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()
96 gedcom_error(_("Line too long, max %d characters allowed"), MAXGEDCLINELEN);
99 static void error_level_leading_zero()
101 gedcom_error (_("Level number with leading zero not allowed"));
104 static void error_level_out_of_range()
106 gedcom_error (_("Level number out of range [0..%d]"), MAXGEDCLEVEL);
109 static void error_level_too_high(int level_diff)
111 gedcom_error (_("GEDCOM level number is %d higher than previous"),
115 static void error_tag_too_long(const char *tag)
117 gedcom_error(_("Tag '%s' too long, max %d characters allowed"),
121 static void error_invalid_character(const char *str, char ch)
123 gedcom_error(_("Invalid character for encoding: '%s' (0x%02x)"), str, ch);
126 static void error_pointer_too_long(const char *ptr)
128 gedcom_error(_("Pointer '%s' too long, max %d characters allowed"),
132 static void error_at_character()
134 gedcom_error(_("'@' character should be written as '@@' in values"));
137 static void error_unexpected_character(const char* str, char ch)
139 gedcom_error(_("Unexpected character: '%s' (0x%02x)"), str, ch);
142 /* This is to bypass the iconv conversion (if the input is UTF-8 coming
144 static int dummy_conv = 0;
146 #elif LEX_SECTION == 2
148 #define TO_INTERNAL(STR,OUTBUF) \
149 (dummy_conv ? STR : to_internal(STR, yyleng, OUTBUF))
151 #define INIT_LINE_LEN \
154 #define CHECK_LINE_LEN \
155 { if (line_len != (size_t)-1) { \
156 line_len += strlen(yytext); \
157 if (line_len > MAXGEDCLINELEN * encoding_width) { \
158 error_line_too_long(); \
159 line_len = (size_t)-1; \
165 #define MKTAGACTION(THETAG) \
167 gedcom_lval.tag.string = TO_INTERNAL(yytext, tag_buffer); \
168 gedcom_lval.tag.value = TAG_##THETAG; \
171 return TAG_##THETAG; \
174 /* The GEDCOM level number is converted into a sequence of opening
175 and closing brackets. Simply put, the following GEDCOM fragment:
188 { SOUR genes (1 higher: no closing brackets)
189 { VERS 1.6 (1 higher: no closing brackets)
190 } { NAME Genes (same level: 1 closing bracket)
191 } } { DATE 07 OCT 2001 (1 lower: 2 closing brackets)
205 But because this means that one token is converted into a series
206 of tokens, there is some initial code following immediately here
207 that returns "pending" tokens. */
209 #define ACTION_BEFORE_REGEXPS \
210 { if (level_diff < 1) { \
214 else if (level_diff == 1) { \
216 gedcom_lval.number = current_level; \
220 /* out of brackets... */ \
225 #define ACTION_INITIAL_WHITESPACE \
227 /* ignore initial whitespace further */ \
231 #define ACTION_0_DIGITS \
232 { error_level_leading_zero(); \
237 #define ACTION_DIGITS \
238 { int level = atoi(TO_INTERNAL(yytext, str_buffer)); \
240 if ((level < 0) || (level > MAXGEDCLEVEL)) { \
241 error_level_out_of_range(); \
245 level_diff = level - current_level; \
247 current_level = level; \
248 if (level_diff < 1) { \
252 else if (level_diff == 1) { \
254 gedcom_lval.number = current_level; \
258 /* should never happen (error to GEDCOM spec) */ \
259 error_level_too_high(level_diff); \
266 #define ACTION_ALPHANUM \
267 { if (strlen(yytext) > MAXGEDCTAGLEN * encoding_width) { \
268 error_tag_too_long(yytext); \
273 gedcom_lval.tag.string = TO_INTERNAL(yytext, tag_buffer); \
274 gedcom_lval.tag.value = USERTAG; \
281 #define ACTION_DELIM \
283 gedcom_lval.string = TO_INTERNAL(yytext, str_buffer); \
291 tmp = TO_INTERNAL(yytext, str_buffer); \
293 /* Something went wrong during conversion... */ \
294 error_invalid_character(yytext, yytext[0]); \
298 gedcom_lval.string = tmp; \
299 /* Due to character conversions, it is possible that the current \
300 character will be combined with the next, and so now we don't have a \
302 In principle, this is only applicable to the 1byte case (e.g. ANSEL),\
303 but it doesn't harm the unicode case. \
305 if (strlen(gedcom_lval.string) > 0) \
311 #define ACTION_ESCAPE \
313 gedcom_lval.string = TO_INTERNAL(yytext, str_buffer); \
318 #define ACTION_POINTER \
320 if (strlen(yytext) > MAXGEDCPTRLEN * encoding_width) { \
321 error_pointer_too_long(yytext); \
324 gedcom_lval.string = TO_INTERNAL(yytext, ptr_buffer); \
329 /* Due to the conversion of level numbers into brackets, the
330 terminator is not important, so no token is returned here.
331 Although not strictly according to the GEDCOM spec, we'll ignore
332 whitespace just before the terminator.
335 #define ACTION_TERMINATOR \
342 /* Eventually we have to return 1 closing bracket (for the trailer).
343 We can detect whether we have sent the closing bracket using the
344 level_diff (at eof, first it is 2, then we increment it ourselves)
348 { if (level_diff == 2) { \
353 char* ptr; int size; \
354 /* ... terminate lex */ \
356 /* Get rid of f*cking compiler warning from lex generated code */ \
357 /* yyterminate does return(), so program will never come here */ \
358 yy_flex_realloc(ptr, size); \
362 #define ACTION_NORMAL_AT \
363 { if (compat_mode(C_NO_DOUBLE_AT)) { \
365 char *yycopy = strdup(yytext); \
367 for (i = 0; i < 2; i++) \
368 for (j = yyleng - 1; j >= 0; --j) \
377 error_at_character(); \
382 #define ACTION_UNEXPECTED \
383 { error_unexpected_character(yytext, yytext[0]); \
387 #elif LEX_SECTION == 3
394 static void free_conv_buffers()
396 free_conv_buffer(ptr_buffer);
397 free_conv_buffer(tag_buffer);
398 free_conv_buffer(str_buffer);
401 static void yylex_cleanup()
403 /* fix memory leak in lex */
404 yy_delete_buffer(yy_current_buffer);
405 yy_current_buffer = NULL;
409 static void init_conv_buffers()
412 ptr_buffer = create_conv_buffer(INITIAL_PTR_BUFFER_LEN);
413 tag_buffer = create_conv_buffer(INITIAL_TAG_BUFFER_LEN);
414 str_buffer = create_conv_buffer(INITIAL_STR_BUFFER_LEN);
418 static int exitfuncregistered = 0;
420 void yymyinit(FILE *f)
422 if (! exitfuncregistered && atexit(yylex_cleanup) == 0)
423 exitfuncregistered = 1;
427 /* Reset our state */
429 level_diff = MAXGEDCLEVEL;