We deprecate unspecified locality as was done for Hint.

Close #13724

By default Coq stdlib warnings raise an error, so this is really required.

As noted by Hugo Herbelin, Dec is rather used for "decidable".

Previously real constants were parsed by an unproved OCaml code. The
parser and printer are now implemented in Coq, which will enable a
proof and hopefully make it easier to maintain / make evolve.

Previously reals were all parsed as an integer, an integer multiplied
by a power of ten or an integer divided by a power of ten. This means
1.02 and 102e-2 were both parsed as 102 / 100 and could not be tell
apart when printing. So the printer had to choose between two
representations : without exponent or without decimal dot. The choice
was made heuristically toward a most compact representation.

Now, decimal dot is parsed as a rational and exponents are parsed as a
product or division by a power of ten. For instance, 1.02 is parsed as
Q2R (102 # 100) whereas 102e-2 is parsed as IZR 102 / IZR (Z.pow_pos
10 2).

1.02 and 102e-2 remain equal (proved by reflexivity) but 1.02e1 = Q2R
(102 # 100) * 10 and 10.2 = Q2R (102 # 10) no longer are.

Previously rationals were all parsed as a pair numerator, denominator.
This means 1.02 and 102e-2 were both parsed as 102 # 100 and could not
be tell apart when printing. So the printer had to choose between two
representations : without exponent or without decimal dot. The choice
was made heuristically toward a most compact representation.

Now, decimal dot is still parsed as a power of ten denominator but
exponents are parsed as a product or division by Z.pow_pos. For
instance, 1.02 is parsed as 102 # 100 whereas 102e-2 is parsed as
(102 # 1) / (Z.pow_pos 10 2 # 1).

1.02 and 102e-2 remain equal (proved by reflexivity) but 1.02e1
= (102 # 100) * (10 # 1) = 1020 # 100 and 10.2 = 102 # 10 no longer are.

A nice side effect is that exponents can no longer blow up during
parsing. Previously 1e1_000_000 literally produced a numerator with a
million digits, now it just yields (1 # 1) * (Z.pow_pos 10 1_000_000 # 1).

Keep Numeral Notation wit a deprecation warning.

We add hexadecimal numerals according to the following regexp
0[xX][0-9a-fA-F][0-9a-fA-F_]*(\.[0-9a-fA-F_]+)?([pP][+-]?[0-9][0-9_]*)?

This is unfortunately a rather large commit. I suggest reading it in
the following order:

* test-suite/output/ZSyntax.{v,out}       new test
* test-suite/output/Int63Syntax.{v,out}   ''
* test-suite/output/QArithSyntax.{v,out}  ''
* test-suite/output/RealSyntax.{v,out}    ''
* test-suite/output/FloatSyntax.{v,out}   ''

* interp/numTok.ml{i,}  extending numeral tokens

* theories/Init/Hexadecimal.v  adaptation of Decimal.v
                               for the new hexadecimal Numeral Notation
* theories/Init/Numeral.v  new interface for Numeral Notation (basically,
                           a numeral is either a decimal or an hexadecimal)

* theories/Init/Nat.v            add hexadecimal numeral notation to nat
* theories/PArith/BinPosDef.v    ''                                  positive
* theories/ZArith/BinIntDef.v    ''                                  Z
* theories/NArith/BinNatDef.v    ''                                  N
* theories/QArith/QArith_base.v  ''                                  Q

* interp/notation.ml{i,}     adapting implementation of numeral notations
* plugins/syntax/numeral.ml  ''

* plugins/syntax/r_syntax.ml  adapt parser for real numbers

* plugins/syntax/float_syntax.ml  adapt parser for primitive floats

* theories/Init/Prelude.v  register parser for nat
* adapting the test-suite (test-suite/output/NumeralNotations.{v,out}
  and test-suite/output/SearchPattern.out)
* remaining ml files (interp/constrex{tern,pr_ops}.ml where two open
  had to be permuted)