X-Git-Url: https://git.dlugolecki.net.pl/?a=blobdiff_plain;f=gedcom%2Fmultilex.c;h=393516adc8fdde69a30ad81e005b708cd1dedc33;hb=6103dd898c4de86c68891cc0222543988a2caab7;hp=727c17959cb8f7aa03a6c9040ecf415e91d0cedf;hpb=9cf569cd68164b6611101fefce7604edbad0b3ce;p=gedcom-parse.git

diff --git a/gedcom/multilex.c b/gedcom/multilex.c
index 727c179..393516a 100644
--- a/gedcom/multilex.c
+++ b/gedcom/multilex.c
@@ -24,31 +24,34 @@
 #include "gedcom_internal.h"
 #include "multilex.h"
 #include "encoding.h"
+#include "encoding_state.h"
 #include "xref.h"
 
-int line_no;
+int line_no = 0;
 
 typedef int (*lex_func)(void);
 lex_func lf;
 
-int lexer_init(ENCODING enc, FILE* f)
+#define NEW_MODEL_FILE "new.ged"
+
+int lexer_init(Encoding enc, FILE* f)
 {
   if (enc == ONE_BYTE) {
-    gedcom_1byte_in = f;
-    lf = &gedcom_1byte_lex;
-    set_encoding_width(enc);
+    lf  = &gedcom_1byte_lex;
+    gedcom_1byte_myinit(f);
+    set_read_encoding_width(enc);
     return open_conv_to_internal("ASCII");
   }
   else if (enc == TWO_BYTE_HILO) {
-    gedcom_hilo_in = f;
-    lf = &gedcom_hilo_lex;
-    set_encoding_width(enc);
+    lf  = &gedcom_hilo_lex;
+    gedcom_hilo_myinit(f);
+    set_read_encoding_width(enc);
     return open_conv_to_internal("UNICODE");
   }
   else if (enc == TWO_BYTE_LOHI) {
-    gedcom_lohi_in = f;
-    lf = &gedcom_lohi_lex;
-    set_encoding_width(enc);
+    lf  = &gedcom_lohi_lex;
+    gedcom_lohi_myinit(f);
+    set_read_encoding_width(enc);
     return open_conv_to_internal("UNICODE");
   }
   else {
@@ -66,86 +69,208 @@ int gedcom_lex()
   return (*lf)();
 }
 
+void rewind_file(FILE* f)
+{
+  if (fseek(f, 0, 0) != 0)
+    gedcom_warning(_("Error positioning input file: %s"), strerror(errno));
+}
+
 int determine_encoding(FILE* f)
 {
   char first[2];
+  int read;
 
-  fread(first, 1, 2, f);
-  if ((first[0] == '0') && (first[1] == ' ')) {
-    gedcom_debug_print(_("One-byte encoding"));
-    fseek(f, 0, 0);
+  set_read_encoding_bom(WITHOUT_BOM);
+  read = fread(first, 1, 2, f);
+  if (read != 2) {
+    gedcom_warning(_("Error reading from input file: %s"), strerror(errno));
+    rewind_file(f);
+    return ONE_BYTE;
+  }
+  else if ((first[0] == '0') && (first[1] == ' ')) {
+    gedcom_debug_print("One-byte encoding");
+    rewind_file(f);
     return ONE_BYTE;
   }
-  else if ((first[0] == '\0') && (first[1] == '0'))
-  {
-    gedcom_debug_print(_("Two-byte encoding, high-low"));
-    fseek(f, 0, 0);
+  else if ((first[0] == '\0') && (first[1] == '0')) {
+    gedcom_debug_print("Two-byte encoding, high-low");
+    rewind_file(f);
     return TWO_BYTE_HILO;
   }
-  else if ((first[0] == '\xFE') && (first[1] == '\xFF'))
-  {
-    gedcom_debug_print(_("Two-byte encoding, high-low, with BOM"));
+  else if ((first[0] == '\xFE') && (first[1] == '\xFF')) {
+    gedcom_debug_print("Two-byte encoding, high-low, with BOM");
+    set_read_encoding_bom(WITH_BOM);
     return TWO_BYTE_HILO;
   }
-  else if ((first[0] == '0') && (first[1] == '\0'))
-  {
-    gedcom_debug_print(_("Two-byte encoding, low-high"));
-    fseek(f, 0, 0);
+  else if ((first[0] == '0') && (first[1] == '\0')) {
+    gedcom_debug_print("Two-byte encoding, low-high");
+    rewind_file(f);
     return TWO_BYTE_LOHI;
   }
-  else if ((first[0] == '\xFF') && (first[1] == '\xFE'))
-  {
-    gedcom_debug_print(_("Two-byte encoding, low-high, with BOM"));
+  else if ((first[0] == '\xFF') && (first[1] == '\xFE')) {
+    gedcom_debug_print("Two-byte encoding, low-high, with BOM");
+    set_read_encoding_bom(WITH_BOM);
     return TWO_BYTE_LOHI;
   }
+  else if ((first[0] == '\xEF') && (first[1] == '\xBB')) {
+    read = fread(first, 1, 1, f);
+    if (read != 1) {
+      gedcom_warning(_("Error reading from input file: %s"), strerror(errno));
+      rewind_file(f);
+    }
+    else if (first[0] == '\xBF') {
+      set_read_encoding_bom(WITH_BOM);
+      gedcom_debug_print("UTF-8 encoding, with BOM");
+    }
+    else {
+      gedcom_warning(_("Unknown encoding, falling back to one-byte"));
+      rewind_file(f);
+    }
+    return ONE_BYTE;
+  }
   else {
     gedcom_warning(_("Unknown encoding, falling back to one-byte"));
-    fseek(f, 0, 0);
+    rewind_file(f);
     return ONE_BYTE;
   }
 }
 
