175897234Sdrh<html> 275897234Sdrh<head> 375897234Sdrh<title>The Lemon Parser Generator</title> 475897234Sdrh</head> 59a243e69Sdrh<body bgcolor='white'> 69a243e69Sdrh<h1 align='center'>The Lemon Parser Generator</h1> 775897234Sdrh 89bccde3dSdrh<p>Lemon is an LALR(1) parser generator for C. 99bccde3dSdrhIt does the same job as "bison" and "yacc". 109a243e69SdrhBut Lemon is not a bison or yacc clone. Lemon 1175897234Sdrhuses a different grammar syntax which is designed to 129bccde3dSdrhreduce the number of coding errors. Lemon also uses a 139bccde3dSdrhparsing engine that is faster than yacc and 149bccde3dSdrhbison and which is both reentrant and threadsafe. 159bccde3dSdrh(Update: Since the previous sentence was written, bison 169bccde3dSdrhhas also been updated so that it too can generate a 179bccde3dSdrhreentrant and threadsafe parser.) 189bccde3dSdrhLemon also implements features that can be used 199a243e69Sdrhto eliminate resource leaks, making it suitable for use 2075897234Sdrhin long-running programs such as graphical user interfaces 2175897234Sdrhor embedded controllers.</p> 2275897234Sdrh 2375897234Sdrh<p>This document is an introduction to the Lemon 2475897234Sdrhparser generator.</p> 2575897234Sdrh 26c5e56b34Sdrh<h2>Security Note</h2> 27c5e56b34Sdrh 28c5e56b34Sdrh<p>The language parser code created by Lemon is very robust and 29c5e56b34Sdrhis well-suited for use in internet-facing applications that need to 30c5e56b34Sdrhsafely process maliciously crafted inputs. 31c5e56b34Sdrh 32c5e56b34Sdrh<p>The "lemon.exe" command-line tool itself works great when given a valid 33c5e56b34Sdrhinput grammar file and almost always gives helpful 34c5e56b34Sdrherror messages for malformed inputs. However, it is possible for 35c5e56b34Sdrha malicious user to craft a grammar file that will cause 36c5e56b34Sdrhlemon.exe to crash. 37c5e56b34SdrhWe do not see this as a problem, as lemon.exe is not intended to be used 38c5e56b34Sdrhwith hostile inputs. 39c5e56b34SdrhTo summarize:</p> 40c5e56b34Sdrh 41c5e56b34Sdrh<ul> 42c5e56b34Sdrh<li>Parser code generated by lemon → Robust and secure 43c5e56b34Sdrh<li>The "lemon.exe" command line tool itself → Not so much 44c5e56b34Sdrh</ul> 45c5e56b34Sdrh 4675897234Sdrh<h2>Theory of Operation</h2> 4775897234Sdrh 4875897234Sdrh<p>The main goal of Lemon is to translate a context free grammar (CFG) 4975897234Sdrhfor a particular language into C code that implements a parser for 5075897234Sdrhthat language. 5175897234SdrhThe program has two inputs: 5275897234Sdrh<ul> 5375897234Sdrh<li>The grammar specification. 5475897234Sdrh<li>A parser template file. 5575897234Sdrh</ul> 5675897234SdrhTypically, only the grammar specification is supplied by the programmer. 5775897234SdrhLemon comes with a default parser template which works fine for most 5875897234Sdrhapplications. But the user is free to substitute a different parser 5975897234Sdrhtemplate if desired.</p> 6075897234Sdrh 619a243e69Sdrh<p>Depending on command-line options, Lemon will generate up to 629a243e69Sdrhthree output files. 6375897234Sdrh<ul> 6475897234Sdrh<li>C code to implement the parser. 6575897234Sdrh<li>A header file defining an integer ID for each terminal symbol. 6675897234Sdrh<li>An information file that describes the states of the generated parser 6775897234Sdrh automaton. 6875897234Sdrh</ul> 6975897234SdrhBy default, all three of these output files are generated. 709bccde3dSdrhThe header file is suppressed if the "-m" command-line option is 719bccde3dSdrhused and the report file is omitted when "-q" is selected.</p> 7275897234Sdrh 739bccde3dSdrh<p>The grammar specification file uses a ".y" suffix, by convention. 7475897234SdrhIn the examples used in this document, we'll assume the name of the 759bccde3dSdrhgrammar file is "gram.y". A typical use of Lemon would be the 7675897234Sdrhfollowing command: 7775897234Sdrh<pre> 7875897234Sdrh lemon gram.y 7975897234Sdrh</pre> 809bccde3dSdrhThis command will generate three output files named "gram.c", 819bccde3dSdrh"gram.h" and "gram.out". 8275897234SdrhThe first is C code to implement the parser. The second 8375897234Sdrhis the header file that defines numerical values for all 8475897234Sdrhterminal symbols, and the last is the report that explains 8575897234Sdrhthe states used by the parser automaton.</p> 8675897234Sdrh 8775897234Sdrh<h3>Command Line Options</h3> 8875897234Sdrh 8975897234Sdrh<p>The behavior of Lemon can be modified using command-line options. 9075897234SdrhYou can obtain a list of the available command-line options together 9175897234Sdrhwith a brief explanation of what each does by typing 9275897234Sdrh<pre> 939a243e69Sdrh lemon "-?" 9475897234Sdrh</pre> 9575897234SdrhAs of this writing, the following command-line options are supported: 9675897234Sdrh<ul> 979bccde3dSdrh<li><b>-b</b> 989bccde3dSdrhShow only the basis for each parser state in the report file. 999bccde3dSdrh<li><b>-c</b> 1009a243e69SdrhDo not compress the generated action tables. The parser will be a 1019a243e69Sdrhlittle larger and slower, but it will detect syntax errors sooner. 102fb32c44eSdrh<li><b>-d</b><i>directory</i> 103fb32c44eSdrhWrite all output files into <i>directory</i>. Normally, output files 104fb32c44eSdrhare written into the directory that contains the input grammar file. 1059bccde3dSdrh<li><b>-D<i>name</i></b> 1069a243e69SdrhDefine C preprocessor macro <i>name</i>. This macro is usable by 1079a243e69Sdrh"<tt><a href='#pifdef'>%ifdef</a></tt>" and 1089a243e69Sdrh"<tt><a href='#pifdef'>%ifndef</a></tt>" lines 1099a243e69Sdrhin the grammar file. 1109bccde3dSdrh<li><b>-g</b> 1119bccde3dSdrhDo not generate a parser. Instead write the input grammar to standard 1129bccde3dSdrhoutput with all comments, actions, and other extraneous text removed. 1139bccde3dSdrh<li><b>-l</b> 114dfe4e6bbSdrhOmit "#line" directives in the generated parser C code. 1159bccde3dSdrh<li><b>-m</b> 1169bccde3dSdrhCause the output C source code to be compatible with the "makeheaders" 1179bccde3dSdrhprogram. 1189bccde3dSdrh<li><b>-p</b> 1199bccde3dSdrhDisplay all conflicts that are resolved by 1209bccde3dSdrh<a href='#precrules'>precedence rules</a>. 1219bccde3dSdrh<li><b>-q</b> 1229bccde3dSdrhSuppress generation of the report file. 1239bccde3dSdrh<li><b>-r</b> 1249bccde3dSdrhDo not sort or renumber the parser states as part of optimization. 1259bccde3dSdrh<li><b>-s</b> 1269bccde3dSdrhShow parser statistics before existing. 1279bccde3dSdrh<li><b>-T<i>file</i></b> 1289bccde3dSdrhUse <i>file</i> as the template for the generated C-code parser implementation. 1299bccde3dSdrh<li><b>-x</b> 1309bccde3dSdrhPrint the Lemon version number. 13175897234Sdrh</ul> 13275897234Sdrh 13375897234Sdrh<h3>The Parser Interface</h3> 13475897234Sdrh 13575897234Sdrh<p>Lemon doesn't generate a complete, working program. It only generates 13675897234Sdrha few subroutines that implement a parser. This section describes 13775897234Sdrhthe interface to those subroutines. It is up to the programmer to 13875897234Sdrhcall these subroutines in an appropriate way in order to produce a 13975897234Sdrhcomplete system.