niceSOFT/sqlite: Official Git mirror of the SQLite source tree

doc/lemon.html
1 282 строки49 KB
drh
Experimental changes that prevent parser stack overflows by growing the
26 янв 2024, 23:34
26 янв 2024, 23:3482bf137
<html>
<head>
<title>The Lemon Parser Generator</title>
</head>
<body>
<a id="main"></a>
<h1 align='center'>The Lemon Parser Generator</h1>

<p>Lemon is an LALR(1) parser generator for C.
It does the same job as "bison" and "yacc".
But Lemon is not a bison or yacc clone.  Lemon
uses a different grammar syntax which is designed to
reduce the number of coding errors.  Lemon also uses a
parsing engine that is faster than yacc and
bison and which is both reentrant and threadsafe.
(Update: Since the previous sentence was written, bison
has also been updated so that it too can generate a
reentrant and threadsafe parser.)
Lemon also implements features that can be used
to eliminate resource leaks, making it suitable for use
in long-running programs such as graphical user interfaces
or embedded controllers.</p>

<p>This document is an introduction to the Lemon
parser generator.</p>

<a id="toc"></a>
<h2>1.0 Table of Contents</h2>
<ul>
<li><a href="#main">Introduction</a>
<li><a href="#toc">1.0 Table of Contents</a>
<li><a href="#secnot">2.0 Security Notes</a><br>
<li><a href="#optheory">3.0 Theory of Operation</a>
    <ul>
    <li><a href="#options">3.1 Command Line Options</a>
    <li><a href="#interface">3.2 The Parser Interface</a>
        <ul>
        <li><a href="#onstack">3.2.1 Allocating The Parse Object On Stack</a>
        <li><a href="#ifsum">3.2.2 Interface Summary</a>
        </ul>
    <li><a href="#yaccdiff">3.3 Differences With YACC and BISON</a>
    <li><a href="#build">3.4 Building The "lemon" Or "lemon.exe" Executable</a>
    </ul>
<li><a href="#syntax">4.0 Input File Syntax</a>
    <ul>
    <li><a href="#tnt">4.1 Terminals and Nonterminals</a>
    <li><a href="#rules">4.2 Grammar Rules</a>
    <li><a href="#precrules">4.3 Precedence Rules</a>
    <li><a href="#special">4.4 Special Directives</a>
    </ul>
<li><a href="#errors">5.0 Error Processing</a>
<li><a href="#history">6.0 History of Lemon</a>
<li><a href="#copyright">7.0 Copyright</a>
</ul>

<a id="secnot"></a>
<h2>2.0 Security Note</h2>

<p>The language parser code created by Lemon is very robust and
is well-suited for use in internet-facing applications that need to
safely process maliciously crafted inputs.</p>

<p>The "lemon.exe" command-line tool itself works great when given a valid
input grammar file and almost always gives helpful
error messages for malformed inputs.  However,  it is possible for
a malicious user to craft a grammar file that will cause 
lemon.exe to crash.
We do not see this as a problem, as lemon.exe is not intended to be used
with hostile inputs.
To summarize:</p>

<ul>
<li>Parser code generated by lemon &rarr; Robust and secure
<li>The "lemon.exe" command line tool itself &rarr; Not so much
</ul>

<a id="optheory"></a>
<h2>3.0 Theory of Operation</h2>

<p>Lemon is computer program that translates a context free grammar (CFG)
for a particular language into C code that implements a parser for
that language.
The Lemon program has two inputs:</p>
<ul>
<li>The grammar specification.
<li>A parser template file.
</ul>
<p>Typically, only the grammar specification is supplied by the programmer.
Lemon comes with a default parser template
("<a href="https://sqlite.org/src/file/tool/lempar.c">lempar.c</a>")
that works fine for most applications.  But the user is free to substitute
a different parser template if desired.</p>

<p>Depending on command-line options, Lemon will generate up to
three output files.</p>
<ul>
<li>C code to implement a parser for the input grammar.
<li>A header file defining an integer ID for each terminal symbol
    (or "token").
<li>An information file that describes the states of the generated parser
    automaton.
</ul>
<p>By default, all three of these output files are generated.
The header file is suppressed if the "-m" command-line option is
used and the report file is omitted when "-q" is selected.</p>

<p>The grammar specification file uses a ".y" suffix, by convention.
In the examples used in this document, we'll assume the name of the
grammar file is "gram.y".  A typical use of Lemon would be the
following command:</p>
<pre>
   lemon gram.y
</pre>
<p>This command will generate three output files named "gram.c",
"gram.h" and "gram.out".
The first is C code to implement the parser.  The second
is the header file that defines numerical values for all
terminal symbols, and the last is the report that explains
the states used by the parser automaton.</p>

<a id="options"></a>
<h3>3.1 Command Line Options</h3>

<p>The behavior of Lemon can be modified using command-line options.
You can obtain a list of the available command-line options together
with a brief explanation of what each does by typing</p>
<pre>
   lemon "-?"
</pre>
<p>As of this writing, the following command-line options are supported:</p>
<ul>
<li><b>-b</b>
Show only the basis for each parser state in the report file.
<li><b>-c</b>
Do not compress the generated action tables.  The parser will be a
little larger and slower, but it will detect syntax errors sooner.
<li><b>-d</b><i>directory</i>
Write all output files into <i>directory</i>.  Normally, output files
are written into the directory that contains the input grammar file.
<li><b>-D<i>name</i></b>
Define C preprocessor macro <i>name</i>.  This macro is usable by
"<tt><a href='#pifdef'>%ifdef</a></tt>",
"<tt><a href='#pifdef'>%ifndef</a></tt>", and
"<tt><a href="#pifdef">%if</a></tt> lines
in the grammar file.
<li><b>-E</b>
Run the "%if" preprocessor step only and print the revised grammar
file.
<li><b>-g</b>
Do not generate a parser.  Instead write the input grammar to standard
output with all comments, actions, and other extraneous text removed.
<li><b>-l</b>
Omit "#line" directives in the generated parser C code.
<li><b>-m</b>
Cause the output C source code to be compatible with the "makeheaders"
program.
<li><b>-p</b>
Display all conflicts that are resolved by
<a href='#precrules'>precedence rules</a>.
<li><b>-q</b>
Suppress generation of the report file.
<li><b>-r</b>
Do not sort or renumber the parser states as part of optimization.
<li><b>-s</b>
Show parser statistics before exiting.
<li><b>-T<i>file</i></b>
Use <i>file</i> as the template for the generated C-code parser implementation.
<li><b>-x</b>
Print the Lemon version number.
</ul>

