X-Git-Url: https://git.dlugolecki.net.pl/?a=blobdiff_plain;f=gedcom%2Fgedcom_lex_common.c;h=109da8a1c33035ca9d26197fca562e45267c8ba3;hb=87ca3acd1b3a5bdaf51384110d85b7174618e2de;hp=e1c17bb78b12dfb89b510a086d4cdc5d1282d7f4;hpb=dc88e8545a834df924004ad1a4742ee06e3e7416;p=gedcom-parse.git

diff --git a/gedcom/gedcom_lex_common.c b/gedcom/gedcom_lex_common.c
index e1c17bb..109da8a 100644
--- a/gedcom/gedcom_lex_common.c
+++ b/gedcom/gedcom_lex_common.c
@@ -1,5 +1,5 @@
 /* Common lexer code.
-   Copyright (C) 2001 The Genes Development Team
+   Copyright (C) 2001, 2002 The Genes Development Team
    This file is part of the Gedcom parser library.
    Contributed by Peter Verthez <Peter.Verthez@advalvas.be>, 2001.
 
@@ -27,9 +27,8 @@
 #include "multilex.h"
 #include "encoding.h"
 #include "gedcom.h"
-#include "gedcom.tab.h"
-
-#define YY_NO_UNPUT
+#include "gedcom.tabgen.h"
+#include "compat.h"
 
 static size_t encoding_width;
 static int current_level = -1;
@@ -43,12 +42,13 @@ static char str_buf[MAXGEDCLINELEN * UTF_FACTOR + 1];
 #ifdef LEXER_TEST 
 YYSTYPE gedcom_lval;
 int line_no = 1;
+int compat_at = 0;
 
 int gedcom_lex();
 
 void message_handler(Gedcom_msg_type type, char *msg)
 {
-  fprintf(stderr, msg);
+  fprintf(stderr, "(%d) %s\n", type, msg);
 }
 
 int test_loop(ENCODING enc, char* code)
@@ -73,8 +73,8 @@ int test_loop(ENCODING enc, char* code)
       case DELIM: printf("DELIM "); break;
       case ANYCHAR: printf("%s ", gedcom_lval.string); break;
       case POINTER: printf("POINTER(%s) ", gedcom_lval.string); break;
-      case USERTAG: printf("USERTAG(%s) ", gedcom_lval.string); break;
-      default: printf("TAG(%s) ", gedcom_lval.string); break;
+      case USERTAG: printf("USERTAG(%s) ", gedcom_lval.tag.string); break;
+      default: printf("TAG(%s) ", gedcom_lval.tag.string); break;
     }
     tok = gedcom_lex();
   }
@@ -107,7 +107,8 @@ int test_loop(ENCODING enc, char* code)
 
 #define MKTAGACTION(THETAG)                                                  \
   { CHECK_LINE_LEN;                                                          \
-    gedcom_lval.string = TO_INTERNAL(yytext, tag_buf);                       \
+    gedcom_lval.tag.string = TO_INTERNAL(yytext, tag_buf);                   \
+    gedcom_lval.tag.value  = TAG_##THETAG;                                   \
     BEGIN(NORMAL);                                                           \
     return TAG_##THETAG;                                                     \
   }
@@ -211,7 +212,8 @@ int test_loop(ENCODING enc, char* code)
        return BADTOKEN;                                                       \
      }                                                                        \
      CHECK_LINE_LEN;                                                          \
-     gedcom_lval.string = TO_INTERNAL(yytext, tag_buf);                       \
+     gedcom_lval.tag.string = TO_INTERNAL(yytext, tag_buf);                   \
+     gedcom_lval.tag.value  = USERTAG;                                        \
      BEGIN(NORMAL);                                                           \
      return USERTAG;                                                          \
    }
@@ -225,16 +227,26 @@ int test_loop(ENCODING enc, char* code)
 
 
 #define ACTION_ANY                                                            \
-  { CHECK_LINE_LEN;                                                           \
-    gedcom_lval.string = TO_INTERNAL(yytext, str_buf);                        \
-    /* Due to character conversions, it is possible that the current          \
-       character will be combined with the next, and so now we don't have a   \
-       character yet...                                                       \
-       In principle, this is only applicable to the 1byte case (e.g. ANSEL),  \
-       but it doesn't harm the unicode case.                                  \
-    */                                                                        \
-    if (strlen(gedcom_lval.string) > 0)                                       \
-      return ANYCHAR;                                                         \
+  { char* tmp;                                                                \
+    CHECK_LINE_LEN;                                                           \
+    tmp = TO_INTERNAL(yytext, str_buf);                                       \
+    if (!tmp) {                                                               \
+      /* Something went wrong during conversion... */                         \
+          gedcom_error(_("Invalid character for encoding: '%s' (0x%02x)"),    \
+		 yytext, yytext[0]);                                          \
+          return BADTOKEN;                                                    \
+    }                                                                         \
+    else {                                                                    \
+      gedcom_lval.string = tmp;                                               \
+      /* Due to character conversions, it is possible that the current        \
+         character will be combined with the next, and so now we don't have a \
+         character yet...                                                     \
+         In principle, this is only applicable to the 1byte case (e.g. ANSEL),\
+         but it doesn't harm the unicode case.                                \
+      */                                                                      \
+      if (strlen(gedcom_lval.string) > 0)                                     \
+        return ANYCHAR;                                                       \
+    }                                                                         \
   }
 
 
@@ -282,14 +294,32 @@ int test_loop(ENCODING enc, char* code)
       return CLOSE;                                                           \
     }                                                                         \
     else {                                                                    \
+      char* ptr; int size;                                                    \
       /* Reset our state */                                                   \
       current_level = -1;                                                     \
       level_diff = MAXGEDCLEVEL;                                              \
       /* ... then terminate lex */                                            \
       yyterminate();                                                          \
+      /* Get rid of f*cking compiler warning from lex generated code */       \
+      /* yyterminate does return(), so program will never come here  */       \
+      yy_flex_realloc(ptr, size);                                             \
     }                                                                         \
   } 
 
+#define ACTION_NORMAL_AT                                                      \
+  { if (compat_at) {                                                          \
+      int i, j;                                                               \
+      char *yycopy = strdup(yytext);                                          \
+      for (i = 0; i < 2; i++)                                                 \
+        for (j = yyleng - 1; j >= 0; --j)                                     \
+          unput(yycopy[j]);                                                   \
+      free(yycopy);                                                           \
+    }                                                                         \
+    else {                                                                    \
+      gedcom_error(_("'@' character should be written as '@@' in values"));   \
+      return BADTOKEN;                                                        \
+    }                                                                         \
+  }
 
 #define ACTION_UNEXPECTED                                                     \
   { gedcom_error(_("Unexpected character: '%s' (0x%02x)"),                    \