X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=lily%2Fsource-file.cc;h=49d8eb4f214f7f669c94e282d5d1c233aa2e9c8f;hb=5b4b0d6e9a197e8f9eb085b7c2ad78b8be3e5cfc;hp=eb83b0e1de40308ac598ac1312a0d5232e8f6dd0;hpb=7f223f9827a8ace82f1d065fb8cd2f8ed3a143a2;p=lilypond.git

diff --git a/lily/source-file.cc b/lily/source-file.cc
index eb83b0e1de..49d8eb4f21 100644
--- a/lily/source-file.cc
+++ b/lily/source-file.cc
@@ -3,172 +3,223 @@
 
   source file of the GNU LilyPond music typesetter
 
-  (c) 1997--2004 Jan Nieuwenhuizen <janneke@gnu.org>
-  Han-Wen Nienhuys <hanwen@cs.uu.nl>
+  (c) 1997--2008 Jan Nieuwenhuizen <janneke@gnu.org>
+  Han-Wen Nienhuys <hanwen@xs4all.nl>
 */
 
-#include "config.h"
+#if GCC_MAJOR < 4
+#define _GLIBCXX_HAVE_MBSTATE_T
+#include <wchar.h>
+#endif /* GCC_MAJOR < 4 */
+
+#include "source-file.hh"
+
+#include "config.hh"
+
+#if HAVE_UTF8_WCHAR_H
+#include <utf8/wchar.h>  /* mbrtowc */
+#else /* !HAVE_UTF8_WCHAR_H */
+#include <cwchar> /* mbrtowc */
+#endif /* HAVE_UTF8_WCHAR_H */
+
+#include <cstdio>
 
-#include <stdio.h>
-#include <assert.h>
 #if HAVE_SSTREAM
 #include <sstream>
 #else
 #include <strstream>
-#define istringstream(x) istrstream(x, length ()) 
+#define istringstream(x) istrstream (x, length ())
 #endif
+using namespace std;
 
-#include "string.hh"
-#include "flower-proto.hh"
+#include "file-name-map.hh"
+#include "international.hh"
 #include "warn.hh"
-#include "source-file.hh"
-#include "array.hh"
 
 void
 Source_file::load_stdin ()
 {
-  length_ = 0;
-
+  characters_.clear ();
   int c;
-  Array<char> chs;		// ugh.
   while ((c = fgetc (stdin)) != EOF)
-    chs.push (c);
-
-  chs.push (0);
-  length_ = chs.size ();
-  contents_str0_ = chs.remove_array ();
+    characters_.push_back (c);
 }
 
-char*
-gulp_file (String filename, int *filesize)
+/*
+  return contents of FILENAME. *Not 0-terminated!* 
+ */
+vector<char>
+gulp_file (string filename, int desired_size)
 {
   /* "b" must ensure to open literally, avoiding text (CR/LF)
      conversions.  */
-  FILE *f = fopen (filename.to_str0 (), "rb");
+  FILE *f = fopen (filename.c_str (), "rb");
   if (!f)
     {
-      warning (_f ("can't open file: `%s'", filename.to_str0 ()));
-      return 0;
+      warning (_f ("cannot open file: `%s'", filename.c_str ()));
+
+      vector<char> cxx_arr;
+      return cxx_arr;
     }
 
   fseek (f, 0, SEEK_END);
-  *filesize = ftell (f);
+  int real_size = ftell (f);
+  int read_count = real_size;
+
+  if (desired_size > 0)
+    read_count = min (read_count, desired_size);
+  
   rewind (f);
 
-  char *str = new char[*filesize + 1];
-  str[*filesize] = 0;
+  char *str = new char[read_count + 1];
+  str[read_count] = 0;
 
-  int bytes_read = fread (str, sizeof (char), *filesize, f);
-  if (bytes_read != *filesize)
-    warning (_f ("Huh?  Got %d, expected %d characters", bytes_read,
-		 *filesize));
+  int bytes_read = fread (str, sizeof (char), read_count, f);
+  if (bytes_read != read_count)
+    warning (_f ("expected to read %d characters, got %d", bytes_read,
+		 read_count));
   fclose (f);
+  int filesize = bytes_read;
+
+  vector<char> cxx_arr;
+  cxx_arr.resize (filesize);
 