<a id="interface"></a>
<h3>3.2 The Parser Interface</h3>

<p>Lemon doesn't generate a complete, working program.  It only generates
a few subroutines that implement a parser.  This section describes
the interface to those subroutines.  It is up to the programmer to
call these subroutines in an appropriate way in order to produce a
complete system.</p>

<p>Before a program begins using a Lemon-generated parser, the program
must first create the parser.
A new parser is created as follows:</p>
<pre>
   void *pParser = ParseAlloc( malloc );
</pre>
<p>The ParseAlloc() routine allocates and initializes a new parser and
returns a pointer to it.
The actual data structure used to represent a parser is opaque &mdash;
its internal structure is not visible or usable by the calling routine.
For this reason, the ParseAlloc() routine returns a pointer to void
rather than a pointer to some particular structure.
The sole argument to the ParseAlloc() routine is a pointer to the
subroutine used to allocate memory.  Typically this means malloc().</p>

<p>After a program is finished using a parser, it can reclaim all
memory allocated by that parser by calling</p>
<pre>
   ParseFree(pParser, free);
</pre>
<p>The first argument is the same pointer returned by ParseAlloc().  The
second argument is a pointer to the function used to release bulk
memory back to the system.</p>

<p>After a parser has been allocated using ParseAlloc(), the programmer
must supply the parser with a sequence of tokens (terminal symbols) to
be parsed.  This is accomplished by calling the following function
once for each token:<p>
<pre>
   Parse(pParser, hTokenID, sTokenData, pArg);
</pre>
<p>The first argument to the Parse() routine is the pointer returned by
ParseAlloc().
The second argument is a small positive integer that tells the parser the
type of the next token in the data stream.
There is one token type for each terminal symbol in the grammar.
The gram.h file generated by Lemon contains #define statements that
map symbolic terminal symbol names into appropriate integer values.
A value of 0 for the second argument is a special flag to the
parser to indicate that the end of input has been reached.
The third argument is the value of the given token.  By default,
the type of the third argument is "void*", but the grammar will
usually redefine this type to be some kind of structure.
Typically the second argument will be a broad category of tokens
such as "identifier" or "number" and the third argument will
be the name of the identifier or the value of the number.</p>

<p>The Parse() function may have either three or four arguments,
depending on the grammar.  If the grammar specification file requests
it (via the <tt><a href='#extraarg'>%extra_argument</a></tt> directive),
the Parse() function will have a fourth parameter that can be
of any type chosen by the programmer.  The parser doesn't do anything
with this argument except to pass it through to action routines.
This is a convenient mechanism for passing state information down
to the action routines without having to use global variables.</p>

<p>A typical use of a Lemon parser might look something like the
following:</p>
<pre>
    1 ParseTree *ParseFile(const char *zFilename){
    2    Tokenizer *pTokenizer;
    3    void *pParser;
    4    Token sToken;
    5    int hTokenId;
    6    ParserState sState;
    7
    8    pTokenizer = TokenizerCreate(zFilename);
    9    pParser = ParseAlloc( malloc );
   10    InitParserState(&amp;sState);
   11    while( GetNextToken(pTokenizer, &amp;hTokenId, &amp;sToken) ){
   12       Parse(pParser, hTokenId, sToken, &amp;sState);
   13    }
   14    Parse(pParser, 0, sToken, &amp;sState);
   15    ParseFree(pParser, free );
   16    TokenizerFree(pTokenizer);
   17    return sState.treeRoot;
   18 }
</pre>
<p>This example shows a user-written routine that parses a file of
text and returns a pointer to the parse tree.
(All error-handling code is omitted from this example to keep it
simple.)
We assume the existence of some kind of tokenizer which is created
using TokenizerCreate() on line 8 and deleted by TokenizerFree()
on line 16.  The GetNextToken() function on line 11 retrieves the
next token from the input file and puts its type in the
integer variable hTokenId.  The sToken variable is assumed to be
some kind of structure that contains details about each token,
such as its complete text, what line it occurs on, etc.</p>

<p>This example also assumes the existence of a structure of type
ParserState that holds state information about a particular parse.
An instance of such a structure is created on line 6 and initialized
on line 10.  A pointer to this structure is passed into the Parse()
routine as the optional 4th argument.
The action routine specified by the grammar for the parser can use
the ParserState structure to hold whatever information is useful and
appropriate.  In the example, we note that the treeRoot field of
the ParserState structure is left pointing to the root of the parse
tree.</p>

<p>The core of this example as it relates to Lemon is as follows:</p>
<pre>
   ParseFile(){
      pParser = ParseAlloc( malloc );
      while( GetNextToken(pTokenizer,&amp;hTokenId, &amp;sToken) ){
         Parse(pParser, hTokenId, sToken);
      }
      Parse(pParser, 0, sToken);
      ParseFree(pParser, free );
   }
</pre>
<p>Basically, what a program has to do to use a Lemon-generated parser
is first create the parser, then send it lots of tokens obtained by
tokenizing an input source.  When the end of input is reached, the
Parse() routine should be called one last time with a token type
of 0.  This step is necessary to inform the parser that the end of
input has been reached.  Finally, we reclaim memory used by the
parser by calling ParseFree().</p>

<p>There is one other interface routine that should be mentioned
before we move on.
The ParseTrace() function can be used to generate debugging output
from the parser.  A prototype for this routine is as follows:</p>
<pre>
   ParseTrace(FILE *stream, char *zPrefix);
</pre>
<p>After this routine is called, a short (one-line) message is written
to the designated output stream every time the parser changes states
or calls an action routine.  Each such message is prefaced using
the text given by zPrefix.  This debugging output can be turned off
by calling ParseTrace() again with a first argument of NULL (0).</p>

<a id="onstack"></a>
<h4>3.2.1 Allocating The Parse Object On Stack</h4>

<p>If all calls to the Parse() interface are made from within
<a href="#pcode"><tt>%code</tt> directives</a>, then the parse
object can be allocated from the stack rather than from the heap.
These are the steps:

