xref: /llvm-project/llvm/docs/tutorial/MyFirstLanguageFrontend/LangImpl01.rst (revision 63cf7040814e3b190a3b5a65858d8d59ab47b74d)
1=====================================================
2Kaleidoscope: Kaleidoscope Introduction and the Lexer
3=====================================================
4
5.. contents::
6   :local:
7
8The Kaleidoscope Language
9=========================
10
11This tutorial is illustrated with a toy language called
12"`Kaleidoscope <http://en.wikipedia.org/wiki/Kaleidoscope>`_" (derived
13from "meaning beautiful, form, and view"). Kaleidoscope is a procedural
14language that allows you to define functions, use conditionals, math,
15etc. Over the course of the tutorial, we'll extend Kaleidoscope to
16support the if/then/else construct, a for loop, user defined operators,
17JIT compilation with a simple command line interface, debug info, etc.
18
19We want to keep things simple, so the only datatype in Kaleidoscope
20is a 64-bit floating point type (aka 'double' in C parlance). As such,
21all values are implicitly double precision and the language doesn't
22require type declarations. This gives the language a very nice and
23simple syntax. For example, the following simple example computes
24`Fibonacci numbers: <http://en.wikipedia.org/wiki/Fibonacci_number>`_
25
26::
27
28    # Compute the x'th fibonacci number.
29    def fib(x)
30      if x < 3 then
31        1
32      else
33        fib(x-1)+fib(x-2)
34
35    # This expression will compute the 40th number.
36    fib(40)
37
38We also allow Kaleidoscope to call into standard library functions - the
39LLVM JIT makes this really easy. This means that you can use the
40'extern' keyword to define a function before you use it (this is also
41useful for mutually recursive functions).  For example:
42
43::
44
45    extern sin(arg);
46    extern cos(arg);
47    extern atan2(arg1 arg2);
48
49    atan2(sin(.4), cos(42))
50
51A more interesting example is included in Chapter 6 where we write a
52little Kaleidoscope application that `displays a Mandelbrot
53Set <LangImpl06.html#kicking-the-tires>`_ at various levels of magnification.
54
55Let's dive into the implementation of this language!
56
57The Lexer
58=========
59
60When it comes to implementing a language, the first thing needed is the
61ability to process a text file and recognize what it says. The
62traditional way to do this is to use a
63"`lexer <http://en.wikipedia.org/wiki/Lexical_analysis>`_" (aka
64'scanner') to break the input up into "tokens". Each token returned by
65the lexer includes a token code and potentially some metadata (e.g. the
66numeric value of a number). First, we define the possibilities:
67
68.. code-block:: c++
69
70    // The lexer returns tokens [0-255] if it is an unknown character, otherwise one
71    // of these for known things.
72    enum Token {
73      tok_eof = -1,
74
75      // commands
76      tok_def = -2,
77      tok_extern = -3,
78
79      // primary
80      tok_identifier = -4,
81      tok_number = -5,
82    };
83
84    static std::string IdentifierStr; // Filled in if tok_identifier
85    static double NumVal;             // Filled in if tok_number
86
87Each token returned by our lexer will either be one of the Token enum
88values or it will be an 'unknown' character like '+', which is returned
89as its ASCII value. If the current token is an identifier, the
90``IdentifierStr`` global variable holds the name of the identifier. If
91the current token is a numeric literal (like 1.0), ``NumVal`` holds its
92value. We use global variables for simplicity, but this is not the
93best choice for a real language implementation :).
94
95The actual implementation of the lexer is a single function named
96``gettok``. The ``gettok`` function is called to return the next token
97from standard input. Its definition starts as:
98
99.. code-block:: c++
100
101    /// gettok - Return the next token from standard input.
102    static int gettok() {
103      static int LastChar = ' ';
104
105      // Skip any whitespace.
106      while (isspace(LastChar))
107        LastChar = getchar();
108
109``gettok`` works by calling the C ``getchar()`` function to read
110characters one at a time from standard input. It eats them as it
111recognizes them and stores the last character read, but not processed,
112in LastChar. The first thing that it has to do is ignore whitespace
113between tokens. This is accomplished with the loop above.
114
115The next thing ``gettok`` needs to do is recognize identifiers and
116specific keywords like "def". Kaleidoscope does this with this simple
117loop:
118
119.. code-block:: c++
120
121      if (isalpha(LastChar)) { // identifier: [a-zA-Z][a-zA-Z0-9]*
122        IdentifierStr = LastChar;
123        while (isalnum((LastChar = getchar())))
124          IdentifierStr += LastChar;
125
126        if (IdentifierStr == "def")
127          return tok_def;
128        if (IdentifierStr == "extern")
129          return tok_extern;
130        return tok_identifier;
131      }
132
133Note that this code sets the '``IdentifierStr``' global whenever it
134lexes an identifier. Also, since language keywords are matched by the
135same loop, we handle them here inline. Numeric values are similar:
136
137.. code-block:: c++
138
139      if (isdigit(LastChar) || LastChar == '.') {   // Number: [0-9.]+
140        std::string NumStr;
141        do {
142          NumStr += LastChar;
143          LastChar = getchar();
144        } while (isdigit(LastChar) || LastChar == '.');
145
146        NumVal = strtod(NumStr.c_str(), 0);
147        return tok_number;
148      }
149
150This is all pretty straightforward code for processing input. When
151reading a numeric value from input, we use the C ``strtod`` function to
152convert it to a numeric value that we store in ``NumVal``. Note that
153this isn't doing sufficient error checking: it will incorrectly read
154"1.23.45.67" and handle it as if you typed in "1.23". Feel free to
155extend it!  Next we handle comments:
156
157.. code-block:: c++
158
159      if (LastChar == '#') {
160        // Comment until end of line.
161        do
162          LastChar = getchar();
163        while (LastChar != EOF && LastChar != '\n' && LastChar != '\r');
164
165        if (LastChar != EOF)
166          return gettok();
167      }
168
169We handle comments by skipping to the end of the line and then return
170the next token. Finally, if the input doesn't match one of the above
171cases, it is either an operator character like '+' or the end of the
172file. These are handled with this code:
173
174.. code-block:: c++
175
176      // Check for end of file.  Don't eat the EOF.
177      if (LastChar == EOF)
178        return tok_eof;
179
180      // Otherwise, just return the character as its ascii value.
181      int ThisChar = LastChar;
182      LastChar = getchar();
183      return ThisChar;
184    }
185
186With this, we have the complete lexer for the basic Kaleidoscope
187language (the `full code listing <LangImpl02.html#full-code-listing>`_ for the Lexer
188is available in the `next chapter <LangImpl02.html>`_ of the tutorial).
189Next we'll `build a simple parser that uses this to build an Abstract
190Syntax Tree <LangImpl02.html>`_. When we have that, we'll include a
191driver so that you can use the lexer and parser together.
192
193`Next: Implementing a Parser and AST <LangImpl02.html>`_
194
195