2 Copyright (C) 2001, 2002 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.tabgen.h"
33 static size_t encoding_width;
34 static int current_level = -1;
35 static int level_diff=MAXGEDCLEVEL;
36 static size_t line_len = 0;
38 static struct conv_buffer* ptr_buffer = NULL;
39 static struct conv_buffer* tag_buffer = NULL;
40 static struct conv_buffer* str_buffer = NULL;
42 #define INITIAL_PTR_BUFFER_LEN MAXGEDCPTRLEN * UTF_FACTOR + 1
43 #define INITIAL_TAG_BUFFER_LEN MAXGEDCTAGLEN * UTF_FACTOR + 1
44 #define INITIAL_STR_BUFFER_LEN MAXGEDCLINELEN * UTF_FACTOR + 1
52 void message_handler(Gedcom_msg_type type, char *msg)
54 fprintf(stderr, "(%d) %s\n", type, msg);
57 int test_loop(ENCODING enc, const char* code)
61 set_encoding_width(enc);
62 gedcom_set_message_handler(message_handler);
63 res = open_conv_to_internal(code);
65 gedcom_error("Unable to open conversion context: %s",
72 case BADTOKEN: printf("BADTOKEN "); break;
73 case OPEN: printf("OPEN(%d) ", gedcom_lval.number); break;
74 case CLOSE: printf("CLOSE "); break;
75 case ESCAPE: printf("ESCAPE(%s) ", gedcom_lval.string); break;
76 case DELIM: printf("DELIM "); break;
77 case ANYCHAR: printf("%s ", gedcom_lval.string); break;
78 case POINTER: printf("POINTER(%s) ", gedcom_lval.string); break;
79 case USERTAG: printf("USERTAG(%s) ", gedcom_lval.tag.string); break;
80 default: printf("TAG(%s) ", gedcom_lval.tag.string); break;
85 close_conv_to_internal();
89 #endif /* of #ifdef LEXER_TEST */
91 /* These are defined as functions here, because xgettext has trouble
92 extracting the strings out of long pre-processor defined */
94 static void error_line_too_long(const char *line)
96 gedcom_error(_("Line too long, max %d characters allowed: %s"),
97 MAXGEDCLINELEN, line);
100 static void error_level_leading_zero()
102 gedcom_error (_("Level number with leading zero not allowed"));
105 static void error_level_out_of_range()
107 gedcom_error (_("Level number out of range [0..%d]"), MAXGEDCLEVEL);
110 static void error_level_too_high(int level_diff)
112 gedcom_error (_("GEDCOM level number is %d higher than previous"),
116 static void error_tag_too_long(const char *tag)
118 gedcom_error(_("Tag '%s' too long, max %d characters allowed"),
122 static void error_invalid_character(const char *str, char ch)
124 gedcom_error(_("Invalid character for encoding: '%s' (0x%02x)"), str, ch);
127 static void error_pointer_too_long(const char *ptr)
129 gedcom_error(_("Pointer '%s' too long, max %d characters allowed"),
133 static void error_at_character()
135 gedcom_error(_("'@' character should be written as '@@' in values"));
138 static void error_unexpected_character(const char* str, char ch)
140 gedcom_error(_("Unexpected character: '%s' (0x%02x)"), str, ch);
143 /* This is to bypass the iconv conversion (if the input is UTF-8 coming
145 static int dummy_conv = 0;
147 #elif LEX_SECTION == 2
149 #define TO_INTERNAL(STR,OUTBUF) \
150 (dummy_conv ? STR : to_internal(STR, yyleng, OUTBUF))
152 #define INIT_LINE_LEN \
155 #define CHECK_LINE_LEN \
156 { if (line_len != (size_t)-1) { \
157 line_len += strlen(yytext); \
158 if (line_len > MAXGEDCLINELEN * encoding_width) { \
159 error_line_too_long(yytext); \
160 line_len = (size_t)-1; \
166 #define MKTAGACTION(THETAG) \
168 gedcom_lval.tag.string = TO_INTERNAL(yytext, tag_buffer); \
169 gedcom_lval.tag.value = TAG_##THETAG; \
172 return TAG_##THETAG; \
175 /* The GEDCOM level number is converted into a sequence of opening
176 and closing brackets. Simply put, the following GEDCOM fragment:
189 { SOUR genes (1 higher: no closing brackets)
190 { VERS 1.6 (1 higher: no closing brackets)
191 } { NAME Genes (same level: 1 closing bracket)
192 } } { DATE 07 OCT 2001 (1 lower: 2 closing brackets)
