
* remove MySQL specific my_thread_init()/my_thread_end() from pthread_create.
No automagic is done on Unix on pthread_create(). Having the same on Windows will 
improve portability and avoid extra #ifdef's
      
* remove redefinition of getpid() - it was defined as GetCurrentThreadId()

2677 Vladislav Vaintroub	2008-11-04
CMakeLists.txt files cleanup
- remove SAFEMALLOC and SAFE_MUTEX definitions that were 
present in *each* CMakeLists.txt. Instead, put them into top level 
CMakeLists.txt, but disable on Windows, because
a) SAFEMALLOC does not add any functionality that is not already
present in Debug C runtime ( and 2 safe malloc one on top of the other 
only unnecessarily slows down the server) 
      
b)SAFE_MUTEX does not work on Windows  and have been
explicitely  disabled on Windows with #undef previously.  Fortunately,
ntdll does  pretty good  job identifying l problems with 
CRITICAL_SECTIONs.
DebugBreak()s on using uninited critical section, unlocking unowned 
critical section)
      
-Also, remove occationally used -D_DEBUG (added by compiler 
anyway)

We cann connect() in a non-blocking mode to be able to specify a
non-standard timeout.

The problem was that we did not fetch the status from the
non-blocking connect(). We assumed that poll() would not return
a POLLIN flag if the connect failed. But on some platforms this
is not true.

After a successful poll() we do now retrieve the status value
from connect() with getsockopt(...SO_ERROR...). Now we do know
if (and how) the connect failed.

The test case for my investigation was rpl.rlp_ssl1 on an
Ubuntu 9.04 x86_64 machine. Both, IPV4 and IPV6 were active.
'localhost' resolved first for IPV6 and then for IPV4. The
connection over IPV6 was blocked. rpl.rlp_ssl1 timed out
as it did not notice the failed connect(). The first read()
failed, which was interpreted as a master crash and the
connection was tried to reestablish with the same result
until the retry limit was reached.

With the fix, the connect() problem is immediately recognized,
and the connect() is retried on the second resolution for
'localhost', which is successful.

DECIMAL and TIMESTAMP used to have NUM_FLAG, but NEWDECIMAL was forgotten.
It's correct that TIMESTAMP does not have the flag nowadays (manual will be updated, connectors
developers will be notified).

This patch provides performance improvements:
- send_fields() when character_set_results = latin1
  is now about twice faster for column/table/database
  names, consisting on ASCII characters.

Changes:

- Protocol doesn't use "convert" temporary buffer anymore,
  and converts strings directly to "packet".

- General conversion optimization: quick conversion
  of ASCII strings was added.

modified files:

include/m_ctype.h
- Adding a new flag.
- Adding a new function prototype

libmysqld/lib_sql.cc
- Adding quick conversion method for embedded library:
  conversion is now done directly to result buffer,
  without using a temporary buffer.

mysys/charset.c
- Mark all dynamic ucs2 character sets as non-ASCII
- Mark some dymamic 7bit and 8bit charsets as non-ASCII
  (for example swe7 is not fully ASCII compatible).

sql/protocol.cc
- Adding quick method to convert a string directly
  into protocol buffer, without using a temporary buffer.

sql/protocol.h
- Adding a new method prototype

sql/sql_string.cc
  Optimization for conversion between two ASCII-compatible charsets:
- quickly convert ASCII strings,
  switch to mc_wc->wc_mb method only when a non-ASCII character is met.
- copy four ASCII characters at once on i386

strings/conf_to_src.c
- Marking non-ASCII character sets with a flag.

strings/ctype-extra.c
- Regenerating ctype-extra.c by running "conf_to_src".

strings/ctype-uca.c
- Marking UCS2 character set as non-ASCII.

strings/ctype-ucs2.c
- Marking UCS2 character set as non-ASCII.

strings/ctype.c
- A new function to detect if a 7bit or 8bit character set
  is ascii compatible.

http://lists.mysql.com/commits/57725

Vladislav Vaintroub	2008-11-03
Cleanup CMakeLists.txt(s) - remove winsock2 (ws2_32) from 
TARGET_LINK_LIBRARIES. 
      
Every exe or dll linked with mysys needs ws2_32, because
mysys uses winsock function WSAStartup in my_init().
However, there is no need to explicitely add ws2_32 to
the list of TARGET_LINK_LIBRARIES multiple times. 
Visual Studio comes with a handy pragma that tells linker
to add library. So patch replaces bunch of ws2_32 in 
CMakeLists with  single pragma comment(lib,"ws2_32")
in my_init.c
      
Additionally, reference to non-existing "debug" library
has been removed from TARGET_LINK_LIBRARIES. The correct
name of the library is "dbug".

htttp://lists.mysql.com/commits/50957?f=plain
      
Always use TLS functions instead of __declspec(thread) to access 
thread local storage variables.
The change removes the necessity to recomplile the same source
files twice -  with USE_TLS for DLLs and without USE_TLS for EXEs.
Real benefit of this change is better readability and maintainability
of TLS functions within MySQL.
              
There is a performance loss using TlsXXX functions compared to __declspec 
but the difference is negligible in practice. In a sysbench-like benchmark 
I ran with with TlsGetValue, pthread_[get|set]_specific was called 600000000 
times and took 0.17sec of total 35min CPU time, or 0.008%.