-void gedcom_init()
+int init_called = 0;
+
+/** This function initializes the Gedcom parser library and must be called
+    before any other function in this library.
+
+    The function also initializes locale handling by calling
+    <tt> setlocale(LC_ALL, "") </tt>, in case the application would not do this
+    (it doesn't hurt for the application to do the same).
+
+    \attention This function should be called as early as possible.  The
+    requirement
+    is that it should come before the first call to \c iconv_open (part of the
+    generic character set conversion feature) in the program, either by your
+    program itself, or indirectly by the library calls it makes.
+    \attention Practically,
+    it should e.g. come before any calls to any GTK functions, because GTK
+    uses \c iconv_open in its initialization.
+
+    \retval 0 in case of success
+    \retval nonzero in case of failure (e.g. failure to set locale)
+ */
+int gedcom_init()
 {
+  init_called = 1;
   update_gconv_search_path();
+  init_encodings();
+  if (!setlocale(LC_ALL, "")
+      || ! bindtextdomain(PACKAGE, LOCALEDIR)
+      || ! bind_textdomain_codeset(PACKAGE, INTERNAL_ENCODING))
+    return 1;
+  else
+    return 0;
 }
 
-int gedcom_parse_file(char* file_name)
+/** This function parses the given file.  By itself, it doesn't provide any
+    other information than the parse result.
+
+    The function also empties the cross-reference table before parsing, and
+    checks the validity of the
+    cross-references if the parse was successful.
+    The following conditions can occur in the cross-reference table:
+      - An xref was defined, but not used (warning)
+      - An xref was used, but not defined (error)
+      - An xref was used as a different type than the defined type (error)
+
+    \param file_name The name of the Gedcom file to parse
+
+    \retval 0 if the parse was successful and no errors were found in the
+    cross-reference table
+    \retval nonzero on errors, which can include:
+            - \ref gedcom_init() was not called
+	    - The given file was not found
+	    - The parse of the given file failed
+	    - There were errors found in the cross-reference table
+ */
+
+int gedcom_parse_file(const char* file_name)
 {
-  ENCODING enc;
+  Encoding enc;
   int result = 1;
   FILE* file;
-  
-  char *save_locale  = strdup(setlocale(LC_ALL, NULL));
-  char *save_textdom = textdomain(NULL);
-  setlocale(LC_ALL, "");
-  bindtextdomain(PACKAGE, LOCALEDIR);
-  bind_textdomain_codeset(PACKAGE, INTERNAL_ENCODING);
-  textdomain(PACKAGE);
-
-  line_no = 1;
-  file = fopen(file_name, "r");
-  if (!file) {
-    gedcom_error(_("Could not open file '%s'"), file_name);
+
+  if (!init_called) {
+    gedcom_error(_("Internal error: GEDCOM parser not initialized"));
   }
   else {
-    init_encodings();
-    enc = determine_encoding(file);
-    
-    if (lexer_init(enc, file)) {
+    file = fopen(file_name, "r");
+    if (!file) {
+      gedcom_error(_("Could not open file '%s': %s"),
+		   file_name, strerror(errno));
+    }
+    else {
       line_no = 1;
-      make_xref_table();
-      result = gedcom_parse();
-      if (result == 0)
-	result = check_xref_table();
+      enc = determine_encoding(file);
+      
+      if (lexer_init(enc, file)) {
+	line_no = 0;
+	make_xref_table();
+	result = gedcom_parse();
+	line_no = 0;
+	if (result == 0)
+	  result = check_xref_table();
+      }
+      lexer_close();
+      fclose(file);
     }
-    lexer_close();
-    fclose(file);
   }
 
-  textdomain(save_textdom);
-  setlocale(LC_ALL, save_locale);
-  free(save_locale);
   return result;
 }
 
+/** This function starts a new model.  It does this by parsing the \c new.ged
+    file in the data directory of the library (\c $PREFIX/share/gedcom-parse).
+    This can be used to start from an empty model, and to build up the model
+    by adding new records yourself.
+
+    \retval 0 on success
+    \retval nonzero on errors (mainly the errors from \ref gedcom_parse_file()).
+ */
+
+int gedcom_new_model()
+{
+  int result = 1;
+  FILE* file;
+
+  file = fopen(NEW_MODEL_FILE, "r");
+  if (file) {
+    fclose(file);
+    result = gedcom_parse_file(NEW_MODEL_FILE);
+  }
+  else {
+    char* filename = (char*) malloc(strlen(PKGDATADIR) + strlen(NEW_MODEL_FILE)
+				    + 2);
+    if (!filename)
+      MEMORY_ERROR;
+    else {
+      sprintf(filename, "%s/%s", PKGDATADIR, NEW_MODEL_FILE);
+      result = gedcom_parse_file(filename);
+      free(filename);
+    }
+  }
+  return result;
+}
+
+int gedcom_check_version(int major, int minor, int patch)
+{
+  if (major < GEDCOM_PARSE_VERSION_MAJOR)
+    return 1;
+  else if (major > GEDCOM_PARSE_VERSION_MAJOR)
+    return 0;
+  else if (minor < GEDCOM_PARSE_VERSION_MINOR)
+    return 1;
+  else if (minor > GEDCOM_PARSE_VERSION_MINOR)
+    return 0;
+  else if (patch <= GEDCOM_PARSE_VERSION_PATCH)
+    return 1;
+  else
+    return 0;
+}