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"
29 #include "encoding_state.h"
31 #include "gedcom.tabgen.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 struct conv_buffer* ptr_buffer = NULL;
40 static struct conv_buffer* tag_buffer = NULL;
41 static struct conv_buffer* str_buffer = NULL;
43 #define INITIAL_PTR_BUFFER_LEN MAXGEDCPTRLEN * UTF_FACTOR + 1
44 #define INITIAL_TAG_BUFFER_LEN MAXGEDCTAGLEN * UTF_FACTOR + 1
45 #define INITIAL_STR_BUFFER_LEN MAXGEDCLINELEN * UTF_FACTOR + 1
53 void message_handler(Gedcom_msg_type type, char *msg)
55 fprintf(stderr, "(%d) %s\n", type, msg);
58 int test_loop(ENCODING enc, const char* code)
62 set_encoding_width(enc);
63 gedcom_set_message_handler(message_handler);
64 res = open_conv_to_internal(code);
66 gedcom_error("Unable to open conversion context: %s",
73 case BADTOKEN: printf("BADTOKEN "); break;
74 case OPEN: printf("OPEN(%d) ", gedcom_lval.number); break;
75 case CLOSE: printf("CLOSE "); break;
76 case ESCAPE: printf("ESCAPE(%s) ", gedcom_lval.string); break;
77 case DELIM: printf("DELIM "); break;
78 case ANYCHAR: printf("%s ", gedcom_lval.string); break;
79 case POINTER: printf("POINTER(%s) ", gedcom_lval.string); break;
80 case USERTAG: printf("USERTAG(%s) ", gedcom_lval.tag.string); break;
81 default: printf("TAG(%s) ", gedcom_lval.tag.string); break;
86 close_conv_to_internal();
90 #endif /* of #ifdef LEXER_TEST */
92 /* These are defined as functions here, because xgettext has trouble
93 extracting the strings out of long pre-processor defined */
95 static void error_line_too_long(const char *line)
97 gedcom_error(_("Line too long, max %d characters allowed: %s"),
98 MAXGEDCLINELEN, line);
101 static void error_level_leading_zero()
103 gedcom_error (_("Level number with leading zero not allowed"));
106 static void error_level_out_of_range()
108 gedcom_error (_("Level number out of range [0..%d]"), MAXGEDCLEVEL);
111 static void error_level_too_high(int level_diff)
113 gedcom_error (_("GEDCOM level number is %d higher than previous"),
117 static void error_tag_too_long(const char *tag)
119 gedcom_error(_("Tag '%s' too long, max %d characters allowed"),
123 static void error_invalid_character(const char *str, char ch)
125 gedcom_error(_("Invalid character for encoding: '%s' (0x%02x)"), str, ch);
128 static void error_pointer_too_long(const char *ptr)
130 gedcom_error(_("Pointer '%s' too long, max %d characters allowed"),
134 static void error_at_character()
136 gedcom_error(_("'@' character should be written as '@@' in values"));
139 static void error_unexpected_character(const char* str, char ch)
141 gedcom_error(_("Unexpected character: '%s' (0x%02x)"), str, ch);
144 /* This is to bypass the iconv conversion (if the input is UTF-8 coming
146 static int dummy_conv = 0;
148 #elif LEX_SECTION == 2
150 #define TO_INTERNAL(STR,OUTBUF) \
151 (dummy_conv ? STR : to_internal(STR, yyleng, OUTBUF))
153 #define INIT_LINE_LEN \
156 #define CHECK_LINE_LEN \
157 { if (line_len != (size_t)-1) { \
158 line_len += strlen(yytext); \
159 if (line_len > MAXGEDCLINELEN * encoding_width) { \
160 error_line_too_long(yytext); \
161 line_len = (size_t)-1; \
167 #define MKTAGACTION(THETAG) \
169 gedcom_lval.tag.string = TO_INTERNAL(yytext, tag_buffer); \
170 gedcom_lval.tag.value = TAG_##THETAG; \
173 return TAG_##THETAG; \
176 /* The GEDCOM level number is converted into a sequence of opening
177 and closing brackets. Simply put, the following GEDCOM fragment:
190 { SOUR genes (1 higher: no closing brackets)
191 { VERS 1.6 (1 higher: no closing brackets)
192 } { NAME Genes (same level: 1 closing bracket)
193 } } { DATE 07 OCT 2001 (1 lower: 2 closing brackets)
207 But because this means that one token is converted into a series
