This library is a set of APIs defined with module types, and a set of modules and functors implementing one or more of those interfaces.

The APIs define what a character and a string of characters should be.

See the INSTALL file for build instructions and/or the documentation website.

The library is “packed” in the Sosa toplevel module name.

We have, in the sub-module Api:

• BASIC_CHARACTER: characters of any length.
• NATIVE_CONVERSIONS: functions to transform from/to native OCaml strings.
• BASIC_STRING: immutable strings of (potentially abstract) characters:
  • includes NATIVE_CONVERSIONS,
  • contains a functor to provide a thread agnostic output function: Make_output: OUTPUT_MODEL → sig val output: ... end.
• UNSAFELY_MUTABLE: mutability of some string implementations (“unsafe” meaning that they break immutability invariants/assumptions).
• MINIMALISTIC_MUTABLE_STRING: abstract mutable string used as argument of the Of_mutable functor.

The Native_character module implements BASIC_CHARACTER with OCaml's char type.

The Native_string module implements BASIC_STRING with OCaml's string type considered immutable (and hence Native_character).

The Native_bytes module implements BASIC_STRING and UNSAFELY_MUTABLE with OCaml's bytes type.

List_of is a functor: BASIC_CHARACTER → BASIC_STRING, i.e., it creates a string datastructure made of a list of characters.

The functor Of_mutable uses an implementation of MINIMALISTIC_MUTABLE_STRING to build a BASIC_STRING.

The Int_utf8_character module implements BASIC_CHARACTER with OCaml integers (int) representing Utf8 characters (we force the handling of not more than 31 bits, even if RFC 3629 restricts them to end at U+10FFFF, c.f. also wikipedia). Note that the function is_whitespace considers only ASCII whitespace (useful while writing parsers for example).

See the file test/main.ml for usage examples, the library is tested with:

• native strings and characters,
• lists of native characters (List_of(Native_character)),
• lists of integers representing UTF-8 characters (List_of(utf8-int array)),
• arrays of integers representing UTF-8 characters (Of_mutable(utf8-int)),
• bigarrays of 8-bit integers (Of_mutable(int8 Bigarray1.t)).

The tests depend on the Nonstd, unix, and bigarray libraries:

make test
./sosa_tests

and you may add the basic benchmarks to the process with:

./sosa_tests bench