</p> 14075897234Sdrh 14175897234Sdrh<p>Before a program begins using a Lemon-generated parser, the program 14275897234Sdrhmust first create the parser. 14375897234SdrhA new parser is created as follows: 14475897234Sdrh<pre> 14575897234Sdrh void *pParser = ParseAlloc( malloc ); 14675897234Sdrh</pre> 14775897234SdrhThe ParseAlloc() routine allocates and initializes a new parser and 14875897234Sdrhreturns a pointer to it. 1499bccde3dSdrhThe actual data structure used to represent a parser is opaque — 15075897234Sdrhits internal structure is not visible or usable by the calling routine. 15175897234SdrhFor this reason, the ParseAlloc() routine returns a pointer to void 15275897234Sdrhrather than a pointer to some particular structure. 15375897234SdrhThe sole argument to the ParseAlloc() routine is a pointer to the 1549bccde3dSdrhsubroutine used to allocate memory. Typically this means malloc().</p> 15575897234Sdrh 15675897234Sdrh<p>After a program is finished using a parser, it can reclaim all 15775897234Sdrhmemory allocated by that parser by calling 15875897234Sdrh<pre> 15975897234Sdrh ParseFree(pParser, free); 16075897234Sdrh</pre> 16175897234SdrhThe first argument is the same pointer returned by ParseAlloc(). The 16275897234Sdrhsecond argument is a pointer to the function used to release bulk 16375897234Sdrhmemory back to the system.</p> 16475897234Sdrh 16575897234Sdrh<p>After a parser has been allocated using ParseAlloc(), the programmer 16675897234Sdrhmust supply the parser with a sequence of tokens (terminal symbols) to 16775897234Sdrhbe parsed. This is accomplished by calling the following function 16875897234Sdrhonce for each token: 16975897234Sdrh<pre> 17075897234Sdrh Parse(pParser, hTokenID, sTokenData, pArg); 17175897234Sdrh</pre> 17275897234SdrhThe first argument to the Parse() routine is the pointer returned by 17375897234SdrhParseAlloc(). 1749a243e69SdrhThe second argument is a small positive integer that tells the parser the 17575897234Sdrhtype of the next token in the data stream. 17675897234SdrhThere is one token type for each terminal symbol in the grammar. 17775897234SdrhThe gram.h file generated by Lemon contains #define statements that 17875897234Sdrhmap symbolic terminal symbol names into appropriate integer values. 1799bccde3dSdrhA value of 0 for the second argument is a special flag to the 1809bccde3dSdrhparser to indicate that the end of input has been reached. 18175897234SdrhThe third argument is the value of the given token. By default, 1829a243e69Sdrhthe type of the third argument is "void*", but the grammar will 18375897234Sdrhusually redefine this type to be some kind of structure. 18475897234SdrhTypically the second argument will be a broad category of tokens 1859bccde3dSdrhsuch as "identifier" or "number" and the third argument will 18675897234Sdrhbe the name of the identifier or the value of the number.</p> 18775897234Sdrh 18875897234Sdrh<p>The Parse() function may have either three or four arguments, 18945f31be8Sdrhdepending on the grammar. If the grammar specification file requests 1909a243e69Sdrhit (via the <tt><a href='#extraarg'>%extra_argument</a></tt> directive), 19145f31be8Sdrhthe Parse() function will have a fourth parameter that can be 19275897234Sdrhof any type chosen by the programmer. The parser doesn't do anything 19375897234Sdrhwith this argument except to pass it through to action routines. 19475897234SdrhThis is a convenient mechanism for passing state information down 19575897234Sdrhto the action routines without having to use global variables.</p> 19675897234Sdrh 19775897234Sdrh<p>A typical use of a Lemon parser might look something like the 19875897234Sdrhfollowing: 19975897234Sdrh<pre> 2009a243e69Sdrh 1 ParseTree *ParseFile(const char *zFilename){ 2019a243e69Sdrh 2 Tokenizer *pTokenizer; 2029a243e69Sdrh 3 void *pParser; 2039a243e69Sdrh 4 Token sToken; 2049a243e69Sdrh 5 int hTokenId; 2059a243e69Sdrh 6 ParserState sState; 2069a243e69Sdrh 7 2079a243e69Sdrh 8 pTokenizer = TokenizerCreate(zFilename); 2089a243e69Sdrh 9 pParser = ParseAlloc( malloc ); 2099a243e69Sdrh 10 InitParserState(&sState); 2109a243e69Sdrh 11 while( GetNextToken(pTokenizer, &hTokenId, &sToken) ){ 2119a243e69Sdrh 12 Parse(pParser, hTokenId, sToken, &sState); 21275897234Sdrh 13 } 2139a243e69Sdrh 14 Parse(pParser, 0, sToken, &sState); 21475897234Sdrh 15 ParseFree(pParser, free ); 21575897234Sdrh 16 TokenizerFree(pTokenizer); 21675897234Sdrh 17 return sState.treeRoot; 21775897234Sdrh 18 } 21875897234Sdrh</pre> 21975897234SdrhThis example shows a user-written routine that parses a file of 22075897234Sdrhtext and returns a pointer to the parse tree. 2219bccde3dSdrh(All error-handling code is omitted from this example to keep it 22275897234Sdrhsimple.) 22375897234SdrhWe assume the existence of some kind of tokenizer which is created 22475897234Sdrhusing TokenizerCreate() on line 8 and deleted by TokenizerFree() 22575897234Sdrhon line 16. The GetNextToken() function on line 11 retrieves the 22675897234Sdrhnext token from the input file and puts its type in the 22775897234Sdrhinteger variable hTokenId. The sToken variable is assumed to be 22875897234Sdrhsome kind of structure that contains details about each token, 22975897234Sdrhsuch as its complete text, what line it occurs on, etc.</p> 23075897234Sdrh 23175897234Sdrh<p>This example also assumes the existence of structure of type 23275897234SdrhParserState that holds state information about a particular parse. 23375897234SdrhAn instance of such a structure is created on line 6 and initialized 23475897234Sdrhon line 10. A pointer to this structure is passed into the Parse() 23575897234Sdrhroutine as the optional 4th argument. 23675897234SdrhThe action routine specified by the grammar for the parser can use 23775897234Sdrhthe ParserState structure to hold whatever information is useful and 23875897234Sdrhappropriate. In the example, we note that the treeRoot field of 23975897234Sdrhthe ParserState structure is left pointing to the root of the parse 24075897234Sdrhtree.</p> 24175897234Sdrh 24275897234Sdrh<p>The core of this example as it relates to Lemon is as follows: 24375897234Sdrh<pre> 24475897234Sdrh ParseFile(){ 24575897234Sdrh pParser = ParseAlloc( malloc ); 2469a243e69Sdrh while( GetNextToken(pTokenizer,&hTokenId, &sToken) ){ 24775897234Sdrh Parse(pParser, hTokenId, sToken); 24875897234Sdrh } 24975897234Sdrh Parse(pParser, 0, sToken); 25075897234Sdrh ParseFree(pParser, free ); 25175897234Sdrh } 25275897234Sdrh</pre> 25375897234SdrhBasically, what a program has to do to use a Lemon-generated parser 25475897234Sdrhis first create the parser, then send it lots of tokens obtained by 25575897234Sdrhtokenizing an input source. When the end of input is reached, the 25675897234SdrhParse() routine should be called one last time with a token type 25775897234Sdrhof 0. This step is necessary to inform the parser that the end of 25875897234Sdrhinput has been reached. Finally, we reclaim memory used by the 25975897234Sdrhparser by calling ParseFree().