<ul>
<li> Declare a local variable of type "yyParser"
<li> Initialize the variable using ParseInit()
<li> Pass a pointer to the variable in calls to Parse()
<li> Deallocate substructure in the parse variable using ParseFinalize().
</ul>

<p>The following code illustrates how this is done:

<pre>
   ParseFile(){
      yyParser x;
      ParseInit( &x );
      while( GetNextToken(pTokenizer,&amp;hTokenId, &amp;sToken) ){
         Parse(&x, hTokenId, sToken);
      }
      Parse(&x, 0, sToken);
      ParseFinalize( &x );
   }
</pre>

<a id="ifsum"></a>
<h4>3.2.2 Interface Summary</h4>

<p>Here is a quick overview of the C-language interface to a
Lemon-generated parser:</p>

<blockquote><pre>
void *ParseAlloc( (void*(*malloc)(size_t) );
void ParseFree(void *pParser, (void(*free)(void*) );
void Parse(void *pParser, int tokenCode, ParseTOKENTYPE token, ...);
void ParseTrace(FILE *stream, char *zPrefix);
</pre></blockquote>

<p>Notes:</p>
<ul>
<li> Use the <a href="#pname"><tt>%name</tt> directive</a> to change
the "Parse" prefix names of the procedures in the interface.
<li> Use the <a href="#token_type"><tt>%token_type</tt> directive</a>
to define the "ParseTOKENTYPE" type.
<li> Use the <a href="#extraarg"><tt>%extra_argument</tt> directive</a>
to specify the type and name of the 4th parameter to the
Parse() function.
</ul>

<a id="yaccdiff"></a>
<h3>3.3 Differences With YACC and BISON</h3>

<p>Programmers who have previously used the yacc or bison parser
generator will notice several important differences between yacc and/or
bison and Lemon.</p>
<ul>
<li>In yacc and bison, the parser calls the tokenizer.  In Lemon,
    the tokenizer calls the parser.
<li>Lemon uses no global variables.  Yacc and bison use global variables
    to pass information between the tokenizer and parser.
<li>Lemon allows multiple parsers to be running simultaneously.  Yacc
    and bison do not.
</ul>
<p>These differences may cause some initial confusion for programmers
with prior yacc and bison experience.
But after years of experience using Lemon, I firmly
believe that the Lemon way of doing things is better.</p>

<p><i>Updated as of 2016-02-16:</i>
The text above was written in the 1990s.
We are told that Bison has lately been enhanced to support the
tokenizer-calls-parser paradigm used by Lemon, eliminating the
need for global variables.</p>

<a id="build"><a>
<h3>3.4 Building The "lemon" or "lemon.exe" Executable</h3>

<p>The "lemon" or "lemon.exe" program is built from a single file
of C-code named 
"<a href="https://sqlite.org/src/tool/lemon.c">lemon.c</a>". 
The Lemon source code is generic C89 code that uses
no unusual or non-standard libraries.  Any
reasonable C compiler should suffice to compile the lemon program.  
A command-line like the following will usually work:</p>

<blockquote><pre>
cc -o lemon lemon.c
</pre></blockquote

<p>On Windows machines with Visual C++ installed, bring up a
"VS20<i>NN</i> x64 Native Tools Command Prompt" window and enter:

<blockquote><pre>
cl lemon.c
</pre></blockquote>

<p>Compiling Lemon really is that simple. 
Additional compiler options such as
"-O2" or "-g" or "-Wall" can be added if desired, but they are not
necessary.</p>


<a id="syntax"></a>
<h2>4.0 Input File Syntax</h2>

<p>The main purpose of the grammar specification file for Lemon is
to define the grammar for the parser.  But the input file also
specifies additional information Lemon requires to do its job.
Most of the work in using Lemon is in writing an appropriate
grammar file.</p>

<p>The grammar file for Lemon is, for the most part, a free format.
It does not have sections or divisions like yacc or bison.  Any
declaration can occur at any point in the file.  Lemon ignores
whitespace (except where it is needed to separate tokens), and it
honors the same commenting conventions as C and C++.</p>

<a id="tnt"></a>
<h3>4.1 Terminals and Nonterminals</h3>

<p>A terminal symbol (token) is any string of alphanumeric
and/or underscore characters
that begins with an uppercase letter.
A terminal can contain lowercase letters after the first character,
but the usual convention is to make terminals all uppercase.
A nonterminal, on the other hand, is any string of alphanumeric
and underscore characters than begins with a lowercase letter.
Again, the usual convention is to make nonterminals use all lowercase
letters.</p>

<p>In Lemon, terminal and nonterminal symbols do not need to
be declared or identified in a separate section of the grammar file.
Lemon is able to generate a list of all terminals and nonterminals
by examining the grammar rules, and it can always distinguish a
terminal from a nonterminal by checking the case of the first
character of the name.</p>

<p>Yacc and bison allow terminal symbols to have either alphanumeric
names or to be individual characters included in single quotes, like
this: ')' or '$'.  Lemon does not allow this alternative form for
terminal symbols.  With Lemon, all symbols, terminals and nonterminals,
must have alphanumeric names.</p>

<a id="rules"></a>
<h3>4.2 Grammar Rules</h3>

<p>The main component of a Lemon grammar file is a sequence of grammar
rules.
Each grammar rule consists of a nonterminal symbol followed by
the special symbol "::=" and then a list of terminals and/or nonterminals.
The rule is terminated by a period.
The list of terminals and nonterminals on the right-hand side of the
rule can be empty.
Rules can occur in any order, except that the left-hand side of the
first rule is assumed to be the start symbol for the grammar (unless
specified otherwise using the <tt><a href='#start_symbol'>%start_symbol</a></tt>
directive described below.)
A typical sequence of grammar rules might look something like this:</p>
<pre>
  expr ::= expr PLUS expr.
  expr ::= expr TIMES expr.
  expr ::= LPAREN expr RPAREN.
  expr ::= VALUE.
</pre>

<p>There is one non-terminal in this example, "expr", and five
terminal symbols or tokens: "PLUS", "TIMES", "LPAREN",
"RPAREN" and "VALUE".</p>