-  return str;
+  copy (str, str + filesize, cxx_arr.begin ());
+  
+  delete[] str;
+  return cxx_arr;
 }
 
-/*
-  Unused.
- */
-Source_file::Source_file (String filename, String data)
+void
+Source_file::init ()
 {
-  name_ = filename;
   istream_ = 0;
-  contents_str0_ = data.get_copy_str0();
-  length_ = data.length ();
-  pos_str0_ = to_str0 ();
+  line_offset_ = 0;
+  str_port_ = SCM_EOL;
+  self_scm_ = SCM_EOL;
+  smobify_self ();
+}
+
+Source_file::Source_file (string filename, string data)
+{
+  init ();
+  
+  name_ = filename;
+
+  characters_.resize (data.length ());
+  copy (data.begin (), data.end (), characters_.begin ());
+
+  characters_.push_back (0);
+  
   init_port ();
+
+  for (vsize i = 0; i < characters_.size (); i++)
+    if (characters_[i] == '\n')
+      newline_locations_.push_back (&characters_[0] + i);
 }
 
-Source_file::Source_file (String filename_string)
+Source_file::Source_file (string filename_string)
 {
+  init ();
+  
   name_ = filename_string;
-  istream_ = 0;
-  contents_str0_ = 0;
 
   if (filename_string == "-")
     load_stdin ();
   else
-    contents_str0_ = gulp_file (filename_string, &length_);
-  
-  pos_str0_ = to_str0 ();
+    {
+      characters_ = gulp_file (filename_string, -1);
+    }
+
+  characters_.push_back (0);
 
   init_port ();
 
-  for (int i = 0; i < length_; i++)
-    if (contents_str0_[i] == '\n')
-      newline_locations_.push (contents_str0_ + i);
+  for (vsize i = 0; i < characters_.size (); i++)
+    if (characters_[i] == '\n')
+      newline_locations_.push_back (&characters_[0] + i);
 }
 
 void
 Source_file::init_port ()
 {
-  SCM str =scm_makfrom0str (contents_str0_);
-  
-  str_port_ = scm_mkstrport (SCM_INUM0, str, SCM_OPN | SCM_RDNG,
-			     __FUNCTION__);
-  scm_set_port_filename_x (str_port_,
-			   scm_makfrom0str (name_.get_str0()));
+  SCM str = scm_from_locale_string (c_str ());
+  str_port_ = scm_mkstrport (SCM_INUM0, str, SCM_OPN | SCM_RDNG, __FUNCTION__);
+  scm_set_port_filename_x (str_port_, ly_string2scm (name_));
 }
 
-int
-Source_file::tell () const
-{
-  return pos_str0_  - contents_str0_; 
-}
 
-std::istream*
+istream *
 Source_file::get_istream ()
 {
   if (!istream_)
     {
       if (length ()) // can-t this be done without such a hack?
-	istream_ = new std::istringstream (to_str0 ());
+	istream_ = new istringstream (c_str ());
       else
 	{
-	  istream_ = new std::istringstream ("");
-	  istream_->setstate (std::ios::eofbit);
+	  istream_ = new istringstream ("");
+	  istream_->setstate (ios::eofbit);
 	  //	  istream_->set (ios::eofbit);
 	}
     }
   return istream_;
 }
 
-String
+string
 Source_file::file_line_column_string (char const *context_str0) const
 {
-  if (!to_str0 ())
+  if (!c_str ())
     return " (" + _ ("position unknown") + ")";
   else
-    return name_string () + ":" + to_string (get_line (context_str0))
-      + ":" + to_string (get_char (context_str0));
+    {
+      int l, ch, col;
+      get_counts (context_str0, &l, &ch, &col);
+
+      return name_string () + ":" + to_string (l)
+	+ ":" + to_string (col);
+    }
 }
 
-String
+string
+Source_file::quote_input (char const *pos_str0) const
+{
+  if (!contains (pos_str0))
+    return " (" + _ ("position unknown") + ")";
+
+  int l, ch, col;
+  get_counts (pos_str0, &l, &ch, &col);
+  string line = line_string (pos_str0);
+  string context = line.substr (0, ch)
+    + to_string ('\n')
+    + to_string (' ', col)
+    + line.substr (ch, line.length ()-ch);
+  return context;
+}
+
+string
 Source_file::name_string () const
 {
-  return name_;
+  return map_file_name (name_);
 }
 
