1 =================================================
2 Kaleidoscope: Tutorial Introduction and the Lexer
3 =================================================
8 Written by `Chris Lattner <mailto:sabre@nondot.org>`_ and `Erick
9 Tryzelaar <mailto:idadesub@users.sourceforge.net>`_
14 Welcome to the "Implementing a language with LLVM" tutorial. This
15 tutorial runs through the implementation of a simple language, showing
16 how fun and easy it can be. This tutorial will get you up and started as
17 well as help to build a framework you can extend to other languages. The
18 code in this tutorial can also be used as a playground to hack on other
21 The goal of this tutorial is to progressively unveil our language,
22 describing how it is built up over time. This will let us cover a fairly
23 broad range of language design and LLVM-specific usage issues, showing
24 and explaining the code for it all along the way, without overwhelming
25 you with tons of details up front.
27 It is useful to point out ahead of time that this tutorial is really
28 about teaching compiler techniques and LLVM specifically, *not* about
29 teaching modern and sane software engineering principles. In practice,
30 this means that we'll take a number of shortcuts to simplify the
31 exposition. For example, the code leaks memory, uses global variables
32 all over the place, doesn't use nice design patterns like
33 `visitors <http://en.wikipedia.org/wiki/Visitor_pattern>`_, etc... but
34 it is very simple. If you dig in and use the code as a basis for future
35 projects, fixing these deficiencies shouldn't be hard.
37 I've tried to put this tutorial together in a way that makes chapters
38 easy to skip over if you are already familiar with or are uninterested
39 in the various pieces. The structure of the tutorial is:
41 - `Chapter #1 <#language>`_: Introduction to the Kaleidoscope
42 language, and the definition of its Lexer - This shows where we are
43 going and the basic functionality that we want it to do. In order to
44 make this tutorial maximally understandable and hackable, we choose
45 to implement everything in Objective Caml instead of using lexer and
46 parser generators. LLVM obviously works just fine with such tools,
47 feel free to use one if you prefer.
48 - `Chapter #2 <OCamlLangImpl2.html>`_: Implementing a Parser and
49 AST - With the lexer in place, we can talk about parsing techniques
50 and basic AST construction. This tutorial describes recursive descent
51 parsing and operator precedence parsing. Nothing in Chapters 1 or 2
52 is LLVM-specific, the code doesn't even link in LLVM at this point.
54 - `Chapter #3 <OCamlLangImpl3.html>`_: Code generation to LLVM IR -
55 With the AST ready, we can show off how easy generation of LLVM IR
57 - `Chapter #4 <OCamlLangImpl4.html>`_: Adding JIT and Optimizer
58 Support - Because a lot of people are interested in using LLVM as a
59 JIT, we'll dive right into it and show you the 3 lines it takes to
60 add JIT support. LLVM is also useful in many other ways, but this is
61 one simple and "sexy" way to shows off its power. :)
62 - `Chapter #5 <OCamlLangImpl5.html>`_: Extending the Language:
63 Control Flow - With the language up and running, we show how to
64 extend it with control flow operations (if/then/else and a 'for'
65 loop). This gives us a chance to talk about simple SSA construction
67 - `Chapter #6 <OCamlLangImpl6.html>`_: Extending the Language:
68 User-defined Operators - This is a silly but fun chapter that talks
69 about extending the language to let the user program define their own
70 arbitrary unary and binary operators (with assignable precedence!).
71 This lets us build a significant piece of the "language" as library
73 - `Chapter #7 <OCamlLangImpl7.html>`_: Extending the Language:
74 Mutable Variables - This chapter talks about adding user-defined
75 local variables along with an assignment operator. The interesting
76 part about this is how easy and trivial it is to construct SSA form
77 in LLVM: no, LLVM does *not* require your front-end to construct SSA
79 - `Chapter #8 <OCamlLangImpl8.html>`_: Conclusion and other useful
80 LLVM tidbits - This chapter wraps up the series by talking about
81 potential ways to extend the language, but also includes a bunch of
82 pointers to info about "special topics" like adding garbage
83 collection support, exceptions, debugging, support for "spaghetti
84 stacks", and a bunch of other tips and tricks.
86 By the end of the tutorial, we'll have written a bit less than 700 lines
87 of non-comment, non-blank, lines of code. With this small amount of
88 code, we'll have built up a very reasonable compiler for a non-trivial
89 language including a hand-written lexer, parser, AST, as well as code
90 generation support with a JIT compiler. While other systems may have
91 interesting "hello world" tutorials, I think the breadth of this
92 tutorial is a great testament to the strengths of LLVM and why you
93 should consider it if you're interested in language or compiler design.
95 A note about this tutorial: we expect you to extend the language and
96 play with it on your own. Take the code and go crazy hacking away at it,
97 compilers don't need to be scary creatures - it can be a lot of fun to
103 This tutorial will be illustrated with a toy language that we'll call
104 "`Kaleidoscope <http://en.wikipedia.org/wiki/Kaleidoscope>`_" (derived
105 from "meaning beautiful, form, and view"). Kaleidoscope is a procedural
106 language that allows you to define functions, use conditionals, math,
107 etc. Over the course of the tutorial, we'll extend Kaleidoscope to
108 support the if/then/else construct, a for loop, user defined operators,
109 JIT compilation with a simple command line interface, etc.
111 Because we want to keep things simple, the only datatype in Kaleidoscope
112 is a 64-bit floating point type (aka 'float' in O'Caml parlance). As
113 such, all values are implicitly double precision and the language
114 doesn't require type declarations. This gives the language a very nice
115 and simple syntax. For example, the following simple example computes
116 `Fibonacci numbers: <http://en.wikipedia.org/wiki/Fibonacci_number>`_
120 # Compute the x'th fibonacci number.