208 of tokens, there is some initial code following immediately here
209 that returns "pending" tokens. */
211 #define ACTION_BEFORE_REGEXPS \
212 { if (level_diff < 1) { \
216 else if (level_diff == 1) { \
218 gedcom_lval.number = current_level; \
222 /* out of brackets... */ \
227 #define ACTION_INITIAL_WHITESPACE \
229 /* ignore initial whitespace further */ \
233 #define ACTION_0_DIGITS \
234 { error_level_leading_zero(); \
239 #define ACTION_DIGITS \
240 { int level = atoi(TO_INTERNAL(yytext, str_buffer)); \
242 if ((level < 0) || (level > MAXGEDCLEVEL)) { \
243 error_level_out_of_range(); \
247 level_diff = level - current_level; \
249 current_level = level; \
250 if (level_diff < 1) { \
254 else if (level_diff == 1) { \
256 gedcom_lval.number = current_level; \
260 /* should never happen (error to GEDCOM spec) */ \
261 error_level_too_high(level_diff); \
268 #define ACTION_ALPHANUM \
269 { if (strlen(yytext) > MAXGEDCTAGLEN * encoding_width) { \
270 error_tag_too_long(yytext); \
275 gedcom_lval.tag.string = TO_INTERNAL(yytext, tag_buffer); \
276 gedcom_lval.tag.value = USERTAG; \
283 #define ACTION_DELIM \
285 gedcom_lval.string = TO_INTERNAL(yytext, str_buffer); \
293 tmp = TO_INTERNAL(yytext, str_buffer); \
295 /* Something went wrong during conversion... */ \
296 error_invalid_character(yytext, yytext[0]); \
300 gedcom_lval.string = tmp; \
301 /* Due to character conversions, it is possible that the current \
302 character will be combined with the next, and so now we don't have a \
304 In principle, this is only applicable to the 1byte case (e.g. ANSEL),\
305 but it doesn't harm the unicode case. \
307 if (strlen(gedcom_lval.string) > 0) \
313 #define ACTION_ESCAPE \
315 gedcom_lval.string = TO_INTERNAL(yytext, str_buffer); \
320 #define ACTION_POINTER \
322 if (strlen(yytext) > MAXGEDCPTRLEN * encoding_width) { \
323 error_pointer_too_long(yytext); \
326 gedcom_lval.string = TO_INTERNAL(yytext, ptr_buffer); \
331 /* Due to the conversion of level numbers into brackets, the
332 terminator is not important, so no token is returned here.
333 Although not strictly according to the GEDCOM spec, we'll ignore
334 whitespace just before the terminator.
337 #define ACTION_TERMINATOR \
341 set_read_encoding_terminator(TO_INTERNAL(yytext, str_buffer)); \
346 /* Eventually we have to return 1 closing bracket (for the trailer).
347 We can detect whether we have sent the closing bracket using the
348 level_diff (at eof, first it is 2, then we increment it ourselves)
352 { if (level_diff == 2) { \
357 char* ptr; int size; \
358 /* ... terminate lex */ \
360 /* Get rid of f*cking compiler warning from lex generated code */ \
361 /* yyterminate does return(), so program will never come here */ \
362 yy_flex_realloc(ptr, size); \
366 #define ACTION_NORMAL_AT \
367 { if (compat_mode(C_NO_DOUBLE_AT)) { \
369 char *yycopy = strdup(yytext); \
371 for (i = 0; i < 2; i++) \
372 for (j = yyleng - 1; j >= 0; --j) \
381 error_at_character(); \
386 #define ACTION_UNEXPECTED \
387 { error_unexpected_character(yytext, yytext[0]); \
391 #elif LEX_SECTION == 3
398 static void free_conv_buffers()
400 free_conv_buffer(ptr_buffer);
401 free_conv_buffer(tag_buffer);
402 free_conv_buffer(str_buffer);
405 static void yylex_cleanup()
407 /* fix memory leak in lex */
408 yy_delete_buffer(yy_current_buffer);
409 yy_current_buffer = NULL;
413 static void init_conv_buffers()
416 ptr_buffer = create_conv_buffer(INITIAL_PTR_BUFFER_LEN);
417 tag_buffer = create_conv_buffer(INITIAL_TAG_BUFFER_LEN);
418 str_buffer = create_conv_buffer(INITIAL_STR_BUFFER_LEN);
422 static int exitfuncregistered = 0;
424 void yymyinit(FILE *f)
426 if (! exitfuncregistered && atexit(yylex_cleanup) == 0)
427 exitfuncregistered = 1;
431 /* Reset our state */
433 level_diff = MAXGEDCLEVEL;