<p>Like yacc and bison, Lemon allows the grammar to specify a block
of C code that will be executed whenever a grammar rule is reduced
by the parser.
In Lemon, this action is specified by putting the C code (contained
within curly braces <tt>{...}</tt>) immediately after the
period that closes the rule.
For example:</p>
<pre>
  expr ::= expr PLUS expr.   { printf("Doing an addition...\n"); }
</pre>

<p>In order to be useful, grammar actions must normally be linked to
their associated grammar rules.
In yacc and bison, this is accomplished by embedding a "$$" in the
action to stand for the value of the left-hand side of the rule and
symbols "$1", "$2", and so forth to stand for the value of
the terminal or nonterminal at position 1, 2 and so forth on the
right-hand side of the rule.
This idea is very powerful, but it is also very error-prone.  The
single most common source of errors in a yacc or bison grammar is
to miscount the number of symbols on the right-hand side of a grammar
rule and say "$7" when you really mean "$8".</p>

<p>Lemon avoids the need to count grammar symbols by assigning symbolic
names to each symbol in a grammar rule and then using those symbolic
names in the action.
In yacc or bison, one would write this:</p>
<pre>
  expr -&gt; expr PLUS expr  { $$ = $1 + $3; };
</pre>
<p>But in Lemon, the same rule becomes the following:</p>
<pre>
  expr(A) ::= expr(B) PLUS expr(C).  { A = B+C; }
</pre>
<p>In the Lemon rule, any symbol in parentheses after a grammar rule
symbol becomes a place holder for that symbol in the grammar rule.
This place holder can then be used in the associated C action to
stand for the value of that symbol.</p>

<p>The Lemon notation for linking a grammar rule with its reduce
action is superior to yacc/bison on several counts.
First, as mentioned above, the Lemon method avoids the need to
count grammar symbols.
Secondly, if a terminal or nonterminal in a Lemon grammar rule
includes a linking symbol in parentheses but that linking symbol
is not actually used in the reduce action, then an error message
is generated.
For example, the rule</p>
<pre>
  expr(A) ::= expr(B) PLUS expr(C).  { A = B; }
</pre>
<p>will generate an error because the linking symbol "C" is used
in the grammar rule but not in the reduce action.</p>

<p>The Lemon notation for linking grammar rules to reduce actions
also facilitates the use of destructors for reclaiming memory
allocated by the values of terminals and nonterminals on the
right-hand side of a rule.</p>

<a id='precrules'></a>
<h3>4.3 Precedence Rules</h3>

<p>Lemon resolves parsing ambiguities in exactly the same way as
yacc and bison.  A shift-reduce conflict is resolved in favor
of the shift, and a reduce-reduce conflict is resolved by reducing
whichever rule comes first in the grammar file.</p>

<p>Just like in
yacc and bison, Lemon allows a measure of control
over the resolution of parsing conflicts using precedence rules.
A precedence value can be assigned to any terminal symbol
using the
<tt><a href='#pleft'>%left</a></tt>,
<tt><a href='#pright'>%right</a></tt> or
<tt><a href='#pnonassoc'>%nonassoc</a></tt> directives.  Terminal symbols
mentioned in earlier directives have a lower precedence than
terminal symbols mentioned in later directives.  For example:</p>

<pre>
   %left AND.
   %left OR.
   %nonassoc EQ NE GT GE LT LE.
   %left PLUS MINUS.
   %left TIMES DIVIDE MOD.
   %right EXP NOT.
</pre>

<p>In the preceding sequence of directives, the AND operator is
defined to have the lowest precedence.  The OR operator is one
precedence level higher.  And so forth.  Hence, the grammar would
attempt to group the ambiguous expression</p>
<pre>
     a AND b OR c
</pre>
<p>like this</p>
<pre>
     a AND (b OR c).
</pre>
<p>The associativity (left, right or nonassoc) is used to determine
the grouping when the precedence is the same.  AND is left-associative
in our example, so</p>
<pre>
     a AND b AND c
</pre>
<p>is parsed like this</p>
<pre>
     (a AND b) AND c.
</pre>
<p>The EXP operator is right-associative, though, so</p>
<pre>
     a EXP b EXP c
</pre>
<p>is parsed like this</p>
<pre>
     a EXP (b EXP c).
</pre>
<p>The nonassoc precedence is used for non-associative operators.
So</p>
<pre>
     a EQ b EQ c
</pre>
<p>is an error.</p>

<p>The precedence of non-terminals is transferred to rules as follows:
The precedence of a grammar rule is equal to the precedence of the
left-most terminal symbol in the rule for which a precedence is
defined.  This is normally what you want, but in those cases where
you want the precedence of a grammar rule to be something different,
you can specify an alternative precedence symbol by putting the
symbol in square braces after the period at the end of the rule and
before any C-code.  For example:</p>

<pre>
   expr = MINUS expr.  [NOT]
</pre>

<p>This rule has a precedence equal to that of the NOT symbol, not the
MINUS symbol as would have been the case by default.</p>

<p>With the knowledge of how precedence is assigned to terminal
symbols and individual
grammar rules, we can now explain precisely how parsing conflicts
are resolved in Lemon.  Shift-reduce conflicts are resolved
as follows:</p>
<ul>
<li> If either the token to be shifted or the rule to be reduced
     lacks precedence information, then resolve in favor of the
     shift, but report a parsing conflict.
<li> If the precedence of the token to be shifted is greater than
     the precedence of the rule to reduce, then resolve in favor
     of the shift.  No parsing conflict is reported.
<li> If the precedence of the token to be shifted is less than the
     precedence of the rule to reduce, then resolve in favor of the
     reduce action.  No parsing conflict is reported.
<li> If the precedences are the same and the shift token is
     right-associative, then resolve in favor of the shift.
     No parsing conflict is reported.
<li> If the precedences are the same and the shift token is
     left-associative, then resolve in favor of the reduce.
     No parsing conflict is reported.
<li> Otherwise, resolve the conflict by doing the shift, and
     report a parsing conflict.
</ul>
<p>Reduce-reduce conflicts are resolved this way:</p>
<ul>
<li> If either reduce rule
     lacks precedence information, then resolve in favor of the
     rule that appears first in the grammar, and report a parsing
     conflict.
<li> If both rules have precedence and the precedence is different,
     then resolve the dispute in favor of the rule with the highest
     precedence, and do not report a conflict.
<li> Otherwise, resolve the conflict by reducing by the rule that
     appears first in the grammar, and report a parsing conflict.
</ul>