</p> 26075897234Sdrh 26175897234Sdrh<p>There is one other interface routine that should be mentioned 26275897234Sdrhbefore we move on. 26375897234SdrhThe ParseTrace() function can be used to generate debugging output 26475897234Sdrhfrom the parser. A prototype for this routine is as follows: 26575897234Sdrh<pre> 26675897234Sdrh ParseTrace(FILE *stream, char *zPrefix); 26775897234Sdrh</pre> 26875897234SdrhAfter this routine is called, a short (one-line) message is written 26975897234Sdrhto the designated output stream every time the parser changes states 27075897234Sdrhor calls an action routine. Each such message is prefaced using 27175897234Sdrhthe text given by zPrefix. This debugging output can be turned off 27275897234Sdrhby calling ParseTrace() again with a first argument of NULL (0).</p> 27375897234Sdrh 27475897234Sdrh<h3>Differences With YACC and BISON</h3> 27575897234Sdrh 27675897234Sdrh<p>Programmers who have previously used the yacc or bison parser 27775897234Sdrhgenerator will notice several important differences between yacc and/or 27875897234Sdrhbison and Lemon. 27975897234Sdrh<ul> 28075897234Sdrh<li>In yacc and bison, the parser calls the tokenizer. In Lemon, 28175897234Sdrh the tokenizer calls the parser. 28275897234Sdrh<li>Lemon uses no global variables. Yacc and bison use global variables 28375897234Sdrh to pass information between the tokenizer and parser. 28475897234Sdrh<li>Lemon allows multiple parsers to be running simultaneously. Yacc 28575897234Sdrh and bison do not. 28675897234Sdrh</ul> 28775897234SdrhThese differences may cause some initial confusion for programmers 28875897234Sdrhwith prior yacc and bison experience. 28975897234SdrhBut after years of experience using Lemon, I firmly 29075897234Sdrhbelieve that the Lemon way of doing things is better.</p> 29175897234Sdrh 29245f31be8Sdrh<p><i>Updated as of 2016-02-16:</i> 29345f31be8SdrhThe text above was written in the 1990s. 29445f31be8SdrhWe are told that Bison has lately been enhanced to support the 29545f31be8Sdrhtokenizer-calls-parser paradigm used by Lemon, and to obviate the 29645f31be8Sdrhneed for global variables.</p> 29745f31be8Sdrh 29875897234Sdrh<h2>Input File Syntax</h2> 29975897234Sdrh 30075897234Sdrh<p>The main purpose of the grammar specification file for Lemon is 30175897234Sdrhto define the grammar for the parser. But the input file also 30275897234Sdrhspecifies additional information Lemon requires to do its job. 30375897234SdrhMost of the work in using Lemon is in writing an appropriate 30475897234Sdrhgrammar file.</p> 30575897234Sdrh 3069a243e69Sdrh<p>The grammar file for Lemon is, for the most part, free format. 30775897234SdrhIt does not have sections or divisions like yacc or bison. Any 30875897234Sdrhdeclaration can occur at any point in the file. 30975897234SdrhLemon ignores whitespace (except where it is needed to separate 3109a243e69Sdrhtokens), and it honors the same commenting conventions as C and C++.</p> 31175897234Sdrh 31275897234Sdrh<h3>Terminals and Nonterminals</h3> 31375897234Sdrh 31475897234Sdrh<p>A terminal symbol (token) is any string of alphanumeric 3159bccde3dSdrhand/or underscore characters 31675897234Sdrhthat begins with an uppercase letter. 317c8eee5e5SdrhA terminal can contain lowercase letters after the first character, 31875897234Sdrhbut the usual convention is to make terminals all uppercase. 31975897234SdrhA nonterminal, on the other hand, is any string of alphanumeric 32075897234Sdrhand underscore characters than begins with a lowercase letter. 3219a243e69SdrhAgain, the usual convention is to make nonterminals use all lowercase 3229a243e69Sdrhletters.</p> 32375897234Sdrh 32475897234Sdrh<p>In Lemon, terminal and nonterminal symbols do not need to 32575897234Sdrhbe declared or identified in a separate section of the grammar file. 32675897234SdrhLemon is able to generate a list of all terminals and nonterminals 32775897234Sdrhby examining the grammar rules, and it can always distinguish a 32875897234Sdrhterminal from a nonterminal by checking the case of the first 32975897234Sdrhcharacter of the name.</p> 33075897234Sdrh 33175897234Sdrh<p>Yacc and bison allow terminal symbols to have either alphanumeric 33275897234Sdrhnames or to be individual characters included in single quotes, like 33375897234Sdrhthis: ')' or '$'. Lemon does not allow this alternative form for 33475897234Sdrhterminal symbols. With Lemon, all symbols, terminals and nonterminals, 33575897234Sdrhmust have alphanumeric names.</p> 33675897234Sdrh 33775897234Sdrh<h3>Grammar Rules</h3> 33875897234Sdrh 33975897234Sdrh<p>The main component of a Lemon grammar file is a sequence of grammar 34075897234Sdrhrules. 34175897234SdrhEach grammar rule consists of a nonterminal symbol followed by 3429bccde3dSdrhthe special symbol "::=" and then a list of terminals and/or nonterminals. 34375897234SdrhThe rule is terminated by a period. 34475897234SdrhThe list of terminals and nonterminals on the right-hand side of the 34575897234Sdrhrule can be empty. 34675897234SdrhRules can occur in any order, except that the left-hand side of the 34775897234Sdrhfirst rule is assumed to be the start symbol for the grammar (unless 3489a243e69Sdrhspecified otherwise using the <tt><a href='#start_symbol'>%start_symbol</a></tt> 3499a243e69Sdrhdirective described below.) 35075897234SdrhA typical sequence of grammar rules might look something like this: 35175897234Sdrh<pre> 35275897234Sdrh expr ::= expr PLUS expr. 35375897234Sdrh expr ::= expr TIMES expr. 35475897234Sdrh expr ::= LPAREN expr RPAREN. 35575897234Sdrh expr ::= VALUE. 35675897234Sdrh</pre> 35775897234Sdrh</p> 35875897234Sdrh 3599bccde3dSdrh<p>There is one non-terminal in this example, "expr", and five 3609bccde3dSdrhterminal symbols or tokens: "PLUS", "TIMES", "LPAREN", 3619bccde3dSdrh"RPAREN" and "VALUE".</p> 36275897234Sdrh 36375897234Sdrh<p>Like yacc and bison, Lemon allows the grammar to specify a block 36475897234Sdrhof C code that will be executed whenever a grammar rule is reduced 36575897234Sdrhby the parser. 36675897234SdrhIn Lemon, this action is specified by putting the C code (contained 36775897234Sdrhwithin curly braces <tt>{...}</tt>) immediately after the 36875897234Sdrhperiod that closes the rule. 36975897234SdrhFor example: 37075897234Sdrh<pre> 37175897234Sdrh expr ::= expr PLUS expr. { printf("Doing an addition...\n"); } 37275897234Sdrh</pre> 37375897234Sdrh</p> 37475897234Sdrh 37575897234Sdrh<p>In order to be useful, grammar actions must normally be linked to 37675897234Sdrhtheir associated grammar rules. 3779bccde3dSdrhIn yacc and bison, this is accomplished by embedding a "$$" in the 37875897234Sdrhaction to stand for the value of the left-hand side of the rule and 3799bccde3dSdrhsymbols "$1", "$2", and so forth to stand for the value of 38075897234Sdrhthe terminal or nonterminal at position 1, 2 and so forth on the 38175897234Sdrhright-hand side of the rule. 38275897234SdrhThis idea is very powerful, but it is also very error-prone. The 38375897234Sdrhsingle most common source of errors in a yacc or bison grammar is 38475897234Sdrhto miscount the number of symbols on the right-hand side of a grammar 3859bccde3dSdrhrule and say "$7" when you really mean "$8".</p> 38675897234Sdrh 38775897234Sdrh<p>Lemon avoids the need to count grammar symbols by assigning symbolic 38875897234Sdrhnames to each symbol in a grammar rule and then using those symbolic 38975897234Sdrhnames in the action. 