package base

README.org

* Base

Base is a standard library for OCaml. It provides a standard set of
general purpose modules that are well-tested, performant, and
fully-portable across any environment that can run OCaml code. Unlike
other standard library projects, Base is meant to be used as a
wholesale replacement of the standard library distributed with the
OCaml compiler. In particular it makes different choices and doesn't
re-export features that are not fully portable such as I/O, which are
left to other libraries.

You also might want to browse the [[https://ocaml.janestreet.com/ocaml-core/latest/doc/base/index.html][API Documentation]].

** Installation

Install Base via [[https://opam.ocaml.org][OPAM]]:

#+begin_src
$ opam install base
#+end_src

Base has no runtime dependencies and is fast to build. Its sole build
dependencies are [[https://github.com/ocaml/dune][dune]], which itself requires nothing more than the
compiler, and [[https://github.com/janestreet/sexplib0][sexplib0]].

** Using the OCaml standard library with Base

Base is intended as a full stdlib replacement.  As a result, after an
=open Base=, all the modules, values, types, ... coming from the OCaml
standard library that one normally gets in the default environment are
deprecated.

In order to access these values, one must use the =Caml= library,
which re-exports them all through the toplevel name =Caml=:
=Caml.String=, =Caml.print_string=, ...

The recommended way to build code using Base is as follows:

#+begin_src ocaml
$ ocamlc -open Base
#+end_src

** Differences between Base and the OCaml standard library

Programmers who are used to the OCaml standard library should read
through this section to understand major differences between the two
libraries that one should be aware of when switching to Base.

*** Comparison operators

The comparison operators exposed by the OCaml standard library are
polymorphic:

#+begin_src ocaml
val compare : 'a -> 'a -> int
val ( <= ) : 'a -> 'a -> bool
...
#+end_src

What they implement is structural comparison of the runtime
representation of values. Since these are often error-prone,
i.e. they don't correspond to what the user expects, they are not
exposed directly by Base.

To use polymorphic comparison with Base, one should use the
=Polymorphic_compare= module. The default comparison operators exposed
by Base are the integer ones, just like the default arithmetic
operators are the integer ones.

The recommended way to compare arbitrary complex data structures is to
use the specific =compare= functions. For instance:

#+begin_src ocaml
List.compare String.compare x y
#+end_src

The [[https://github.com/janestreet/ppx_compare][ppx_compare]] rewriter
offers an alternative way to write this:

#+begin_src ocaml
[%compare: string list] x y
#+end_src

** Base and ppx code generators

Base uses a few ppx code generators to implement:

- reliable and customizable comparison of OCaml values
- reliable and customizable hash of OCaml values
- conversions between OCaml values and s-expression

However, it doesn't need these code generators to build. What it does
instead is use ppx as a code verification tool during development. It
works in a very similar fashion to
[[https://github.com/janestreet/ppx_expect][expectation tests]].

Whenever you see this in the code source:

#+begin_src ocaml
type t = ... [@@deriving_inline sexp_of]
let sexp_of_t = ...
[@@@end]
#+end_src

the code between the =[@@deriving_inline]= and the =[@@@end]= is
generated code. The generated code is currently quite big and hard to
read, however we are working on making it look like human-written
code.

You can put the following elisp code in your =~/.emacs= file to hide
these blocks:

#+begin_src scheme
(defun deriving-inline-forward-sexp (&optional arg)
  (search-forward-regexp "\\[@@@end\\]") nil nil arg)

(defun setup-hide-deriving-inline ()
  (inline)
  (hs-minor-mode t)
  (let ((hs-hide-comments-when-hiding-all nil))
    (hs-hide-all)))

(require 'hideshow)
(add-to-list 'hs-special-modes-alist
             '(tuareg-mode "\\[@@deriving_inline[^]]*\\]" "\\[@@@end\\]" nil
                           deriving-inline-forward-sexp nil))
(add-hook 'tuareg-mode-hook 'setup-hide-deriving-inline)
#+end_src

Things are not yet setup in the git repository to make it convenient
to change types and update the generated code, but they will be setup
soon.

** Base coding rules

There are a few coding rules across the code base that are enforced by
lint tools.

These rules are:

- Opening the =Caml= module is not allowed. Inside Base, the OCaml
  stdlib is shadowed and accessible through the =Caml= module. We
  forbid opening =Caml= so that we know exactly where things come
  from.
- =Caml.Foo= modules cannot be aliased, one must use =Caml.Foo=
  explicitly. This is to avoid having to remember a list of aliases
  at the beginning of each file.
- For some modules that are both in the OCaml stdlib and Base, such as
  =String=, we define a module =String0= for common functions that
  cannot be defined directly in =Base.String= to avoid creating a
  circular dependency.  Except for =String= itself, other modules
  are not allowed to use =Caml.String= and must use either =String= or
  =String0= instead.
- Indentation is exactly the one of =ocp-indent=.
- A few other coding style rules enforced by
  [[https://github.com/janestreet/ppx_js_style][ppx_js_style]].

The Base specific coding rules are checked by =ppx_base_lint=, in the
=lint= subfolder. The indentation rules are checked by a wrapper around
=ocp-indent= and the coding style rules are checked by =ppx_js_style=.

These checks are currently not run by =dune=, but it will soon get a
=-dev= flag to run them automatically.

** Roadmap

Following is the current plan for a stable version 1 of Base.

*** Add more integer types

Add support for ={,u}int{8,16,32,64}=. These are always useful when
implementing binary protocols.

Initially they should be implemented with C stubs and eventually we
should propose their inclusion in the compiler.

*** 80 columns limit

Currently lines in Base are limited to a maximum width of 90
characters. To make things more standard, we should use an 80 columns
limit.  The only thing needed for this is to extend the style checker
to enforce a maximum line width.

*** Improve the generated code

Improve our code generators to produce code that looks more like
hand-written code.