<a id="special"></a>
<h3>4.4 Special Directives</h3>

<p>The input grammar to Lemon consists of grammar rules and special
directives.  We've described all the grammar rules, so now we'll
talk about the special directives.</p>

<p>Directives in Lemon can occur in any order.  You can put them before
the grammar rules, or after the grammar rules, or in the midst of the
grammar rules.  It doesn't matter.  The relative order of
directives used to assign precedence to terminals is important, but
other than that, the order of directives in Lemon is arbitrary.</p>

<p>Lemon supports the following special directives:</p>
<ul>
<li><tt><a href='#pcode'>%code</a></tt>
<li><tt><a href='#default_destructor'>%default_destructor</a></tt>
<li><tt><a href='#default_type'>%default_type</a></tt>
<li><tt><a href='#destructor'>%destructor</a></tt>
<li><tt><a href='#pifdef'>%else</a></tt>
<li><tt><a href='#pifdef'>%endif</a></tt>
<li><tt><a href='#extraarg'>%extra_argument</a></tt>
<li><tt><a href='#pfallback'>%fallback</a></tt>
<li><tt><a href='#reallc'>%free</a></tt>
<li><tt><a href='#pifdef'>%if</a></tt>
<li><tt><a href='#pifdef'>%ifdef</a></tt>
<li><tt><a href='#pifdef'>%ifndef</a></tt>
<li><tt><a href='#pinclude'>%include</a></tt>
<li><tt><a href='#pleft'>%left</a></tt>
<li><tt><a href='#pname'>%name</a></tt>
<li><tt><a href='#pnonassoc'>%nonassoc</a></tt>
<li><tt><a href='#parse_accept'>%parse_accept</a></tt>
<li><tt><a href='#parse_failure'>%parse_failure</a></tt>
<li><tt><a href='#pright'>%right</a></tt>
<li><tt><a href='#reallc'>%realloc</a></tt>
<li><tt><a href='#stack_overflow'>%stack_overflow</a></tt>
<li><tt><a href='#stack_size'>%stack_size</a></tt>
<li><tt><a href='#start_symbol'>%start_symbol</a></tt>
<li><tt><a href='#syntax_error'>%syntax_error</a></tt>
<li><tt><a href='#token'>%token</a></tt>
<li><tt><a href='#token_class'>%token_class</a></tt>
<li><tt><a href='#token_destructor'>%token_destructor</a></tt>
<li><tt><a href='#token_prefix'>%token_prefix</a></tt>
<li><tt><a href='#token_type'>%token_type</a></tt>
<li><tt><a href='#ptype'>%type</a></tt>
<li><tt><a href='#pwildcard'>%wildcard</a></tt>
</ul>
<p>Each of these directives will be described separately in the
following sections:</p>

<a id='pcode'></a>
<h4>4.4.1 The <tt>%code</tt> directive</h4>

<p>The <tt>%code</tt> directive is used to specify additional C code that
is added to the end of the main output file.  This is similar to
the <tt><a href='#pinclude'>%include</a></tt> directive except that
<tt>%include</tt> is inserted at the beginning of the main output file.</p>

<p><tt>%code</tt> is typically used to include some action routines or perhaps
a tokenizer or even the "main()" function
as part of the output file.</p>

<p>There can be multiple <tt>%code</tt> directives.  The arguments of
all <tt>%code</tt> directives are concatenated.</p>

<a id='default_destructor'></a>
<h4>4.4.2 The <tt>%default_destructor</tt> directive</h4>

<p>The <tt>%default_destructor</tt> directive specifies a destructor to
use for non-terminals that do not have their own destructor
specified by a separate <tt>%destructor</tt> directive.  See the documentation
on the <tt><a href='#destructor'>%destructor</a></tt> directive below for
additional information.</p>

<p>In some grammars, many different non-terminal symbols have the
same data type and hence the same destructor.  This directive is
a convenient way to specify the same destructor for all those
non-terminals using a single statement.</p>

<a id='default_type'></a>
<h4>4.4.3 The <tt>%default_type</tt> directive</h4>

<p>The <tt>%default_type</tt> directive specifies the data type of non-terminal
symbols that do not have their own data type defined using a separate
<tt><a href='#ptype'>%type</a></tt> directive.</p>

<a id='destructor'></a>
<h4>4.4.4 The <tt>%destructor</tt> directive</h4>

<p>The <tt>%destructor</tt> directive is used to specify a destructor for
a non-terminal symbol.
(See also the <tt><a href='#token_destructor'>%token_destructor</a></tt>
directive which is used to specify a destructor for terminal symbols.)</p>

<p>A non-terminal's destructor is called to dispose of the
non-terminal's value whenever the non-terminal is popped from
the stack.  This includes all of the following circumstances:</p>
<ul>
<li> When a rule reduces and the value of a non-terminal on
     the right-hand side is not linked to C code.
<li> When the stack is popped during error processing.
<li> When the ParseFree() function runs.
</ul>
<p>The destructor can do whatever it wants with the value of
the non-terminal, but its design is to deallocate memory
or other resources held by that non-terminal.</p>

<p>Consider an example:</p>
<pre>
   %type nt {void*}
   %destructor nt { free($$); }
   nt(A) ::= ID NUM.   { A = malloc( 100 ); }
</pre>
<p>This example is a bit contrived, but it serves to illustrate how
destructors work.  The example shows a non-terminal named
"nt" that holds values of type "void*".  When the rule for
an "nt" reduces, it sets the value of the non-terminal to
space obtained from malloc().  Later, when the nt non-terminal
is popped from the stack, the destructor will fire and call
free() on this malloced space, thus avoiding a memory leak.
(Note that the symbol "$$" in the destructor code is replaced
by the value of the non-terminal.)</p>

<p>It is important to note that the value of a non-terminal is passed
to the destructor whenever the non-terminal is removed from the
stack, unless the non-terminal is used in a C-code action.  If
the non-terminal is used by C-code, then it is assumed that the
C-code will take care of destroying it.
More commonly, the value is used to build some
larger structure, and we don't want to destroy it, which is why
the destructor is not called in this circumstance.</p>