39075897234SdrhIn yacc or bison, one would write this: 39175897234Sdrh<pre> 3929a243e69Sdrh expr -> expr PLUS expr { $$ = $1 + $3; }; 39375897234Sdrh</pre> 39475897234SdrhBut in Lemon, the same rule becomes the following: 39575897234Sdrh<pre> 39675897234Sdrh expr(A) ::= expr(B) PLUS expr(C). { A = B+C; } 39775897234Sdrh</pre> 39875897234SdrhIn the Lemon rule, any symbol in parentheses after a grammar rule 39975897234Sdrhsymbol becomes a place holder for that symbol in the grammar rule. 40075897234SdrhThis place holder can then be used in the associated C action to 40175897234Sdrhstand for the value of that symbol.<p> 40275897234Sdrh 40375897234Sdrh<p>The Lemon notation for linking a grammar rule with its reduce 40475897234Sdrhaction is superior to yacc/bison on several counts. 40575897234SdrhFirst, as mentioned above, the Lemon method avoids the need to 40675897234Sdrhcount grammar symbols. 40775897234SdrhSecondly, if a terminal or nonterminal in a Lemon grammar rule 40875897234Sdrhincludes a linking symbol in parentheses but that linking symbol 40975897234Sdrhis not actually used in the reduce action, then an error message 41075897234Sdrhis generated. 41175897234SdrhFor example, the rule 41275897234Sdrh<pre> 41375897234Sdrh expr(A) ::= expr(B) PLUS expr(C). { A = B; } 41475897234Sdrh</pre> 4159bccde3dSdrhwill generate an error because the linking symbol "C" is used 41675897234Sdrhin the grammar rule but not in the reduce action.</p> 41775897234Sdrh 41875897234Sdrh<p>The Lemon notation for linking grammar rules to reduce actions 41975897234Sdrhalso facilitates the use of destructors for reclaiming memory 42075897234Sdrhallocated by the values of terminals and nonterminals on the 42175897234Sdrhright-hand side of a rule.</p> 42275897234Sdrh 4239bccde3dSdrh<a name='precrules'></a> 42475897234Sdrh<h3>Precedence Rules</h3> 42575897234Sdrh 42675897234Sdrh<p>Lemon resolves parsing ambiguities in exactly the same way as 42775897234Sdrhyacc and bison. A shift-reduce conflict is resolved in favor 42875897234Sdrhof the shift, and a reduce-reduce conflict is resolved by reducing 42975897234Sdrhwhichever rule comes first in the grammar file.</p> 43075897234Sdrh 43175897234Sdrh<p>Just like in 43275897234Sdrhyacc and bison, Lemon allows a measure of control 4339a243e69Sdrhover the resolution of parsing conflicts using precedence rules. 43475897234SdrhA precedence value can be assigned to any terminal symbol 4359bccde3dSdrhusing the 4369a243e69Sdrh<tt><a href='#pleft'>%left</a></tt>, 4379a243e69Sdrh<tt><a href='#pright'>%right</a></tt> or 4389a243e69Sdrh<tt><a href='#pnonassoc'>%nonassoc</a></tt> directives. Terminal symbols 4399a243e69Sdrhmentioned in earlier directives have a lower precedence than 44075897234Sdrhterminal symbols mentioned in later directives. For example:</p> 44175897234Sdrh 44275897234Sdrh<p><pre> 44375897234Sdrh %left AND. 44475897234Sdrh %left OR. 44575897234Sdrh %nonassoc EQ NE GT GE LT LE. 44675897234Sdrh %left PLUS MINUS. 44775897234Sdrh %left TIMES DIVIDE MOD. 44875897234Sdrh %right EXP NOT. 44975897234Sdrh</pre></p> 45075897234Sdrh 45175897234Sdrh<p>In the preceding sequence of directives, the AND operator is 45275897234Sdrhdefined to have the lowest precedence. The OR operator is one 45375897234Sdrhprecedence level higher. And so forth. Hence, the grammar would 45475897234Sdrhattempt to group the ambiguous expression 45575897234Sdrh<pre> 45675897234Sdrh a AND b OR c 45775897234Sdrh</pre> 45875897234Sdrhlike this 45975897234Sdrh<pre> 46075897234Sdrh a AND (b OR c). 46175897234Sdrh</pre> 46275897234SdrhThe associativity (left, right or nonassoc) is used to determine 46375897234Sdrhthe grouping when the precedence is the same. AND is left-associative 46475897234Sdrhin our example, so 46575897234Sdrh<pre> 46675897234Sdrh a AND b AND c 46775897234Sdrh</pre> 46875897234Sdrhis parsed like this 46975897234Sdrh<pre> 47075897234Sdrh (a AND b) AND c. 47175897234Sdrh</pre> 47275897234SdrhThe EXP operator is right-associative, though, so 47375897234Sdrh<pre> 47475897234Sdrh a EXP b EXP c 47575897234Sdrh</pre> 47675897234Sdrhis parsed like this 47775897234Sdrh<pre> 47875897234Sdrh a EXP (b EXP c). 47975897234Sdrh</pre> 48075897234SdrhThe nonassoc precedence is used for non-associative operators. 48175897234SdrhSo 48275897234Sdrh<pre> 48375897234Sdrh a EQ b EQ c 48475897234Sdrh</pre> 48575897234Sdrhis an error.</p> 48675897234Sdrh 48775897234Sdrh<p>The precedence of non-terminals is transferred to rules as follows: 48875897234SdrhThe precedence of a grammar rule is equal to the precedence of the 48975897234Sdrhleft-most terminal symbol in the rule for which a precedence is 49075897234Sdrhdefined. This is normally what you want, but in those cases where 49175897234Sdrhyou want to precedence of a grammar rule to be something different, 49275897234Sdrhyou can specify an alternative precedence symbol by putting the 49375897234Sdrhsymbol in square braces after the period at the end of the rule and 49475897234Sdrhbefore any C-code. For example:</p> 49575897234Sdrh 49675897234Sdrh<p><pre> 49775897234Sdrh expr = MINUS expr. [NOT] 49875897234Sdrh</pre></p> 49975897234Sdrh 50075897234Sdrh<p>This rule has a precedence equal to that of the NOT symbol, not the 50175897234SdrhMINUS symbol as would have been the case by default.</p> 50275897234Sdrh 50375897234Sdrh<p>With the knowledge of how precedence is assigned to terminal 50475897234Sdrhsymbols and individual 50575897234Sdrhgrammar rules, we can now explain precisely how parsing conflicts 50675897234Sdrhare resolved in Lemon. Shift-reduce conflicts are resolved 50775897234Sdrhas follows: 50875897234Sdrh<ul> 50975897234Sdrh<li> If either the token to be shifted or the rule to be reduced 51075897234Sdrh lacks precedence information, then resolve in favor of the 51175897234Sdrh shift, but report a parsing conflict. 51275897234Sdrh<li> If the precedence of the token to be shifted is greater than 51375897234Sdrh the precedence of the rule to reduce, then resolve in favor 51475897234Sdrh of the shift. No parsing conflict is reported. 5159a243e69Sdrh<li> If the precedence of the token to be shifted is less than the 51675897234Sdrh precedence of the rule to reduce, then resolve in favor of the 51775897234Sdrh reduce action. No parsing conflict is reported. 51875897234Sdrh<li> If the precedences are the same and the shift token is 51975897234Sdrh right-associative, then resolve in favor of the shift. 52075897234Sdrh No parsing conflict is reported. 5219a243e69Sdrh<li> If the precedences are the same and the shift token is 52275897234Sdrh left-associative, then resolve in favor of the reduce. 52375897234Sdrh No parsing conflict is reported. 5249a243e69Sdrh<li> Otherwise, resolve the conflict by doing the shift, and 5259a243e69Sdrh report a parsing conflict. 52675897234Sdrh</ul> 52775897234SdrhReduce-reduce conflicts are resolved this way: 52875897234Sdrh<ul> 52975897234Sdrh<li> If either reduce rule 53075897234Sdrh lacks precedence information, then resolve in favor of the 5319a243e69Sdrh rule that appears first in the grammar, and report a parsing 53275897234Sdrh conflict. 