xref: /openbsd-src/gnu/usr.bin/perl/parser.h (revision 50b7afb2c2c0993b0894d4e34bf857cb13ed9c80)
1 /*    parser.h
2  *
3  *    Copyright (c) 2006, 2007, 2009, 2010, 2011 Larry Wall and others
4  *
5  *    You may distribute under the terms of either the GNU General Public
6  *    License or the Artistic License, as specified in the README file.
7  *
8  * This file defines the layout of the parser object used by the parser
9  * and lexer (perly.c, toke.c).
10  */
11 
12 #define YYEMPTY		(-2)
13 
14 typedef struct {
15     YYSTYPE val;    /* semantic value */
16     short   state;
17     I32     savestack_ix;	/* size of savestack at this state */
18     CV	    *compcv; /* value of PL_compcv when this value was created */
19 #ifdef DEBUGGING
20     const char  *name; /* token/rule name for -Dpv */
21 #endif
22 } yy_stack_frame;
23 
24 /* Fields that need to be shared with (i.e., visible to) inner lex-
25    ing scopes. */
26 typedef struct yy_lexshared {
27     struct yy_lexshared	*ls_prev;
28     SV			*ls_linestr;	/* mirrors PL_parser->linestr */
29     char		*ls_bufptr;	/* mirrors PL_parser->bufptr */
30     char		*re_eval_start;	/* start of "(?{..." text */
31     SV			*re_eval_str;	/* "(?{...})" text */
32     line_t		herelines;	/* number of lines in here-doc */
33 } LEXSHARED;
34 
35 typedef struct yy_parser {
36 
37     /* parser state */
38 
39     struct yy_parser *old_parser; /* previous value of PL_parser */
40     YYSTYPE	    yylval;	/* value of lookahead symbol, set by yylex() */
41     int		    yychar;	/* The lookahead symbol.  */
42 
43     /* Number of tokens to shift before error messages enabled.  */
44     int		    yyerrstatus;
45 
46     int		    stack_size;
47     int		    yylen;	/* length of active reduction */
48     yy_stack_frame  *stack;	/* base of stack */
49     yy_stack_frame  *ps;	/* current stack frame */
50 
51     /* lexer state */
52 
53     I32		lex_brackets;	/* square and curly bracket count */
54     I32		lex_casemods;	/* casemod count */
55     char	*lex_brackstack;/* what kind of brackets to pop */
56     char	*lex_casestack;	/* what kind of case mods in effect */
57     U8		lex_defer;	/* state after determined token */
58     bool	lex_dojoin;	/* doing an array interpolation */
59     U8		lex_expect;	/* expect after determined token */
60     U8		expect;		/* how to interpret ambiguous tokens */
61     I32		lex_formbrack;	/* bracket count at outer format level */
62     OP		*lex_inpat;	/* in pattern $) and $| are special */
63     OP		*lex_op;	/* extra info to pass back on op */
64     SV		*lex_repl;	/* runtime replacement from s/// */
65     U16		lex_inwhat;	/* what kind of quoting are we in */
66     OPCODE	last_lop_op;	/* last named list or unary operator */
67     I32		lex_starts;	/* how many interps done on level */
68     SV		*lex_stuff;	/* runtime pattern from m// or s/// */
69     I32		multi_start;	/* 1st line of multi-line string */
70     I32		multi_end;	/* last line of multi-line string */
71     char	multi_open;	/* delimiter of said string */
72     char	multi_close;	/* delimiter of said string */
73     bool	preambled;
74     bool        lex_re_reparsing; /* we're doing G_RE_REPARSING */
75     I32		lex_allbrackets;/* (), [], {}, ?: bracket count */
76     SUBLEXINFO	sublex_info;
77     LEXSHARED	*lex_shared;
78     SV		*linestr;	/* current chunk of src text */
79     char	*bufptr;	/* carries the cursor (current parsing
80 				   position) from one invocation of yylex
81 				   to the next */
82     char	*oldbufptr;	/* in yylex, beginning of current token */