<p>Destructors help avoid memory leaks by automatically freeing
allocated objects when they go out of scope.
To do the same using yacc or bison is much more difficult.</p>

<a id='extraarg'></a>
<h4>4.4.5 The <tt>%extra_argument</tt> directive</h4>

<p>The <tt>%extra_argument</tt> directive instructs Lemon to add a 4th parameter
to the parameter list of the Parse() function it generates.  Lemon
doesn't do anything itself with this extra argument, but it does
make the argument available to C-code action routines, destructors,
and so forth.  For example, if the grammar file contains:</p>

<pre>
    %extra_argument { MyStruct *pAbc }
</pre>

<p>Then the Parse() function generated will have an 4th parameter
of type "MyStruct*" and all action routines will have access to
a variable named "pAbc" that is the value of the 4th parameter
in the most recent call to Parse().</p>

<p>The <tt>%extra_context</tt> directive works the same except that it
is passed in on the ParseAlloc() or ParseInit() routines instead of
on Parse().</p>

<a id='extractx'></a>
<h4>4.4.6 The <tt>%extra_context</tt> directive</h4>

<p>The <tt>%extra_context</tt> directive instructs Lemon to add a 2nd parameter
to the parameter list of the ParseAlloc() and ParseInit() functions.  Lemon
doesn't do anything itself with these extra argument, but it does
store the value make it available to C-code action routines, destructors,
and so forth.  For example, if the grammar file contains:</p>

<pre>
    %extra_context { MyStruct *pAbc }
</pre>

<p>Then the ParseAlloc() and ParseInit() functions will have an 2nd parameter
of type "MyStruct*" and all action routines will have access to
a variable named "pAbc" that is the value of that 2nd parameter.</p>

<p>The <tt>%extra_argument</tt> directive works the same except that it
is passed in on the Parse() routine instead of on ParseAlloc()/ParseInit().</p>

<a id='pfallback'></a>
<h4>4.4.7 The <tt>%fallback</tt> directive</h4>

<p>The <tt>%fallback</tt> directive specifies an alternative meaning for one
or more tokens.  The alternative meaning is tried if the original token
would have generated a syntax error.</p>

<p>The <tt>%fallback</tt> directive was added to support robust parsing of SQL
syntax in <a href='https://www.sqlite.org/'>SQLite</a>.
The SQL language contains a large assortment of keywords, each of which
appears as a different token to the language parser.  SQL contains so
many keywords that it can be difficult for programmers to keep up with
them all.  Programmers will, therefore, sometimes mistakenly use an
obscure language keyword for an identifier.  The <tt>%fallback</tt> directive
provides a mechanism to tell the parser:  "If you are unable to parse
this keyword, try treating it as an identifier instead."</p>

<p>The syntax of <tt>%fallback</tt> is as follows:</p>

<blockquote>
<tt>%fallback</tt> <i>ID</i> <i>TOKEN...</i> <b>.</b>
</blockquote></p>

<p>In words, the <tt>%fallback</tt> directive is followed by a list of token
names terminated by a period.
The first token name is the fallback token &mdash; the
token to which all the other tokens fall back to.  The second and subsequent
arguments are tokens which fall back to the token identified by the first
argument.</p>

<a id='pifdef'></a>
<h4>4.4.8 The <tt>%if</tt> directive and its friends</h4>

<p>The <tt>%if</tt>, <tt>%ifdef</tt>, <tt>%ifndef</tt>, <tt>%else</tt>,
and <tt>%endif</tt> directives
are similar to #if, #ifdef, #ifndef, #else, and #endif in the C-preprocessor,
just not as general.
Each of these directives must begin at the left margin.  No whitespace
is allowed between the "%" and the directive name.</p>

<p>Grammar text in between "<tt>%ifdef MACRO</tt>" and the next nested
"<tt>%endif</tt>" is
ignored unless the "-DMACRO" command-line option is used.  Grammar text
betwen "<tt>%ifndef MACRO</tt>" and the next nested "<tt>%endif</tt>" is
included except when the "-DMACRO" command-line option is used.<p>

<p>The text in between "<tt>%if</tt> <i>CONDITIONAL</i>" and its
corresponding <tt>%endif</tt> is included only if <i>CONDITIONAL</i>
is true.  The CONDITION is one or more macro names, optionally connected
using the "||" and "&amp;&amp;" binary operators, the "!" unary operator,
and grouped using balanced parentheses.  Each term is true if the
corresponding macro exists, and false if it does not exist.</p>

<p>An optional "<tt>%else</tt>" directive can occur anywhere in between a 
<tt>%ifdef</tt>, <tt>%ifndef</tt>, or <tt>%if</tt> directive and
its corresponding <tt>%endif</tt>.</p>

<p>Note that the argument to <tt>%ifdef</tt> and <tt>%ifndef</tt> is
intended to be a single preprocessor symbol name, not a general expression.
Use the "<tt>%if</tt>" directive for general expressions.</p>

<a id='pinclude'></a>
<h4>4.4.9 The <tt>%include</tt> directive</h4>

<p>The <tt>%include</tt> directive specifies C code that is included at the
top of the generated parser.  You can include any text you want &mdash;
the Lemon parser generator copies it blindly.  If you have multiple
<tt>%include</tt> directives in your grammar file, their values are concatenated
so that all <tt>%include</tt> code ultimately appears near the top of the
generated parser, in the same order as it appeared in the grammar.</p>

<p>The <tt>%include</tt> directive is very handy for getting some extra #include
preprocessor statements at the beginning of the generated parser.
For example:</p>

<pre>
   %include {#include &lt;unistd.h&gt;}
</pre>

<p>This might be needed, for example, if some of the C actions in the
grammar call functions that are prototyped in unistd.h.</p>

<p>Use the <tt><a href="#pcode">%code</a></tt> directive to add code to
the end of the generated parser.</p>

<a id='pleft'></a>
<h4>4.4.10 The <tt>%left</tt> directive</h4>

The <tt>%left</tt> directive is used (along with the
<tt><a href='#pright'>%right</a></tt> and
<tt><a href='#pnonassoc'>%nonassoc</a></tt> directives) to declare
precedences of terminal symbols.
Every terminal symbol whose name appears after
a <tt>%left</tt> directive but before the next period (".") is
given the same left-associative precedence value.  Subsequent
<tt>%left</tt> directives have higher precedence.  For example:</p>