5339a243e69Sdrh<li> If both rules have precedence and the precedence is different, 53475897234Sdrh then resolve the dispute in favor of the rule with the highest 5359a243e69Sdrh precedence, and do not report a conflict. 53675897234Sdrh<li> Otherwise, resolve the conflict by reducing by the rule that 5379a243e69Sdrh appears first in the grammar, and report a parsing conflict. 53875897234Sdrh</ul> 53975897234Sdrh 54075897234Sdrh<h3>Special Directives</h3> 54175897234Sdrh 54275897234Sdrh<p>The input grammar to Lemon consists of grammar rules and special 54375897234Sdrhdirectives. We've described all the grammar rules, so now we'll 54475897234Sdrhtalk about the special directives.</p> 54575897234Sdrh 5469a243e69Sdrh<p>Directives in Lemon can occur in any order. You can put them before 5479a243e69Sdrhthe grammar rules, or after the grammar rules, or in the midst of the 54875897234Sdrhgrammar rules. It doesn't matter. The relative order of 54975897234Sdrhdirectives used to assign precedence to terminals is important, but 55075897234Sdrhother than that, the order of directives in Lemon is arbitrary.</p> 55175897234Sdrh 55275897234Sdrh<p>Lemon supports the following special directives: 55375897234Sdrh<ul> 5549a243e69Sdrh<li><tt><a href='#pcode'>%code</a></tt> 5559a243e69Sdrh<li><tt><a href='#default_destructor'>%default_destructor</a></tt> 5569a243e69Sdrh<li><tt><a href='#default_type'>%default_type</a></tt> 5579a243e69Sdrh<li><tt><a href='#destructor'>%destructor</a></tt> 5589a243e69Sdrh<li><tt><a href='#pifdef'>%endif</a></tt> 5599a243e69Sdrh<li><tt><a href='#extraarg'>%extra_argument</a></tt> 5609a243e69Sdrh<li><tt><a href='#pfallback'>%fallback</a></tt> 5619a243e69Sdrh<li><tt><a href='#pifdef'>%ifdef</a></tt> 5629a243e69Sdrh<li><tt><a href='#pifdef'>%ifndef</a></tt> 5639a243e69Sdrh<li><tt><a href='#pinclude'>%include</a></tt> 5649a243e69Sdrh<li><tt><a href='#pleft'>%left</a></tt> 5659a243e69Sdrh<li><tt><a href='#pname'>%name</a></tt> 5669a243e69Sdrh<li><tt><a href='#pnonassoc'>%nonassoc</a></tt> 5679a243e69Sdrh<li><tt><a href='#parse_accept'>%parse_accept</a></tt> 5689a243e69Sdrh<li><tt><a href='#parse_failure'>%parse_failure</a></tt> 5699a243e69Sdrh<li><tt><a href='#pright'>%right</a></tt> 5709a243e69Sdrh<li><tt><a href='#stack_overflow'>%stack_overflow</a></tt> 5719a243e69Sdrh<li><tt><a href='#stack_size'>%stack_size</a></tt> 5729a243e69Sdrh<li><tt><a href='#start_symbol'>%start_symbol</a></tt> 5739a243e69Sdrh<li><tt><a href='#syntax_error'>%syntax_error</a></tt> 5749a243e69Sdrh<li><tt><a href='#token_class'>%token_class</a></tt> 5759a243e69Sdrh<li><tt><a href='#token_destructor'>%token_destructor</a></tt> 5769a243e69Sdrh<li><tt><a href='#token_prefix'>%token_prefix</a></tt> 5779a243e69Sdrh<li><tt><a href='#token_type'>%token_type</a></tt> 5789a243e69Sdrh<li><tt><a href='#ptype'>%type</a></tt> 5799a243e69Sdrh<li><tt><a href='#pwildcard'>%wildcard</a></tt> 58075897234Sdrh</ul> 58175897234SdrhEach of these directives will be described separately in the 58275897234Sdrhfollowing sections:</p> 58375897234Sdrh 5849bccde3dSdrh<a name='pcode'></a> 585f2340fc7Sdrh<h4>The <tt>%code</tt> directive</h4> 586f2340fc7Sdrh 5879a243e69Sdrh<p>The <tt>%code</tt> directive is used to specify additional C code that 588f2340fc7Sdrhis added to the end of the main output file. This is similar to 5899a243e69Sdrhthe <tt><a href='#pinclude'>%include</a></tt> directive except that 5909a243e69Sdrh<tt>%include</tt> is inserted at the beginning of the main output file.</p> 591f2340fc7Sdrh 5929a243e69Sdrh<p><tt>%code</tt> is typically used to include some action routines or perhaps 5939bccde3dSdrha tokenizer or even the "main()" function 5949bccde3dSdrhas part of the output file.</p> 595f2340fc7Sdrh 5969bccde3dSdrh<a name='default_destructor'></a> 597f2340fc7Sdrh<h4>The <tt>%default_destructor</tt> directive</h4> 598f2340fc7Sdrh 5999a243e69Sdrh<p>The <tt>%default_destructor</tt> directive specifies a destructor to 600f2340fc7Sdrhuse for non-terminals that do not have their own destructor 6019a243e69Sdrhspecified by a separate <tt>%destructor</tt> directive. See the documentation 6029a243e69Sdrhon the <tt><a name='#destructor'>%destructor</a></tt> directive below for 6039bccde3dSdrhadditional information.</p> 604f2340fc7Sdrh 6059a243e69Sdrh<p>In some grammars, many different non-terminal symbols have the 606f2340fc7Sdrhsame data type and hence the same destructor. This directive is 6079a243e69Sdrha convenient way to specify the same destructor for all those 608f2340fc7Sdrhnon-terminals using a single statement.</p> 609f2340fc7Sdrh 6109bccde3dSdrh<a name='default_type'></a> 611f2340fc7Sdrh<h4>The <tt>%default_type</tt> directive</h4> 612f2340fc7Sdrh 6139a243e69Sdrh<p>The <tt>%default_type</tt> directive specifies the data type of non-terminal 6149a243e69Sdrhsymbols that do not have their own data type defined using a separate 6159a243e69Sdrh<tt><a href='#ptype'>%type</a></tt> directive.</p> 616f2340fc7Sdrh 6179bccde3dSdrh<a name='destructor'></a> 61875897234Sdrh<h4>The <tt>%destructor</tt> directive</h4> 61975897234Sdrh 6209a243e69Sdrh<p>The <tt>%destructor</tt> directive is used to specify a destructor for 62175897234Sdrha non-terminal symbol. 6229a243e69Sdrh(See also the <tt><a href='#token_destructor'>%token_destructor</a></tt> 6239bccde3dSdrhdirective which is used to specify a destructor for terminal symbols.)</p> 62475897234Sdrh 62575897234Sdrh<p>A non-terminal's destructor is called to dispose of the 62675897234Sdrhnon-terminal's value whenever the non-terminal is popped from 62775897234Sdrhthe stack. This includes all of the following circumstances: 62875897234Sdrh<ul> 62975897234Sdrh<li> When a rule reduces and the value of a non-terminal on 63075897234Sdrh the right-hand side is not linked to C code. 63175897234Sdrh<li> When the stack is popped during error processing. 63275897234Sdrh<li> When the ParseFree() function runs. 63375897234Sdrh</ul> 63475897234SdrhThe destructor can do whatever it wants with the value of 63575897234Sdrhthe non-terminal, but its design is to deallocate memory 63675897234Sdrhor other resources held by that non-terminal.</p> 63775897234Sdrh 63875897234Sdrh<p>Consider an example: 63975897234Sdrh<pre> 64075897234Sdrh %type nt {void*} 64175897234Sdrh %destructor nt { free($$); } 64275897234Sdrh nt(A) ::= ID NUM. { A = malloc( 100 ); } 64375897234Sdrh</pre> 6449a243e69SdrhThis example is a bit contrived, but it serves to illustrate how 64575897234Sdrhdestructors work. The example shows a non-terminal named 6469bccde3dSdrh"nt" that holds values of type "void*". When the rule for 6479bccde3dSdrhan "nt" reduces, it sets the value of the non-terminal to 64875897234Sdrhspace obtained from malloc(). Later, when the nt non-terminal 64975897234Sdrhis popped from the stack, the destructor will fire and call 65075897234Sdrhfree() on this malloced space, thus avoiding a memory leak. 6519bccde3dSdrh(Note that the symbol "$$" in the destructor code is replaced 65275897234Sdrhby the value of the non-terminal.)