Use -Wno-bitwise-op-parentheses only if the compiler has them.

Do not use bundled libobjc on Darwin.
Refactor cmake scritps: introduce os.cmake, arch.cmake,
compiler.cmake, libobjc.cmake, luajit.cmake.

Remove the hack with building libobjc/luajit as a
dependency on cfg.

Do not specify -ggdb when compiling with clang.

Add -fno-objc-legacy-dispatch and -fobjc-nonfragile-abi
when compiling with CLang on Linux.

Remove -fgnu-runtime when compiling with gcc on Mac.

Many minor cleanups.

libobjc2 has a notion of "small objects". The type
of a small object is detected by checking its address:
it should not be word-aligned.

When performing a custom memory allocation for Request,
make sure it's not a small object, by aligning it around
(void *) pointer (word) boundary.

object_dispose in libobjc2 runtime doesn't check whether
the passed in object is nil, so check for nil outside.

Add a makefile to libobjc.

Fix a typo in the code which would move the code
to the Objective C 2.0 runtime. Correctly check
for the size of area necessary to store the object (vs.
allocated area) to make a decision whether or not we
need to re-allocate.

In the new Objective C 2.0 runtime, class Object was trimmed
and no longer contains + alloc, - init, - free and - subclassResponsibility,
which we use.

Implement our own base class and use it instead of the old Object.

This should fix compilation with gcc 4.7 everywhere and clang on
Darwin.

The current code can still be built with gcc 4.6.

macosx2: merging in (manually) previously unmerged changes (from macosx-port1), non-working tests: box/lua.test, box_big/lua.test connector_c/xlog_rpl.test

A fix for Bug#1017905 "replication failure may leave garbage in the read
buffer".

No test case since this is a runtime error which is hard to emulate.

Clear the read buffer whenever closing a file desriptor, to remove
garbage data from the closed connection.

Tarantool/Box crashed during shutdown if a non running LUA fiber
was created.

Fiber garbage collection in Lua, initiated during shutdown,
was performed after the fiber subsystem was destroyed.

Additionally, fix a wrongly formulated assert which checks
that the fiber which is being cancelled has not been
cancelled already.

A fix for bug https://bugs.launchpad.net/tarantool/+bug/790688
"ER_TUPLE_IS_RO should be more rare".

Rewrite the transaction processor to optimistically
perform changes in memory. In case of an error, transactions
are rolled back in order, reverse to the order of execution,
and gradually restore the original state of memory.
The wal writer therad orchestrates the rollback, by not
initiating it until all queries which entered the processing queue
before the error has occurred were not added to the rollback queue.

While in the state of rollback (i.e. collecting all requests
for rollback) new requests are not accepted into the processing
queue (see changes in wal_write()).

XXX: This patch makes dirty reads possible when we are in rollback,
since SELECTs do not enter the wal writer queue and can therefore
see changes which later will be rolled back. Since this is a
very rare situation, it's considered acceptable to fix this problem
in a separate patch (a separate bug will be reported).

The field made optional by the previous refactoring
patch were used for monitoring.
Put them back.

Extract struct recovery_state members which are used
only when replication is on to struct remote.

No actual code changes, simply split a big file in two.

No code changes.

Do not use buffered file I/O which doesn't
give a precise account of failure, when it happens.

Instead, batch writes to WAL or SNAP files and
write batches using writev(). If writev() fails,
we can accurately evaluate which requests (those that
belong to the batch) were not written to disk.

In case of a partial write, rollback the failed requests,
and seek back to the end of the last fully written
request.

To support writev() and IO with file descriptors,
implement nio library, which provides a set of
EINTR- and error log- aware wrappers around syscalls.

Extract the WAL rotation into a separate function.

Merge common code in wal_write() and snapshot_write_row()
by introducing a structure to represent a single
snapshot or WAL row to be written.

No code changes.

iproto packages validation routine was added.

Continue code cleanup in log_io.m, remove more dead code.

Remove log_io_iter from snapshot IO. Do not use coroutines
and extra buffers. This opens the way to more agressive
buffering of snapshot IO, as compared to no buffering of WAL IO.

Cleanup work with eof/row markers. Remove unnecessary members
of log_io_class.

Fix a regression introduced by the yesterday's cleanup:

recover_wal() can return 0, -1, or 1:
0 - ok, eof read
-1 - error reading wal
1 - can't read eof marker from wal.

recover_remaining_wals() looks at this, and when doing final recovery
deletes wal which don't have eof marker (return value from recover_wal()
== 1).

Yesterday I mistakenly merged 0 and 1 return values into one.
Revert this part of the yesterday's patch.

On the same token, log_io_cursor_next() can't throw an exception,
ever. Take this into account and simplify code.

recover_remaining_wals() continues to stay a mess.

- remove unused parameters
- add SystemError exception (derived form net-split tree)
- improve error printouts.