<pre>
   %left AND.
   %left OR.
   %nonassoc EQ NE GT GE LT LE.
   %left PLUS MINUS.
   %left TIMES DIVIDE MOD.
   %right EXP NOT.
</pre>

<p>Note the period that terminates each <tt>%left</tt>,
<tt>%right</tt> or <tt>%nonassoc</tt>
directive.</p>

<p>LALR(1) grammars can get into a situation where they require
a large amount of stack space if you make heavy use or right-associative
operators.  For this reason, it is recommended that you use <tt>%left</tt>
rather than <tt>%right</tt> whenever possible.</p>

<a id='pname'></a>
<h4>4.4.11 The <tt>%name</tt> directive</h4>

<p>By default, the functions generated by Lemon all begin with the
five-character string "Parse".  You can change this string to something
different using the <tt>%name</tt> directive.  For instance:</p>

<pre>
   %name Abcde
</pre>

<p>Putting this directive in the grammar file will cause Lemon to generate
functions named</p>
<ul>
<li> AbcdeAlloc(),
<li> AbcdeFree(),
<li> AbcdeTrace(), and
<li> Abcde().
</ul>
</p>The <tt>%name</tt> directive allows you to generate two or more different
parsers and link them all into the same executable.</p>

<a id='pnonassoc'></a>
<h4>4.4.12 The <tt>%nonassoc</tt> directive</h4>

<p>This directive is used to assign non-associative precedence to
one or more terminal symbols.  See the section on
<a href='#precrules'>precedence rules</a>
or on the <tt><a href='#pleft'>%left</a></tt> directive
for additional information.</p>

<a id='parse_accept'></a>
<h4>4.4.13 The <tt>%parse_accept</tt> directive</h4>

<p>The <tt>%parse_accept</tt> directive specifies a block of C code that is
executed whenever the parser accepts its input string.  To "accept"
an input string means that the parser was able to process all tokens
without error.</p>

<p>For example:</p>

<pre>
   %parse_accept {
      printf("parsing complete!\n");
   }
</pre>

<a id='parse_failure'></a>
<h4>4.4.14 The <tt>%parse_failure</tt> directive</h4>

<p>The <tt>%parse_failure</tt> directive specifies a block of C code that
is executed whenever the parser fails complete.  This code is not
executed until the parser has tried and failed to resolve an input
error using is usual error recovery strategy.  The routine is
only invoked when parsing is unable to continue.</p>

<pre>
   %parse_failure {
     fprintf(stderr,"Giving up.  Parser is hopelessly lost...\n");
   }
</pre>

<a id='pright'></a>
<h4>4.4.15 The <tt>%right</tt> directive</h4>

<p>This directive is used to assign right-associative precedence to
one or more terminal symbols.  See the section on
<a href='#precrules'>precedence rules</a>
or on the <a href='#pleft'>%left</a> directive for additional information.</p>

<a id='stack_overflow'></a>
<h4>4.4.16 The <tt>%stack_overflow</tt> directive</h4>

<p>The <tt>%stack_overflow</tt> directive specifies a block of C code that
is executed if the parser's internal stack ever overflows.  Typically
this just prints an error message.  After a stack overflow, the parser
will be unable to continue and must be reset.</p>

<pre>
   %stack_overflow {
     fprintf(stderr,"Giving up.  Parser stack overflow\n");
   }
</pre>

<p>You can help prevent parser stack overflows by avoiding the use
of right recursion and right-precedence operators in your grammar.
Use left recursion and and left-precedence operators instead to
encourage rules to reduce sooner and keep the stack size down.
For example, do rules like this:</p>
<pre>
   list ::= list element.      // left-recursion.  Good!
   list ::= .
</pre>
<p>Not like this:</p>
<pre>
   list ::= element list.      // right-recursion.  Bad!
   list ::= .
</pre>

<a id='stack_size'></a>
<h4>4.4.17 The <tt>%stack_size</tt> directive</h4>

<p>If stack overflow is a problem and you can't resolve the trouble
by using left-recursion, then you might want to increase the size
of the parser's stack using this directive.  Put an positive integer
after the <tt>%stack_size</tt> directive and Lemon will generate a parse
with a stack of the requested size.  The default value is 100.</p>

<pre>
   %stack_size 2000
</pre>

<a id='start_symbol'></a>
<h4>4.4.18 The <tt>%start_symbol</tt> directive</h4>

<p>By default, the start symbol for the grammar that Lemon generates
is the first non-terminal that appears in the grammar file.  But you
can choose a different start symbol using the
<tt>%start_symbol</tt> directive.</p>

<pre>
   %start_symbol  prog
</pre>

<a id='syntax_error'></a>
<h4>4.4.19 The <tt>%syntax_error</tt> directive</h4>

<p>See <a href='#errors'>Error Processing</a>.</p>

<a id='token'></a>
<h4>4.4.20 The <tt>%token</tt> directive</h4>

<p>Tokens are normally created automatically, the first time they are used.
Any identifier that begins with an upper-case letter is a token.

<p>Sometimes it is useful to declare tokens in advance, however.  The
integer values assigned to each token determined by the order in which
the tokens are seen.  So by declaring tokens in advance, it is possible to
cause some tokens to have low-numbered values, which might be desirable in
some grammers, or to have sequential values assigned to a sequence of
related tokens.  For this reason, the %token directive is provided to
declare tokens in advance.  The syntax is as follows:

<blockquote>
<tt>%token</tt> <i>TOKEN</i> <i>TOKEN...</i> <b>.</b>
</blockquote></p>

<p>The %token directive is followed by zero or more token symbols and
terminated by a single ".".  Each token named is created if it does not
already exist.  Tokens are created in order.