</p> 65375897234Sdrh 65475897234Sdrh<p>It is important to note that the value of a non-terminal is passed 65575897234Sdrhto the destructor whenever the non-terminal is removed from the 65675897234Sdrhstack, unless the non-terminal is used in a C-code action. If 65775897234Sdrhthe non-terminal is used by C-code, then it is assumed that the 6589bccde3dSdrhC-code will take care of destroying it. 6599bccde3dSdrhMore commonly, the value is used to build some 6609a243e69Sdrhlarger structure, and we don't want to destroy it, which is why 66175897234Sdrhthe destructor is not called in this circumstance.</p> 66275897234Sdrh 6639bccde3dSdrh<p>Destructors help avoid memory leaks by automatically freeing 6649bccde3dSdrhallocated objects when they go out of scope. 66575897234SdrhTo do the same using yacc or bison is much more difficult.</p> 66675897234Sdrh 6679a243e69Sdrh<a name='extraarg'></a> 66875897234Sdrh<h4>The <tt>%extra_argument</tt> directive</h4> 66975897234Sdrh 6709a243e69SdrhThe <tt>%extra_argument</tt> directive instructs Lemon to add a 4th parameter 67175897234Sdrhto the parameter list of the Parse() function it generates. Lemon 67275897234Sdrhdoesn't do anything itself with this extra argument, but it does 67375897234Sdrhmake the argument available to C-code action routines, destructors, 67475897234Sdrhand so forth. For example, if the grammar file contains:</p> 67575897234Sdrh 67675897234Sdrh<p><pre> 67775897234Sdrh %extra_argument { MyStruct *pAbc } 67875897234Sdrh</pre></p> 67975897234Sdrh 68075897234Sdrh<p>Then the Parse() function generated will have an 4th parameter 6819bccde3dSdrhof type "MyStruct*" and all action routines will have access to 6829bccde3dSdrha variable named "pAbc" that is the value of the 4th parameter 68375897234Sdrhin the most recent call to Parse().</p> 68475897234Sdrh 685fb32c44eSdrh<p>The <tt>%extra_context</tt> directive works the same except that it 686fb32c44eSdrhis passed in on the ParseAlloc() or ParseInit() routines instead of 687fb32c44eSdrhon Parse(). 688fb32c44eSdrh 689fb32c44eSdrh<a name='extractx'></a> 690fb32c44eSdrh<h4>The <tt>%extra_context</tt> directive</h4> 691fb32c44eSdrh 692fb32c44eSdrhThe <tt>%extra_context</tt> directive instructs Lemon to add a 2th parameter 693fb32c44eSdrhto the parameter list of the ParseAlloc() and ParseInif() functions. Lemon 694fb32c44eSdrhdoesn't do anything itself with these extra argument, but it does 695fb32c44eSdrhstore the value make it available to C-code action routines, destructors, 696fb32c44eSdrhand so forth. For example, if the grammar file contains:</p> 697fb32c44eSdrh 698fb32c44eSdrh<p><pre> 699fb32c44eSdrh %extra_context { MyStruct *pAbc } 700fb32c44eSdrh</pre></p> 701fb32c44eSdrh 702fb32c44eSdrh<p>Then the ParseAlloc() and ParseInit() functions will have an 2th parameter 703fb32c44eSdrhof type "MyStruct*" and all action routines will have access to 704fb32c44eSdrha variable named "pAbc" that is the value of that 2th parameter.</p> 705fb32c44eSdrh 706fb32c44eSdrh<p>The <tt>%extra_argument</tt> directive works the same except that it 707fb32c44eSdrhis passed in on the Parse() routine instead of on ParseAlloc()/ParseInit(). 708fb32c44eSdrh 7099bccde3dSdrh<a name='pfallback'></a> 7109bccde3dSdrh<h4>The <tt>%fallback</tt> directive</h4> 7119bccde3dSdrh 7129a243e69Sdrh<p>The <tt>%fallback</tt> directive specifies an alternative meaning for one 7139bccde3dSdrhor more tokens. The alternative meaning is tried if the original token 7149a243e69Sdrhwould have generated a syntax error.</p> 7159bccde3dSdrh 7169a243e69Sdrh<p>The <tt>%fallback</tt> directive was added to support robust parsing of SQL 7179a243e69Sdrhsyntax in <a href='https://www.sqlite.org/'>SQLite</a>. 7189bccde3dSdrhThe SQL language contains a large assortment of keywords, each of which 7199bccde3dSdrhappears as a different token to the language parser. SQL contains so 7209a243e69Sdrhmany keywords that it can be difficult for programmers to keep up with 7219bccde3dSdrhthem all. Programmers will, therefore, sometimes mistakenly use an 7229a243e69Sdrhobscure language keyword for an identifier. The <tt>%fallback</tt> directive 7239bccde3dSdrhprovides a mechanism to tell the parser: "If you are unable to parse 7249a243e69Sdrhthis keyword, try treating it as an identifier instead."</p> 7259bccde3dSdrh 7269a243e69Sdrh<p>The syntax of <tt>%fallback</tt> is as follows: 7279bccde3dSdrh 7289bccde3dSdrh<blockquote> 7299bccde3dSdrh<tt>%fallback</tt> <i>ID</i> <i>TOKEN...</i> <b>.</b> 7309a243e69Sdrh</blockquote></p> 7319bccde3dSdrh 7329a243e69Sdrh<p>In words, the <tt>%fallback</tt> directive is followed by a list of token 7339a243e69Sdrhnames terminated by a period. 7349a243e69SdrhThe first token name is the fallback token — the 7359bccde3dSdrhtoken to which all the other tokens fall back to. The second and subsequent 7369bccde3dSdrharguments are tokens which fall back to the token identified by the first 7379a243e69Sdrhargument.</p> 7389bccde3dSdrh 7399bccde3dSdrh<a name='pifdef'></a> 7409a243e69Sdrh<h4>The <tt>%ifdef</tt>, <tt>%ifndef</tt>, and <tt>%endif</tt> directives</h4> 7419bccde3dSdrh 7429a243e69Sdrh<p>The <tt>%ifdef</tt>, <tt>%ifndef</tt>, and <tt>%endif</tt> directives 7439a243e69Sdrhare similar to #ifdef, #ifndef, and #endif in the C-preprocessor, 7449a243e69Sdrhjust not as general. 7459bccde3dSdrhEach of these directives must begin at the left margin. No whitespace 7469a243e69Sdrhis allowed between the "%" and the directive name.</p> 7479bccde3dSdrh 7489a243e69Sdrh<p>Grammar text in between "<tt>%ifdef MACRO</tt>" and the next nested 7499a243e69Sdrh"<tt>%endif</tt>" is 7509bccde3dSdrhignored unless the "-DMACRO" command-line option is used. Grammar text 7519a243e69Sdrhbetwen "<tt>%ifndef MACRO</tt>" and the next nested "<tt>%endif</tt>" is 7529a243e69Sdrhincluded except when the "-DMACRO" command-line option is used.</p> 7539bccde3dSdrh 7549a243e69Sdrh<p>Note that the argument to <tt>%ifdef</tt> and <tt>%ifndef</tt> must 7559a243e69Sdrhbe a single preprocessor symbol name, not a general expression. 7569a243e69SdrhThere is no "<tt>%else</tt>" directive.</p> 7579bccde3dSdrh 7589bccde3dSdrh 7599bccde3dSdrh<a name='pinclude'></a> 76075897234Sdrh<h4>The <tt>%include</tt> directive</h4> 76175897234Sdrh 7629a243e69Sdrh<p>The <tt>%include</tt> directive specifies C code that is included at the 7639a243e69Sdrhtop of the generated parser. You can include any text you want — 764f2340fc7Sdrhthe Lemon parser generator copies it blindly. If you have multiple 7659a243e69Sdrh<tt>%include</tt> directives in your grammar file, their values are concatenated 7669a243e69Sdrhso that all <tt>%include</tt> code ultimately appears near the top of the 7679a243e69Sdrhgenerated parser, in the same order as it appeared in the grammar.</p> 76875897234Sdrh 7699a243e69Sdrh<p>The <tt>%include</tt> directive is very handy for getting some extra #include 77075897234Sdrhpreprocessor statements at the beginning of the generated parser. 77175897234SdrhFor example:</p> 77275897234Sdrh 77375897234Sdrh<p><pre> 77475897234Sdrh %include {#include <unistd.h>} 77575897234Sdrh</pre></p> 77675897234Sdrh 77775897234Sdrh<p>This might be needed, for example, if some of the C actions in the 7789a243e69Sdrhgrammar call functions that are prototyped in unistd.h.