A GC step may invoke some finalizer, which may error and emit
a warning due to stack overflfow.

- more organized handling of 'nCcalls'
- comments
- deprecation of 'setcstacklimit'

A "with stack" implementation gains too little in performance to be
worth all the noise from C-stack overflows.

This commit is almost a sketch, to test performance. There are several
pending stuff:

- review control of C-stack overflow and error messages;
- what to do with setcstacklimit;
- review comments;
- review unroll of Lua calls.

The macro 'luaL_pushfail' documents all places in the standard libraries
that return nil to signal some kind of failure. It is defined as
'lua_pushnil'. The manual also got a notation (@fail) to document those
returns. The tests were changed to be agnostic regarding whether 'fail'
is 'nil' or 'false'.

Added new functions to dynamically set the C-stack limit
('lua_setCstacklimit' in the C-API, 'debug.setCstacklimit' in Lua).

Bug introduced in commit 28d829c8: OP_TAILCALL might raise an
error without saving 'pc'. (This commit also fixes a detail in
'testes/uf8.lua'.)

"Better" and similar to error messages for invalid function arguments.
  *old message: 'for' limit must be a number
  *new message: bad 'for' limit (number expected, got table)

From the point of view of 'git', all names are relative to the root
directory of the project. So, file names in '$Id:' also should be
relative to that directory: the proper name for test file 'all.lua'
is 'testes/all.lua'.

When creating code for a jump on a 'not' condition, the code generator
was removing an instruction (the OP_NOT) without adjusting its
corresponding line information.

This fix also added tests for this case and extra functionality in
the test library to debug line info. structures.