 Source_file::~Source_file ()
 {
   delete istream_;
-  istream_ = 0;
-  delete[] contents_str0_;
 }
 
 Slice
-Source_file::line_slice (char const* pos_str0) const
+Source_file::line_slice (char const *pos_str0) const
 {
   if (!contains (pos_str0))
-    return Slice (0,0);
+    return Slice (0, 0);
 
-  char const* data_str0 = to_str0 ();
-  char const * eof_C_ = data_str0 + length ();
+  char const *data_str0 = c_str ();
+  char const *eof_C_ = data_str0 + length ();
 
   if (pos_str0 == eof_C_)
-    pos_str0 --;
-  char const* begin_str0 = pos_str0;
+    pos_str0--;
+  char const *begin_str0 = pos_str0;
   while (begin_str0 > data_str0)
     if (*--begin_str0 == '\n')
       {
@@ -176,7 +227,7 @@ Source_file::line_slice (char const* pos_str0) const
 	break;
       }
 
-  char const* end_str0 = pos_str0;
+  char const *end_str0 = pos_str0;
   while (end_str0 < eof_C_)
     if (*end_str0++ == '\n')
       {
@@ -187,147 +238,161 @@ Source_file::line_slice (char const* pos_str0) const
   return Slice (begin_str0 - data_str0, end_str0 - data_str0);
 }
 
-String
-Source_file::line_string (char const* pos_str0) const
+string
+Source_file::line_string (char const *pos_str0) const
 {
   if (!contains (pos_str0))
     return "";
 
   Slice line = line_slice (pos_str0);
-  char const* data_str0 = to_str0 ();
-  return String ((Byte const*)data_str0 + line[LEFT], line.length ());
+  char const *data_str0 = c_str ();
+  return string (data_str0 + line[LEFT], line.length ());
 }
 
-int
-Source_file::get_char (char const* pos_str0) const
+void
+Source_file::get_counts (char const *pos_str0,
+			 int *line_number,
+			 int *line_char,
+			 int *column) const
 {
+  *line_number = 0;
+  *line_char = 0;
+  *column = 0;
+    
   if (!contains (pos_str0))
-    return 0;
+    return;
 
-  char const* data_str0 = to_str0 ();
-  return pos_str0 - (line_slice (pos_str0)[SMALLER] + data_str0);
-}
+  *line_number = get_line (pos_str0);
 
-int
-Source_file::get_column (char const* pos_str0) const
-{
-  if (!contains (pos_str0))
-    return 0;
+  Slice line = line_slice (pos_str0);
+  char const *data = c_str ();
+  char const *line_start = (char const *)data + line[LEFT];
 
-  int ch_i = get_char (pos_str0);
-  String line = line_string (pos_str0);
+  ssize left = (char const *) pos_str0 - line_start;
+  string line_begin (line_start, left);
+  char const *line_chars = line_begin.c_str ();
 
-  int col_i = 0;
-  for (int i = 0; i < ch_i; i++)
-    if (line[i] == '\t')
-      col_i = (col_i / 8 + 1) * 8;
-    else
-      col_i++;
+  *column = 0;
+  *line_char = 0;
 
-  return col_i;
-}
+  mbstate_t state;
 
-String
-Source_file::error_string (char const* pos_str0) const
-{
-  if (!contains (pos_str0))
-    return " (" + _ ("position unknown") + ")";
+  /* Initialize the state.  */
+  memset (&state, '\0', sizeof (state));
 
-  int ch_i = get_char (pos_str0);
-  String line = line_string (pos_str0);
-  String context = line.left_string (ch_i)
-    + to_string ('\n')
-    + to_string (' ', get_column (pos_str0))
-    + line.cut_string (ch_i, INT_MAX);
+  while (left > 0)
+    {
+      /*
+	FIXME, this is apparently locale dependent.
+      */
+#if HAVE_MBRTOWC
+      wchar_t multibyte[2];
+      size_t thislen = mbrtowc (multibyte, line_chars, left, &state);
+#else
+      size_t thislen = 1;
+#endif /* !HAVE_MBRTOWC */
 