</p> 77975897234Sdrh 780*60ce5d31Sdrh<p>Use the <tt><a href="#pcode">%code</a></tt> directive to add code to 781*60ce5d31Sdrhthe end of the generated parser.</p> 782*60ce5d31Sdrh 7839bccde3dSdrh<a name='pleft'></a> 78475897234Sdrh<h4>The <tt>%left</tt> directive</h4> 78575897234Sdrh 7869a243e69SdrhThe <tt>%left</tt> directive is used (along with the 7879a243e69Sdrh<tt><a href='#pright'>%right</a></tt> and 7889a243e69Sdrh<tt><a href='#pnonassoc'>%nonassoc</a></tt> directives) to declare 7899a243e69Sdrhprecedences of terminal symbols. 7909a243e69SdrhEvery terminal symbol whose name appears after 7919a243e69Sdrha <tt>%left</tt> directive but before the next period (".") is 79275897234Sdrhgiven the same left-associative precedence value. Subsequent 7939a243e69Sdrh<tt>%left</tt> directives have higher precedence. For example:</p> 79475897234Sdrh 79575897234Sdrh<p><pre> 79675897234Sdrh %left AND. 79775897234Sdrh %left OR. 79875897234Sdrh %nonassoc EQ NE GT GE LT LE. 79975897234Sdrh %left PLUS MINUS. 80075897234Sdrh %left TIMES DIVIDE MOD. 80175897234Sdrh %right EXP NOT. 80275897234Sdrh</pre></p> 80375897234Sdrh 8049a243e69Sdrh<p>Note the period that terminates each <tt>%left</tt>, 8059a243e69Sdrh<tt>%right</tt> or <tt>%nonassoc</tt> 80675897234Sdrhdirective.</p> 80775897234Sdrh 80875897234Sdrh<p>LALR(1) grammars can get into a situation where they require 80975897234Sdrha large amount of stack space if you make heavy use or right-associative 8109a243e69Sdrhoperators. For this reason, it is recommended that you use <tt>%left</tt> 8119a243e69Sdrhrather than <tt>%right</tt> whenever possible.</p> 81275897234Sdrh 8139bccde3dSdrh<a name='pname'></a> 81475897234Sdrh<h4>The <tt>%name</tt> directive</h4> 81575897234Sdrh 81675897234Sdrh<p>By default, the functions generated by Lemon all begin with the 8179bccde3dSdrhfive-character string "Parse". You can change this string to something 8189a243e69Sdrhdifferent using the <tt>%name</tt> directive. For instance:</p> 81975897234Sdrh 82075897234Sdrh<p><pre> 82175897234Sdrh %name Abcde 82275897234Sdrh</pre></p> 82375897234Sdrh 82475897234Sdrh<p>Putting this directive in the grammar file will cause Lemon to generate 82575897234Sdrhfunctions named 82675897234Sdrh<ul> 82775897234Sdrh<li> AbcdeAlloc(), 82875897234Sdrh<li> AbcdeFree(), 82975897234Sdrh<li> AbcdeTrace(), and 83075897234Sdrh<li> Abcde(). 83175897234Sdrh</ul> 8329a243e69SdrhThe <tt>%name</tt> directive allows you to generate two or more different 8339a243e69Sdrhparsers and link them all into the same executable.</p> 83475897234Sdrh 8359bccde3dSdrh<a name='pnonassoc'></a> 83675897234Sdrh<h4>The <tt>%nonassoc</tt> directive</h4> 83775897234Sdrh 83875897234Sdrh<p>This directive is used to assign non-associative precedence to 8399bccde3dSdrhone or more terminal symbols. See the section on 8409bccde3dSdrh<a href='#precrules'>precedence rules</a> 8419a243e69Sdrhor on the <tt><a href='#pleft'>%left</a></tt> directive 8429a243e69Sdrhfor additional information.</p> 84375897234Sdrh 8449bccde3dSdrh<a name='parse_accept'></a> 84575897234Sdrh<h4>The <tt>%parse_accept</tt> directive</h4> 84675897234Sdrh 8479a243e69Sdrh<p>The <tt>%parse_accept</tt> directive specifies a block of C code that is 8489bccde3dSdrhexecuted whenever the parser accepts its input string. To "accept" 84975897234Sdrhan input string means that the parser was able to process all tokens 85075897234Sdrhwithout error.</p> 85175897234Sdrh 85275897234Sdrh<p>For example:</p> 85375897234Sdrh 85475897234Sdrh<p><pre> 85575897234Sdrh %parse_accept { 85675897234Sdrh printf("parsing complete!\n"); 85775897234Sdrh } 85875897234Sdrh</pre></p> 85975897234Sdrh 8609bccde3dSdrh<a name='parse_failure'></a> 86175897234Sdrh<h4>The <tt>%parse_failure</tt> directive</h4> 86275897234Sdrh 8639a243e69Sdrh<p>The <tt>%parse_failure</tt> directive specifies a block of C code that 86475897234Sdrhis executed whenever the parser fails complete. This code is not 86575897234Sdrhexecuted until the parser has tried and failed to resolve an input 86675897234Sdrherror using is usual error recovery strategy. The routine is 86775897234Sdrhonly invoked when parsing is unable to continue.</p> 86875897234Sdrh 86975897234Sdrh<p><pre> 87075897234Sdrh %parse_failure { 87175897234Sdrh fprintf(stderr,"Giving up. Parser is hopelessly lost...\n"); 87275897234Sdrh } 87375897234Sdrh</pre></p> 87475897234Sdrh 8759bccde3dSdrh<a name='pright'></a> 87675897234Sdrh<h4>The <tt>%right</tt> directive</h4> 87775897234Sdrh 87875897234Sdrh<p>This directive is used to assign right-associative precedence to 8799bccde3dSdrhone or more terminal symbols. See the section on 8809bccde3dSdrh<a href='#precrules'>precedence rules</a> 8819bccde3dSdrhor on the <a href='#pleft'>%left</a> directive for additional information.</p> 88275897234Sdrh 8839bccde3dSdrh<a name='stack_overflow'></a> 88475897234Sdrh<h4>The <tt>%stack_overflow</tt> directive</h4> 88575897234Sdrh 8869a243e69Sdrh<p>The <tt>%stack_overflow</tt> directive specifies a block of C code that 88775897234Sdrhis executed if the parser's internal stack ever overflows. Typically 88875897234Sdrhthis just prints an error message. After a stack overflow, the parser 88975897234Sdrhwill be unable to continue and must be reset.</p> 89075897234Sdrh 89175897234Sdrh<p><pre> 89275897234Sdrh %stack_overflow { 89375897234Sdrh fprintf(stderr,"Giving up. Parser stack overflow\n"); 89475897234Sdrh } 89575897234Sdrh</pre></p> 89675897234Sdrh 89775897234Sdrh<p>You can help prevent parser stack overflows by avoiding the use 89875897234Sdrhof right recursion and right-precedence operators in your grammar. 8999a243e69SdrhUse left recursion and and left-precedence operators instead to 90075897234Sdrhencourage rules to reduce sooner and keep the stack size down. 90175897234SdrhFor example, do rules like this: 90275897234Sdrh<pre> 90375897234Sdrh list ::= list element. // left-recursion. Good! 90475897234Sdrh list ::= . 90575897234Sdrh</pre> 90675897234SdrhNot like this: 90775897234Sdrh<pre> 90875897234Sdrh list ::= element list. // right-recursion. Bad! 90975897234Sdrh list ::= . 9109a243e69Sdrh</pre></p> 91175897234Sdrh 9129bccde3dSdrh<a name='stack_size'></a> 91375897234Sdrh<h4>The <tt>%stack_size</tt> directive</h4> 91475897234Sdrh 91575897234Sdrh<p>If stack overflow is a problem and you can't resolve the trouble 91675897234Sdrhby using left-recursion, then you might want to increase the size 91775897234Sdrhof the parser's stack using this directive. Put an positive integer 9189a243e69Sdrhafter the <tt>%stack_size</tt> directive and Lemon will generate a parse 91975897234Sdrhwith a stack of the requested size. The default value is 100.</p> 92075897234Sdrh 92175897234Sdrh<p><pre> 92275897234Sdrh %stack_size 2000 92375897234Sdrh</pre></p> 92475897234Sdrh 9259bccde3dSdrh<a name='start_symbol'></a> 92675897234Sdrh<h4>The <tt>%start_symbol</tt> directive</h4> 92775897234Sdrh 9289a243e69Sdrh<p>By default, the start symbol for the grammar that Lemon generates 92975897234Sdrhis the first non-terminal that appears in the grammar file. But you 9309a243e69Sdrhcan choose a different start symbol using the 9319a243e69Sdrh<tt>%start_symbol</tt> directive.