127 # This expression will compute the 40th number.
130 We also allow Kaleidoscope to call into standard library functions (the
131 LLVM JIT makes this completely trivial). This means that you can use the
132 'extern' keyword to define a function before you use it (this is also
133 useful for mutually recursive functions). For example:
139 extern atan2(arg1 arg2);
141 atan2(sin(.4), cos(42))
143 A more interesting example is included in Chapter 6 where we write a
144 little Kaleidoscope application that `displays a Mandelbrot
145 Set <OCamlLangImpl6.html#example>`_ at various levels of magnification.
147 Lets dive into the implementation of this language!
152 When it comes to implementing a language, the first thing needed is the
153 ability to process a text file and recognize what it says. The
154 traditional way to do this is to use a
155 "`lexer <http://en.wikipedia.org/wiki/Lexical_analysis>`_" (aka
156 'scanner') to break the input up into "tokens". Each token returned by
157 the lexer includes a token code and potentially some metadata (e.g. the
158 numeric value of a number). First, we define the possibilities:
160 .. code-block:: ocaml
162 (* The lexer returns these 'Kwd' if it is an unknown character, otherwise one of
163 * these others for known things. *)
169 | Ident of string | Number of float
174 Each token returned by our lexer will be one of the token variant
175 values. An unknown character like '+' will be returned as
176 ``Token.Kwd '+'``. If the curr token is an identifier, the value will be
177 ``Token.Ident s``. If the current token is a numeric literal (like 1.0),
178 the value will be ``Token.Number 1.0``.
180 The actual implementation of the lexer is a collection of functions
181 driven by a function named ``Lexer.lex``. The ``Lexer.lex`` function is
182 called to return the next token from standard input. We will use
183 `Camlp4 <http://caml.inria.fr/pub/docs/manual-camlp4/index.html>`_ to
184 simplify the tokenization of the standard input. Its definition starts
187 .. code-block:: ocaml
189 (*===----------------------------------------------------------------------===
191 *===----------------------------------------------------------------------===*)
194 (* Skip any whitespace. *)
195 | [< ' (' ' | '\n' | '\r' | '\t'); stream >] -> lex stream
197 ``Lexer.lex`` works by recursing over a ``char Stream.t`` to read
198 characters one at a time from the standard input. It eats them as it
199 recognizes them and stores them in in a ``Token.token`` variant. The
200 first thing that it has to do is ignore whitespace between tokens. This
201 is accomplished with the recursive call above.
203 The next thing ``Lexer.lex`` needs to do is recognize identifiers and
204 specific keywords like "def". Kaleidoscope does this with a pattern
205 match and a helper function.
207 .. code-block:: ocaml
209 (* identifier: [a-zA-Z][a-zA-Z0-9] *)
210 | [< ' ('A' .. 'Z' | 'a' .. 'z' as c); stream >] ->
211 let buffer = Buffer.create 1 in
212 Buffer.add_char buffer c;
213 lex_ident buffer stream
217 and lex_ident buffer = parser
218 | [< ' ('A' .. 'Z' | 'a' .. 'z' | '0' .. '9' as c); stream >] ->
219 Buffer.add_char buffer c;
220 lex_ident buffer stream
221 | [< stream=lex >] ->
222 match Buffer.contents buffer with
223 | "def" -> [< 'Token.Def; stream >]
224 | "extern" -> [< 'Token.Extern; stream >]
225 | id -> [< 'Token.Ident id; stream >]
227 Numeric values are similar:
229 .. code-block:: ocaml
231 (* number: [0-9.]+ *)
232 | [< ' ('0' .. '9' as c); stream >] ->
233 let buffer = Buffer.create 1 in
234 Buffer.add_char buffer c;
235 lex_number buffer stream
239 and lex_number buffer = parser
240 | [< ' ('0' .. '9' | '.' as c); stream >] ->
241 Buffer.add_char buffer c;
242 lex_number buffer stream
243 | [< stream=lex >] ->
244 [< 'Token.Number (float_of_string (Buffer.contents buffer)); stream >]
246 This is all pretty straight-forward code for processing input. When
247 reading a numeric value from input, we use the ocaml ``float_of_string``
248 function to convert it to a numeric value that we store in
249 ``Token.Number``. Note that this isn't doing sufficient error checking:
250 it will raise ``Failure`` if the string "1.23.45.67". Feel free to
251 extend it :). Next we handle comments:
253 .. code-block:: ocaml
255 (* Comment until end of line. *)
256 | [< ' ('#'); stream >] ->
261 and lex_comment = parser
262 | [< ' ('\n'); stream=lex >] -> stream
263 | [< 'c; e=lex_comment >] -> e
266 We handle comments by skipping to the end of the line and then return
267 the next token. Finally, if the input doesn't match one of the above
268 cases, it is either an operator character like '+' or the end of the
269 file. These are handled with this code:
271 .. code-block:: ocaml
273 (* Otherwise, just return the character as its ascii value. *)
274 | [< 'c; stream >] ->
275 [< 'Token.Kwd c; lex stream >]
280 With this, we have the complete lexer for the basic Kaleidoscope
281 language (the `full code listing <OCamlLangImpl2.html#code>`_ for the
282 Lexer is available in the `next chapter <OCamlLangImpl2.html>`_ of the
283 tutorial). Next we'll `build a simple parser that uses this to build an
284 Abstract Syntax Tree <OCamlLangImpl2.html>`_. When we have that, we'll
285 include a driver so that you can use the lexer and parser together.
287 `Next: Implementing a Parser and AST <OCamlLangImpl2.html>`_