<a id='token_class'></a>
<h4>4.4.21 The <tt>%token_class</tt> directive</h4>

<p>Undocumented.  Appears to be related to the MULTITERMINAL concept.
<a href='http://sqlite.org/src/fdiff?v1=796930d5fc2036c7&v2=624b24c5dc048e09&sbs=0'>Implementation</a>.</p>

<a id='token_destructor'></a>
<h4>4.4.22 The <tt>%token_destructor</tt> directive</h4>

<p>The <tt>%destructor</tt> directive assigns a destructor to a non-terminal
symbol.  (See the description of the
<tt><a href='%destructor'>%destructor</a></tt> directive above.)
The <tt>%token_destructor</tt> directive does the same thing
for all terminal symbols.</p>

<p>Unlike non-terminal symbols, which may each have a different data type
for their values, terminals all use the same data type (defined by
the <tt><a href='#token_type'>%token_type</a></tt> directive)
and so they use a common destructor.
Other than that, the token destructor works just like the non-terminal
destructors.</p>

<a id='token_prefix'></a>
<h4>4.4.23 The <tt>%token_prefix</tt> directive</h4>

<p>Lemon generates #defines that assign small integer constants
to each terminal symbol in the grammar.  If desired, Lemon will
add a prefix specified by this directive
to each of the #defines it generates.</p>

<p>So if the default output of Lemon looked like this:</p>
<pre>
    #define AND              1
    #define MINUS            2
    #define OR               3
    #define PLUS             4
</pre>
<p>You can insert a statement into the grammar like this:</p>
<pre>
    %token_prefix    TOKEN_
</pre>
<p>to cause Lemon to produce these symbols instead:</p>
<pre>
    #define TOKEN_AND        1
    #define TOKEN_MINUS      2
    #define TOKEN_OR         3
    #define TOKEN_PLUS       4
</pre>

<a id='token_type'></a><a id='ptype'></a>
<h4>4.4.24 The <tt>%token_type</tt> and <tt>%type</tt> directives</h4>

<p>These directives are used to specify the data types for values
on the parser's stack associated with terminal and non-terminal
symbols.  The values of all terminal symbols must be of the same
type.  This turns out to be the same data type as the 3rd parameter
to the Parse() function generated by Lemon.  Typically, you will
make the value of a terminal symbol be a pointer to some kind of
token structure.  Like this:</p>

<pre>
   %token_type    {Token*}
</pre>

<p>If the data type of terminals is not specified, the default value
is "void*".</p>

<p>Non-terminal symbols can each have their own data types.  Typically
the data type of a non-terminal is a pointer to the root of a parse tree
structure that contains all information about that non-terminal.
For example:</p>

<pre>
   %type   expr  {Expr*}
</pre>

<p>Each entry on the parser's stack is actually a union containing
instances of all data types for every non-terminal and terminal symbol.
Lemon will automatically use the correct element of this union depending
on what the corresponding non-terminal or terminal symbol is.  But
the grammar designer should keep in mind that the size of the union
will be the size of its largest element.  So if you have a single
non-terminal whose data type requires 1K of storage, then your 100
entry parser stack will require 100K of heap space.  If you are willing
and able to pay that price, fine.  You just need to know.</p>

<a id='pwildcard'></a>
<h4>4.4.25 The <tt>%wildcard</tt> directive</h4>

<p>The <tt>%wildcard</tt> directive is followed by a single token name and a
period.  This directive specifies that the identified token should
match any input token.</p>

<p>When the generated parser has the choice of matching an input against
the wildcard token and some other token, the other token is always used.
The wildcard token is only matched if there are no alternatives.</p>

<a id='reallc'></a>
<h4>4.4.26 The <tt>%realloc</tt> and <tt>%free</tt> directives</h4>

<p>The <tt>%realloc</tt> and <tt>%free</tt> directives defines function
that allocate and free heap memory.  The signatures of these functions
should be the same as the realloc() and free() functions from the standard
C library.

<p>If both of these functions are defined
then these functions are used to allocate and free
memory for supplemental parser stack space, if the initial
parse stack space is exceeded.  The initial parser stack size
is specified by either <tt>%stack_size</tt> or the
-DYYSTACKDEPTH compile-time flag.

<a id='errors'></a>
<h2>5.0 Error Processing</h2>

<p>After extensive experimentation over several years, it has been
discovered that the error recovery strategy used by yacc is about
as good as it gets.  And so that is what Lemon uses.</p>

<p>When a Lemon-generated parser encounters a syntax error, it
first invokes the code specified by the <tt>%syntax_error</tt> directive, if
any.  It then enters its error recovery strategy.  The error recovery
strategy is to begin popping the parsers stack until it enters a
state where it is permitted to shift a special non-terminal symbol
named "error".  It then shifts this non-terminal and continues
parsing.  The <tt>%syntax_error</tt> routine will not be called again
until at least three new tokens have been successfully shifted.</p>

<p>If the parser pops its stack until the stack is empty, and it still
is unable to shift the error symbol, then the
<tt><a href='#parse_failure'>%parse_failure</a></tt> routine
is invoked and the parser resets itself to its start state, ready
to begin parsing a new file.  This is what will happen at the very
first syntax error, of course, if there are no instances of the
"error" non-terminal in your grammar.</p>


<a id='history'></a>
<h2>6.0 History of Lemon</h2>

<p>Lemon was originally written by Richard Hipp sometime in the late
1980s on a Sun4 Workstation using K&amp;R C.  
There was a companion LL(1) parser generator program named "Lime".
The Lime source code has been lost.</p>

<p>The lemon.c source file was originally many separate files that were
compiled together to generate the "lemon" executable.  Sometime in the
1990s, the individual source code files were combined together into
the current single large "lemon.c" source file.  You can still see traces
of original filenames in the code.</p>

<p>Since 2001, Lemon has been part of the 
<a href="https://sqlite.org/">SQLite project</a> and the source code
to Lemon has been managed as a part of the 
<a href="https://sqlite.org/src">SQLite source tree</a> in the following
files:</p>

<ul>
<li> <a href="https://sqlite.org/src/file/tool/lemon.c">tool/lemon.c</a>
<li> <a href="https://sqlite.org/src/file/tool/lempar.c">tool/lempar.c</a>
<li> <a href="https://sqlite.org/src/file/doc/lemon.html">doc/lemon.html</a>
</ul>

<a id="copyright"></a>
<h2>7.0 Copyright</h2>

<p>All of the source code to Lemon, including the template parser file
"lempar.c" and this documentation file ("lemon.html") are in the public
domain.  You can use the code for any purpose and without attribution.</p>

<p>The code comes with no warranty.  If it breaks, you get to keep both
pieces.</p>

</body>
</html>