206 But because this means that one token is converted into a series
207 of tokens, there is some initial code following immediately here
208 that returns "pending" tokens. */
210 #define ACTION_BEFORE_REGEXPS \
211 { if (level_diff < 1) { \
215 else if (level_diff == 1) { \
217 gedcom_lval.number = current_level; \
221 /* out of brackets... */ \
226 #define ACTION_INITIAL_WHITESPACE \
228 /* ignore initial whitespace further */ \
232 #define ACTION_0_DIGITS \
233 { error_level_leading_zero(); \
238 #define ACTION_DIGITS \
239 { int level = atoi(TO_INTERNAL(yytext, str_buffer)); \
241 if ((level < 0) || (level > MAXGEDCLEVEL)) { \
242 error_level_out_of_range(); \
246 level_diff = level - current_level; \
248 current_level = level; \
249 if (level_diff < 1) { \
253 else if (level_diff == 1) { \
255 gedcom_lval.number = current_level; \
259 /* should never happen (error to GEDCOM spec) */ \
260 error_level_too_high(level_diff); \
267 #define ACTION_ALPHANUM \
268 { if (strlen(yytext) > MAXGEDCTAGLEN * encoding_width) { \
269 error_tag_too_long(yytext); \
274 gedcom_lval.tag.string = TO_INTERNAL(yytext, tag_buffer); \
275 gedcom_lval.tag.value = USERTAG; \
282 #define ACTION_DELIM \
284 gedcom_lval.string = TO_INTERNAL(yytext, str_buffer); \
292 tmp = TO_INTERNAL(yytext, str_buffer); \
294 /* Something went wrong during conversion... */ \
295 error_invalid_character(yytext, yytext[0]); \
299 gedcom_lval.string = tmp; \
300 /* Due to character conversions, it is possible that the current \
301 character will be combined with the next, and so now we don't have a \
303 In principle, this is only applicable to the 1byte case (e.g. ANSEL),\
304 but it doesn't harm the unicode case. \
306 if (strlen(gedcom_lval.string) > 0) \
312 #define ACTION_ESCAPE \
314 gedcom_lval.string = TO_INTERNAL(yytext, str_buffer); \
319 #define ACTION_POINTER \
321 if (strlen(yytext) > MAXGEDCPTRLEN * encoding_width) { \
322 error_pointer_too_long(yytext); \
325 gedcom_lval.string = TO_INTERNAL(yytext, ptr_buffer); \
330 /* Due to the conversion of level numbers into brackets, the
331 terminator is not important, so no token is returned here.
332 Although not strictly according to the GEDCOM spec, we'll ignore
333 whitespace just before the terminator.
336 #define ACTION_TERMINATOR \
343 /* Eventually we have to return 1 closing bracket (for the trailer).
344 We can detect whether we have sent the closing bracket using the
345 level_diff (at eof, first it is 2, then we increment it ourselves)
349 { if (level_diff == 2) { \
354 char* ptr; int size; \
355 /* ... terminate lex */ \
357 /* Get rid of f*cking compiler warning from lex generated code */ \
358 /* yyterminate does return(), so program will never come here */ \
359 yy_flex_realloc(ptr, size); \
363 #define ACTION_NORMAL_AT \
364 { if (compat_mode(C_NO_DOUBLE_AT)) { \
366 char *yycopy = strdup(yytext); \
368 for (i = 0; i < 2; i++) \
369 for (j = yyleng - 1; j >= 0; --j) \
378 error_at_character(); \
383 #define ACTION_UNEXPECTED \
384 { error_unexpected_character(yytext, yytext[0]); \
388 #elif LEX_SECTION == 3
395 static void free_conv_buffers()
397 free_conv_buffer(ptr_buffer);
398 free_conv_buffer(tag_buffer);
399 free_conv_buffer(str_buffer);
402 static void yylex_cleanup()
404 /* fix memory leak in lex */
405 yy_delete_buffer(yy_current_buffer);
406 yy_current_buffer = NULL;
410 static void init_conv_buffers()
413 ptr_buffer = create_conv_buffer(INITIAL_PTR_BUFFER_LEN);
414 tag_buffer = create_conv_buffer(INITIAL_TAG_BUFFER_LEN);
415 str_buffer = create_conv_buffer(INITIAL_STR_BUFFER_LEN);
419 static int exitfuncregistered = 0;
421 void yymyinit(FILE *f)
423 if (! exitfuncregistered && atexit(yylex_cleanup) == 0)
424 exitfuncregistered = 1;
428 /* Reset our state */
430 level_diff = MAXGEDCLEVEL;