</p> 93275897234Sdrh 93375897234Sdrh<p><pre> 93475897234Sdrh %start_symbol prog 93575897234Sdrh</pre></p> 93675897234Sdrh 9379a243e69Sdrh<a name='syntax_error'></a> 9389a243e69Sdrh<h4>The <tt>%syntax_error</tt> directive</h4> 9399a243e69Sdrh 9409a243e69Sdrh<p>See <a href='#error_processing'>Error Processing</a>.</p> 9419a243e69Sdrh 9429a243e69Sdrh<a name='token_class'></a> 9439a243e69Sdrh<h4>The <tt>%token_class</tt> directive</h4> 9449a243e69Sdrh 9459a243e69Sdrh<p>Undocumented. Appears to be related to the MULTITERMINAL concept. 9469a243e69Sdrh<a href='http://sqlite.org/src/fdiff?v1=796930d5fc2036c7&v2=624b24c5dc048e09&sbs=0'>Implementation</a>.</p> 9479a243e69Sdrh 9489bccde3dSdrh<a name='token_destructor'></a> 94975897234Sdrh<h4>The <tt>%token_destructor</tt> directive</h4> 95075897234Sdrh 9519a243e69Sdrh<p>The <tt>%destructor</tt> directive assigns a destructor to a non-terminal 9529a243e69Sdrhsymbol. (See the description of the 9539a243e69Sdrh<tt><a href='%destructor'>%destructor</a></tt> directive above.) 9549a243e69SdrhThe <tt>%token_destructor</tt> directive does the same thing 9559a243e69Sdrhfor all terminal symbols.</p> 95675897234Sdrh 95775897234Sdrh<p>Unlike non-terminal symbols which may each have a different data type 95875897234Sdrhfor their values, terminals all use the same data type (defined by 9599a243e69Sdrhthe <tt><a href='#token_type'>%token_type</a></tt> directive) 9609a243e69Sdrhand so they use a common destructor. 9619a243e69SdrhOther than that, the token destructor works just like the non-terminal 96275897234Sdrhdestructors.</p> 96375897234Sdrh 9649bccde3dSdrh<a name='token_prefix'></a> 96575897234Sdrh<h4>The <tt>%token_prefix</tt> directive</h4> 96675897234Sdrh 96775897234Sdrh<p>Lemon generates #defines that assign small integer constants 96875897234Sdrhto each terminal symbol in the grammar. If desired, Lemon will 96975897234Sdrhadd a prefix specified by this directive 9709a243e69Sdrhto each of the #defines it generates.</p> 9719a243e69Sdrh 9729a243e69Sdrh<p>So if the default output of Lemon looked like this: 97375897234Sdrh<pre> 97475897234Sdrh #define AND 1 97575897234Sdrh #define MINUS 2 97675897234Sdrh #define OR 3 97775897234Sdrh #define PLUS 4 97875897234Sdrh</pre> 97975897234SdrhYou can insert a statement into the grammar like this: 98075897234Sdrh<pre> 98175897234Sdrh %token_prefix TOKEN_ 98275897234Sdrh</pre> 98375897234Sdrhto cause Lemon to produce these symbols instead: 98475897234Sdrh<pre> 98575897234Sdrh #define TOKEN_AND 1 98675897234Sdrh #define TOKEN_MINUS 2 98775897234Sdrh #define TOKEN_OR 3 98875897234Sdrh #define TOKEN_PLUS 4 9899a243e69Sdrh</pre></p> 99075897234Sdrh 9919bccde3dSdrh<a name='token_type'></a><a name='ptype'></a> 99275897234Sdrh<h4>The <tt>%token_type</tt> and <tt>%type</tt> directives</h4> 99375897234Sdrh 99475897234Sdrh<p>These directives are used to specify the data types for values 99575897234Sdrhon the parser's stack associated with terminal and non-terminal 99675897234Sdrhsymbols. The values of all terminal symbols must be of the same 99775897234Sdrhtype. This turns out to be the same data type as the 3rd parameter 99875897234Sdrhto the Parse() function generated by Lemon. Typically, you will 99975897234Sdrhmake the value of a terminal symbol by a pointer to some kind of 100075897234Sdrhtoken structure. Like this:</p> 100175897234Sdrh 100275897234Sdrh<p><pre> 100375897234Sdrh %token_type {Token*} 100475897234Sdrh</pre></p> 100575897234Sdrh 100675897234Sdrh<p>If the data type of terminals is not specified, the default value 1007dfe4e6bbSdrhis "void*".</p> 100875897234Sdrh 100975897234Sdrh<p>Non-terminal symbols can each have their own data types. Typically 10109a243e69Sdrhthe data type of a non-terminal is a pointer to the root of a parse tree 101175897234Sdrhstructure that contains all information about that non-terminal. 101275897234SdrhFor example:</p> 101375897234Sdrh 101475897234Sdrh<p><pre> 101575897234Sdrh %type expr {Expr*} 101675897234Sdrh</pre></p> 101775897234Sdrh 101875897234Sdrh<p>Each entry on the parser's stack is actually a union containing 101975897234Sdrhinstances of all data types for every non-terminal and terminal symbol. 102075897234SdrhLemon will automatically use the correct element of this union depending 102175897234Sdrhon what the corresponding non-terminal or terminal symbol is. But 102275897234Sdrhthe grammar designer should keep in mind that the size of the union 102375897234Sdrhwill be the size of its largest element. So if you have a single 102475897234Sdrhnon-terminal whose data type requires 1K of storage, then your 100 102575897234Sdrhentry parser stack will require 100K of heap space. If you are willing 102675897234Sdrhand able to pay that price, fine. You just need to know.</p> 102775897234Sdrh 10289bccde3dSdrh<a name='pwildcard'></a> 10299bccde3dSdrh<h4>The <tt>%wildcard</tt> directive</h4> 10309bccde3dSdrh 10319a243e69Sdrh<p>The <tt>%wildcard</tt> directive is followed by a single token name and a 10329bccde3dSdrhperiod. This directive specifies that the identified token should 10339a243e69Sdrhmatch any input token.</p> 10349bccde3dSdrh 10359bccde3dSdrh<p>When the generated parser has the choice of matching an input against 10369bccde3dSdrhthe wildcard token and some other token, the other token is always used. 10379a243e69SdrhThe wildcard token is only matched if there are no alternatives.</p> 10389bccde3dSdrh 10399a243e69Sdrh<a name='error_processing'></a> 104075897234Sdrh<h3>Error Processing</h3> 104175897234Sdrh 104275897234Sdrh<p>After extensive experimentation over several years, it has been 104375897234Sdrhdiscovered that the error recovery strategy used by yacc is about 104475897234Sdrhas good as it gets. And so that is what Lemon uses.</p> 104575897234Sdrh 104675897234Sdrh<p>When a Lemon-generated parser encounters a syntax error, it 10479a243e69Sdrhfirst invokes the code specified by the <tt>%syntax_error</tt> directive, if 104875897234Sdrhany. It then enters its error recovery strategy. The error recovery 104975897234Sdrhstrategy is to begin popping the parsers stack until it enters a 105075897234Sdrhstate where it is permitted to shift a special non-terminal symbol 10519bccde3dSdrhnamed "error". It then shifts this non-terminal and continues 10529a243e69Sdrhparsing. The <tt>%syntax_error</tt> routine will not be called again 105375897234Sdrhuntil at least three new tokens have been successfully shifted.</p> 105475897234Sdrh 105575897234Sdrh<p>If the parser pops its stack until the stack is empty, and it still 10569a243e69Sdrhis unable to shift the error symbol, then the 10579a243e69Sdrh<tt><a href='#parse_failure'>%parse_failure</a></tt> routine 105875897234Sdrhis invoked and the parser resets itself to its start state, ready 105975897234Sdrhto begin parsing a new file. This is what will happen at the very 106075897234Sdrhfirst syntax error, of course, if there are no instances of the 10619bccde3dSdrh"error" non-terminal in your grammar.</p> 106275897234Sdrh 106375897234Sdrh</body> 106475897234Sdrh</html> 1065