This is the first commit for the branch Lua 5.3. All source files
were copied from the official distribution of 5.3.5 in the Lua site.
The test files are the same of 5.3.4. The manual came from the
previous RCS repository, revision 1.167.1.2.

A long bracket with too many equal signs can overflow the 'int' used for
the counting and some arithmetic done on the value. Changing the counter
to 'size_t' avoids that. (Because what is counted goes to a buffer, an
overflow in the counter will first raise a buffer-overflow error.)

Bug-fix: joining an upvalue with itself could cause a use-after-free
crash.

The result of 'string.format("%a", 0.0)' can have multiple zeros
after the dot.

The file 'bugs' reports bugs in several different versions
(corresponding to different branches in the repository), without
a clear division of "this bugs belongs to this version". So, it
doesn't make sense to track it along with one (or many) versions.

While building a new prototype, the GC needs barriers for every object
(strings and nested prototypes) that is attached to the new prototype.

According to ISO C, 'realloc' can fail when shrinking a block. If that
happens, 'l_alloc' simply ignores the fail and returns the original
block.

This directory is used for some tests. As standard Lua has no command
to create directories, it must be present before running the tests.

When "undumping" a long string, the function 'LoadVector' can call the
reader function, which can run the garbage collector, which can collect
the string being read. So, the string must be anchored during the call
to 'LoadVector'. (This commit also fixes the identation in 'l_alloc'.)

ok pctammela@

ok lneto@

ok lneto@

ok lneto@

ok lneto@

Always use GFP_ATOMIC on krealloc because it cannot be called with
GFP_KERNEL while holding a spin_lock.

…rsions >= 5.6.0

This reverts commit 6f12b9f.

- we should not change API semantics like this; instead, we should add
  another one if needed.

- see:
** https://lwn.net/Articles/813350/
** https://lwn.net/Articles/835121/
** linux/Documentation/core-api/timekeeping.rst
** torvalds/linux@c0891ac

…he kernel"

This reverts commit 6742351.

This reverts commit 8ab5aba.

* if needed, we should implement this as a library