Comments re 5.2.2-rc2

["A Royal" <a.royal.public@...>]

>> 1.
On line 164 of lmathlib.c (version 5.2.2-rc2) there is an implicit
conversion of value 'ep'
-- Compiler warns "implicit conversion from floating point to integer"

This has been introduced via changes since 5.2.1. I'm not sure why this has
been changed.

Indeed, what is the intended behaviour if parameter 'ep' for 
l_mathop(ldexp)(x, ep) has a fraction component? I presume silent truncation
of value 'ep'.

Does this incur an additional type conversion? It seems a pity to cast 'ep'
from [probably an integer] (expected type) to [double] (type from
luaL_checknumber) to [integer] (implicit cast).

>> 2.
Macro 'MAXCCALLS' is a little confusing, since there is also the independent
macro LUAI_MAXCCALLS; the former appears to apply only to lstrlib.c ... Good
to see this recursion guard, though!

Regards,

AR