blob: ec0437e3f77618006fcfc23785c50659847b35d2 [file] [log] [blame]
Guido van Rossumf70e43a1991-02-19 12:39:46 +00001/***********************************************************
Guido van Rossumbab9d031992-04-05 14:26:55 +00002Copyright 1991, 1992 by Stichting Mathematisch Centrum, Amsterdam, The
Guido van Rossumf70e43a1991-02-19 12:39:46 +00003Netherlands.
4
5 All Rights Reserved
6
7Permission to use, copy, modify, and distribute this software and its
8documentation for any purpose and without fee is hereby granted,
9provided that the above copyright notice appear in all copies and that
10both that copyright notice and this permission notice appear in
11supporting documentation, and that the names of Stichting Mathematisch
12Centrum or CWI not be used in advertising or publicity pertaining to
13distribution of the software without specific, written prior permission.
14
15STICHTING MATHEMATISCH CENTRUM DISCLAIMS ALL WARRANTIES WITH REGARD TO
16THIS SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND
17FITNESS, IN NO EVENT SHALL STICHTING MATHEMATISCH CENTRUM BE LIABLE
18FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
19WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
20ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT
21OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
22
23******************************************************************/
24
Guido van Rossum85a5fbb1990-10-14 12:07:46 +000025/* Parser-tokenizer link implementation */
26
Guido van Rossum3f5da241990-12-20 15:06:42 +000027#include "pgenheaders.h"
Guido van Rossum85a5fbb1990-10-14 12:07:46 +000028#include "tokenizer.h"
29#include "node.h"
30#include "grammar.h"
31#include "parser.h"
Guido van Rossum3f5da241990-12-20 15:06:42 +000032#include "parsetok.h"
Guido van Rossum85a5fbb1990-10-14 12:07:46 +000033#include "errcode.h"
34
Guido van Rossum3f5da241990-12-20 15:06:42 +000035
36/* Forward */
37static int parsetok PROTO((struct tok_state *, grammar *, int, node **));
38
39
40/* Parse input coming from a string. Return error code, print some errors. */
41
42int
43parsestring(s, g, start, n_ret)
44 char *s;
45 grammar *g;
46 int start;
47 node **n_ret;
48{
49 struct tok_state *tok = tok_setups(s);
50 int ret;
51
52 if (tok == NULL) {
53 fprintf(stderr, "no mem for tok_setups\n");
54 return E_NOMEM;
55 }
56 ret = parsetok(tok, g, start, n_ret);
Guido van Rossum326f5821991-06-03 11:02:09 +000057/*
58XXX Need a more sophisticated way to report the line number.
Guido van Rossum3f5da241990-12-20 15:06:42 +000059 if (ret == E_TOKEN || ret == E_SYNTAX) {
60 fprintf(stderr, "String parsing error at line %d\n",
61 tok->lineno);
62 }
Guido van Rossum326f5821991-06-03 11:02:09 +000063*/
Guido van Rossum3f5da241990-12-20 15:06:42 +000064 tok_free(tok);
65 return ret;
66}
67
68
69/* Parse input coming from a file. Return error code, print some errors. */
70
71int
72parsefile(fp, filename, g, start, ps1, ps2, n_ret)
73 FILE *fp;
74 char *filename;
75 grammar *g;
76 int start;
77 char *ps1, *ps2;
78 node **n_ret;
79{
80 struct tok_state *tok = tok_setupf(fp, ps1, ps2);
81 int ret;
82
83 if (tok == NULL) {
84 fprintf(stderr, "no mem for tok_setupf\n");
85 return E_NOMEM;
86 }
Guido van Rossum91ece421992-03-25 22:32:00 +000087#ifdef macintosh
88 {
89 int tabsize = guesstabsize(filename);
90 if (tabsize > 0)
91 tok->tabsize = tabsize;
92 }
93#endif
Guido van Rossum3f5da241990-12-20 15:06:42 +000094 ret = parsetok(tok, g, start, n_ret);
95 if (ret == E_TOKEN || ret == E_SYNTAX) {
96 char *p;
97 fprintf(stderr, "Parsing error: file %s, line %d:\n",
98 filename, tok->lineno);
Guido van Rossum56b07c81991-06-07 13:58:56 +000099 if (tok->buf == NULL)
100 fprintf(stderr, "(EOF)\n");
101 else {
102 *tok->inp = '\0';
103 if (tok->inp > tok->buf && tok->inp[-1] == '\n')
104 tok->inp[-1] = '\0';
105 fprintf(stderr, "%s\n", tok->buf);
106 for (p = tok->buf; p < tok->cur; p++) {
107 if (*p == '\t')
108 putc('\t', stderr);
109 else
110 putc(' ', stderr);
111 }
112 fprintf(stderr, "^\n");
Guido van Rossum3f5da241990-12-20 15:06:42 +0000113 }
Guido van Rossum3f5da241990-12-20 15:06:42 +0000114 }
115 tok_free(tok);
116 return ret;
117}
Guido van Rossum85a5fbb1990-10-14 12:07:46 +0000118
119
120/* Parse input coming from the given tokenizer structure.
121 Return error code. */
122
123static int
124parsetok(tok, g, start, n_ret)
125 struct tok_state *tok;
126 grammar *g;
127 int start;
128 node **n_ret;
129{
130 parser_state *ps;
131 int ret;
Guido van Rossumd8b1d371992-03-04 16:40:44 +0000132 int started = 0;
Guido van Rossum85a5fbb1990-10-14 12:07:46 +0000133
134 if ((ps = newparser(g, start)) == NULL) {
135 fprintf(stderr, "no mem for new parser\n");
136 return E_NOMEM;
137 }
138
139 for (;;) {
140 char *a, *b;
141 int type;
142 int len;
143 char *str;
144
145 type = tok_get(tok, &a, &b);
146 if (type == ERRORTOKEN) {
147 ret = tok->done;
148 break;
149 }
Guido van Rossumd8b1d371992-03-04 16:40:44 +0000150 if (type == ENDMARKER && started) {
151 type = NEWLINE; /* Add an extra newline */
152 started = 0;
153 }
154 else
155 started = 1;
Guido van Rossum85a5fbb1990-10-14 12:07:46 +0000156 len = b - a;
157 str = NEW(char, len + 1);
158 if (str == NULL) {
159 fprintf(stderr, "no mem for next token\n");
160 ret = E_NOMEM;
161 break;
162 }
163 strncpy(str, a, len);
164 str[len] = '\0';
Guido van Rossum3f5da241990-12-20 15:06:42 +0000165 ret = addtoken(ps, (int)type, str, tok->lineno);
Guido van Rossum85a5fbb1990-10-14 12:07:46 +0000166 if (ret != E_OK) {
Guido van Rossum840bcf11990-11-18 17:39:41 +0000167 if (ret == E_DONE) {
Guido van Rossum85a5fbb1990-10-14 12:07:46 +0000168 *n_ret = ps->p_tree;
Guido van Rossum840bcf11990-11-18 17:39:41 +0000169 ps->p_tree = NULL;
170 }
Guido van Rossum85a5fbb1990-10-14 12:07:46 +0000171 else if (tok->lineno <= 1 && tok->done == E_EOF)
172 ret = E_EOF;
173 break;
174 }
175 }
176
177 delparser(ps);
178 return ret;
179}