-  return context;
+      /* Stop converting at invalid character;
+	 this can mean we have read just the first part
+	 of a valid character.  */
+      if (thislen == (size_t) -1)
+	break;
+
+      /* We want to handle embedded NUL bytes
+	 but the return value is 0.  Correct this.  */
+      if (thislen == 0)
+	thislen = 1;
+
+      if (thislen == 1 && line_chars[0] == '\t')
+	(*column) = (*column / 8 + 1) * 8;
+      else
+	(*column)++;
+
+      (*line_char)++;
+      /* Advance past this character. */
+      line_chars += thislen;
+      left -= thislen;
+    }
 }
 
 bool
-Source_file::contains (char const* pos_str0) const
+Source_file::contains (char const *pos_str0) const
 {
-  return (pos_str0 && (pos_str0 >= to_str0 ()) && (pos_str0 <= to_str0 () + length ()));
+  return (pos_str0 && (pos_str0 >= c_str ()) && (pos_str0 <= c_str () + length ()));
 }
 
 int
-Source_file::get_line (char const* pos_str0) const
+Source_file::get_line (char const *pos_str0) const
 {
   if (!contains (pos_str0))
     return 0;
 
   if (!newline_locations_.size ())
     return 1;
-  
-  int lo=0;
-  int hi = newline_locations_.size ();
 
-  if (newline_locations_[lo] > pos_str0)
-    return 1;
-  
-  if (newline_locations_[hi-1] < pos_str0)
-    return hi;
-  
-  binary_search_bounds (newline_locations_,
-			pos_str0, 
-			Link_array<char>::default_compare,
-			&lo, &hi);
-
-  if (*pos_str0 == '\n')
-    lo --;
-  return lo + 2;
+  /* this will find the '\n' character at the end of our line */
+  vsize lo = lower_bound (newline_locations_,
+			  pos_str0,
+			  less<char const*> ());
+
+  /* the return value will be indexed from 1 */
+  return lo + 1 + line_offset_;
+}
+
+void
+Source_file::set_line (char const *pos_str0, int line)
+{
+  int current_line = get_line (pos_str0);
+  line_offset_ += line - current_line;
+
+  assert (line == get_line (pos_str0));
 }
 
 int
 Source_file::length () const
 {
-  return length_;
+  return characters_.size ();
 }
 
 char const *
-Source_file::to_str0 () const
+Source_file::c_str () const
 {
-  return contents_str0_;
+  return &characters_[0];
 }
 
-void
-Source_file::set_pos (char const * pos_str0)
+SCM
+Source_file::get_port () const
 {
-  if (contains (pos_str0))
-    pos_str0_ = pos_str0;
-  else
-    error (error_string (pos_str0) + "invalid pos");
+  return str_port_;
 }
 
-char const*
-Source_file::seek_str0 (int n)
-{
-  char const* new_str0 = to_str0 () + n;
-  if (n < 0)
-    new_str0 += length ();
-  if (contains (new_str0))
-    pos_str0_ = new_str0;
-  else
-    error (error_string (new_str0) + "seek past eof");
+/****************************************************************/
 
-  return pos_str0_;
-}
+#include "ly-smobs.icc"
+
+IMPLEMENT_SMOBS (Source_file);
+IMPLEMENT_DEFAULT_EQUAL_P (Source_file);
+IMPLEMENT_TYPE_P (Source_file, "ly:source-file?");
 
-char const*
-Source_file::forward_str0 (int n)
+SCM
+Source_file::mark_smob (SCM smob)
 {
-  char const* old_pos = pos_str0_;
-  char const* new_str0 = pos_str0_ + n;
-  if (contains (new_str0))
-    pos_str0_ = new_str0;
-  else
-    error (error_string (new_str0)  + "forward past eof");
+  Source_file *sc = (Source_file *) SCM_CELL_WORD_1 (smob);
 
-  return old_pos;
+  return sc->str_port_;
 }
 
-String
-Source_file::get_string (int n)
+
+int
+Source_file::print_smob (SCM smob, SCM port, scm_print_state *)
 {
-  String str = String ((Byte const*)forward_str0 (n), n);
-  return str;
+  Source_file *sc = (Source_file *) SCM_CELL_WORD_1 (smob);
+
+  scm_puts ("#<Source_file ", port);
+  scm_puts (sc->name_.c_str (), port);
+
+  /* Do not print properties, that is too much hassle.  */
+  scm_puts (" >", port);
+  return 1;
 }
+