83     char	*oldoldbufptr;	/* in yylex, beginning of previous token */
84     char	*bufend;
85     char	*linestart;	/* beginning of most recently read line */
86     char	*last_uni;	/* position of last named-unary op */
87     char	*last_lop;	/* position of last list operator */
88     /* copline is used to pass a specific line number to newSTATEOP.  It
89        is a one-time line number, as newSTATEOP invalidates it (sets it to
90        NOLINE) after using it.  The purpose of this is to report line num-
91        bers in multiline constructs using the number of the first line. */
92     line_t	copline;
93     U16		in_my;		/* we're compiling a "my"/"our" declaration */
94     U8		lex_state;	/* next token is determined */
95     U8		error_count;	/* how many compile errors so far, max 10 */
96     HV		*in_my_stash;	/* declared class of this "my" declaration */
97     PerlIO	*rsfp;		/* current source file pointer */
98     AV		*rsfp_filters;	/* holds chain of active source filters */
99     U8		form_lex_state;	/* remember lex_state when parsing fmt */
100 
101 #ifdef PERL_MAD
102     SV		*endwhite;
103     I32		faketokens;
104     I32		lasttoke;
105     SV		*nextwhite;
106     I32		realtokenstart;
107     SV		*skipwhite;
108     SV		*thisclose;
109     MADPROP *	thismad;
110     SV		*thisopen;
111     SV		*thisstuff;
112     SV		*thistoken;
113     SV		*thiswhite;
114 
115 /* What we know when we're in LEX_KNOWNEXT state. */
116     NEXTTOKE	nexttoke[5];	/* value of next token, if any */
117     I32		curforce;
118 #else
119     YYSTYPE	nextval[5];	/* value of next token, if any */
120     I32		nexttype[5];	/* type of next token */
121     I32		nexttoke;
122 #endif
123 
124     COP		*saved_curcop;	/* the previous PL_curcop */
125     char	tokenbuf[256];
126 
127     U8		lex_fakeeof;	/* precedence at which to fake EOF */
128     U8		lex_flags;
129     PERL_BITFIELD16	in_pod:1;      /* lexer is within a =pod section */
130     PERL_BITFIELD16	filtered:1;    /* source filters in evalbytes */
131 } yy_parser;
132 
133 /* flags for lexer API */
134 #define LEX_STUFF_UTF8		0x00000001
135 #define LEX_KEEP_PREVIOUS	0x00000002
136 
137 #ifdef PERL_CORE
138 # define LEX_START_SAME_FILTER	0x00000001
139 # define LEX_IGNORE_UTF8_HINTS	0x00000002
140 # define LEX_EVALBYTES		0x00000004
141 # define LEX_START_COPIED	0x00000008
142 # define LEX_DONT_CLOSE_RSFP	0x00000010
143 # define LEX_START_FLAGS \
144 	(LEX_START_SAME_FILTER|LEX_START_COPIED \
145 	|LEX_IGNORE_UTF8_HINTS|LEX_EVALBYTES|LEX_DONT_CLOSE_RSFP)
146 #endif
147 
148 /* flags for parser API */
149 #define PARSE_OPTIONAL          0x00000001
150 
151 /* values for lex_fakeeof */
152 enum {
153     LEX_FAKEEOF_NEVER,      /* don't fake EOF */
154     LEX_FAKEEOF_CLOSING,    /* fake EOF at unmatched closing punctuation */
155     LEX_FAKEEOF_NONEXPR,    /* ... and at token that can't be in expression */
156     LEX_FAKEEOF_LOWLOGIC,   /* ... and at low-precedence logic operator */
157     LEX_FAKEEOF_COMMA,      /* ... and at comma */
158     LEX_FAKEEOF_ASSIGN,     /* ... and at assignment operator */
159     LEX_FAKEEOF_IFELSE,     /* ... and at ?: operator */
160     LEX_FAKEEOF_RANGE,      /* ... and at range operator */
161     LEX_FAKEEOF_LOGIC,      /* ... and at logic operator */
162     LEX_FAKEEOF_BITWISE,    /* ... and at bitwise operator */
163     LEX_FAKEEOF_COMPARE,    /* ... and at comparison operator */
164     LEX_FAKEEOF_MAX
165 };
166 
167 /*
168  * Local variables:
169  * c-indentation-style: bsd
170  * c-basic-offset: 4
171  * indent-tabs-mode: nil
172  * End:
173  *
174  * ex: set ts=8 sts=4 sw=4 et:
175  */
176