2 Copyright (C) 2001 The Genes Development Team
3 This file is part of the Gedcom parser library.
4 Contributed by Peter Verthez <Peter.Verthez@advalvas.be>, 2001.
6 The Gedcom parser library is free software; you can redistribute it
7 and/or modify it under the terms of the GNU Lesser General Public
8 License as published by the Free Software Foundation; either
9 version 2.1 of the License, or (at your option) any later version.
11 The Gedcom parser library is distributed in the hope that it will be
12 useful, but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 Lesser General Public License for more details.
16 You should have received a copy of the GNU Lesser General Public
17 License along with the Gedcom parser library; if not, write to the
18 Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
26 #include "gedcom_internal.h"
30 #include "gedcom.tab.h"
34 static size_t encoding_width;
35 static int current_level = -1;
36 static int level_diff=MAXGEDCLEVEL;
37 static size_t line_len = 0;
39 static char ptr_buf[MAXGEDCPTRLEN * UTF_FACTOR + 1];
40 static char tag_buf[MAXGEDCTAGLEN * UTF_FACTOR + 1];
41 static char str_buf[MAXGEDCLINELEN * UTF_FACTOR + 1];
49 void message_handler(Gedcom_msg_type type, char *msg)
54 int test_loop(ENCODING enc, char* code)
58 set_encoding_width(enc);
59 gedcom_set_message_handler(message_handler);
60 res = open_conv_to_internal(code);
62 gedcom_error("Unable to open conversion context: %s",
69 case BADTOKEN: printf("BADTOKEN "); break;
70 case OPEN: printf("OPEN(%d) ", gedcom_lval.number); break;
71 case CLOSE: printf("CLOSE "); break;
72 case ESCAPE: printf("ESCAPE(%s) ", gedcom_lval.string); break;
73 case DELIM: printf("DELIM "); break;
74 case ANYCHAR: printf("%s ", gedcom_lval.string); break;
75 case POINTER: printf("POINTER(%s) ", gedcom_lval.string); break;
76 case USERTAG: printf("USERTAG(%s) ", gedcom_lval.string); break;
77 default: printf("TAG(%s) ", gedcom_lval.string); break;
82 close_conv_to_internal();
86 #endif /* of #ifdef LEXER_TEST */
88 #else /* of #ifndef IN_LEX */
90 #define TO_INTERNAL(STR,OUTBUF) \
91 to_internal(STR, yyleng, OUTBUF, sizeof(OUTBUF))
93 #define INIT_LINE_LEN \
96 #define CHECK_LINE_LEN \
97 { if (line_len != (size_t)-1) { \
98 line_len += strlen(yytext); \
99 if (line_len > MAXGEDCLINELEN * encoding_width) { \
100 gedcom_error("Line too long, max %d characters", \
102 line_len = (size_t)-1; \
108 #define MKTAGACTION(THETAG) \
110 gedcom_lval.string = TO_INTERNAL(yytext, tag_buf); \
112 return TAG_##THETAG; \
115 /* The GEDCOM level number is converted into a sequence of opening
116 and closing brackets. Simply put, the following GEDCOM fragment:
129 { SOUR genes (1 higher: no closing brackets)
130 { VERS 1.6 (1 higher: no closing brackets)
131 } { NAME Genes (same level: 1 closing bracket)
132 } } { DATE 07 OCT 2001 (1 lower: 2 closing brackets)
146 But because this means that one token is converted into a series
147 of tokens, there is some initial code following immediately here
148 that returns "pending" tokens. */
150 #define ACTION_BEFORE_REGEXPS \
151 { if (level_diff < 1) { \
155 else if (level_diff == 1) { \
157 gedcom_lval.number = current_level; \
161 /* out of brackets... */ \
166 #define ACTION_INITIAL_WHITESPACE \
168 /* ignore initial whitespace further */ \
172 #define ACTION_0_DIGITS \
173 { gedcom_error ("Level number with leading zero"); \
178 #define ACTION_DIGITS \
179 { int level = atoi(TO_INTERNAL(yytext, str_buf)); \
181 if ((level < 0) || (level > MAXGEDCLEVEL)) { \
182 gedcom_error ("Level number out of range [0..%d]", \
186 level_diff = level - current_level; \
188 current_level = level; \
189 if (level_diff < 1) { \
193 else if (level_diff == 1) { \
195 gedcom_lval.number = current_level; \
199 /* should never happen (error to GEDCOM spec) */ \
200 gedcom_error ("GEDCOM level number is %d higher than " \
208 #define ACTION_ALPHANUM \
209 { if (strlen(yytext) > MAXGEDCTAGLEN * encoding_width) { \
210 gedcom_error("Tag '%s' too long, max %d characters", \
211 yytext, MAXGEDCTAGLEN); \
215 gedcom_lval.string = TO_INTERNAL(yytext, tag_buf); \
221 #define ACTION_DELIM \
223 gedcom_lval.string = TO_INTERNAL(yytext, str_buf); \
230 gedcom_lval.string = TO_INTERNAL(yytext, str_buf); \
231 /* Due to character conversions, it is possible that the current \
232 character will be combined with the next, and so now we don't have a \
234 In principle, this is only applicable to the 1byte case (e.g. ANSEL), \
235 but it doesn't harm the unicode case. \
237 if (strlen(gedcom_lval.string) > 0) \
242 #define ACTION_ESCAPE \
244 gedcom_lval.string = TO_INTERNAL(yytext, str_buf); \
249 #define ACTION_POINTER \
251 if (strlen(yytext) > MAXGEDCPTRLEN * encoding_width) { \
252 gedcom_error("Pointer '%s' too long, max %d characters", \
253 yytext, MAXGEDCPTRLEN); \
256 gedcom_lval.string = TO_INTERNAL(yytext, ptr_buf); \
261 /* Due to the conversion of level numbers into brackets, the
262 terminator is not important, so no token is returned here.
263 Although not strictly according to the GEDCOM spec, we'll ignore
264 whitespace just before the terminator.
267 #define ACTION_TERMINATOR \
275 /* Eventually we have to return 1 closing bracket (for the trailer).
276 We can detect whether we have sent the closing bracket using the
277 level_diff (at eof, first it is 2, then we increment it ourselves)
281 { if (level_diff == 2) { \
286 /* Reset our state */ \
287 current_level = -1; \
288 level_diff = MAXGEDCLEVEL; \
289 /* ... then terminate lex */ \
295 #define ACTION_UNEXPECTED \
296 { gedcom_error("Unexpected character: '%s' (0x%02x)", \
297 yytext, yytext[0]); \