source file of the GNU LilyPond music typesetter
- (c) 1997--2002 Jan Nieuwenhuizen <janneke@gnu.org>
+ (c) 1997--2005 Jan Nieuwenhuizen <janneke@gnu.org>
Han-Wen Nienhuys <hanwen@cs.uu.nl>
*/
-#include "config.h"
+#include "source-file.hh"
+
+#include "config.hh"
+
+#if HAVE_UTF8_WCHAR_H
+#include <utf8/wchar.h> /* mbrtowc */
+#else
+#include <wchar.h> /* mbrtowc */
+#endif
+
+#include <cstdio>
-#include <stdio.h>
-#include <assert.h>
#if HAVE_SSTREAM
#include <sstream>
#else
-#include <strstream.h>
-#define istringstream(x) istrstream(x, length ())
+#include <strstream>
+#define istringstream(x) istrstream (x, length ())
#endif
-#include "string.hh"
-#include "flower-proto.hh"
#include "warn.hh"
-#include "source-file.hh"
-#include "array.hh"
+#include "file-name-map.hh"
void
Source_file::load_stdin ()
while ((c = fgetc (stdin)) != EOF)
chs.push (c);
+ chs.push (0);
length_ = chs.size ();
contents_str0_ = chs.remove_array ();
}
-
-
char *
-gulp_file (String fn, int* len)
+gulp_file (String filename, int *filesize)
{
- /*
- let's hope that "b" opens anything binary, and does not apply
- CR/LF translation
- */
- FILE * f = fopen (fn.to_str0 (), "rb");
-
+ /* "b" must ensure to open literally, avoiding text (CR/LF)
+ conversions. */
+ FILE *f = fopen (filename.to_str0 (), "rb");
if (!f)
{
- warning (_f ("can't open file: `%s'", fn.to_str0 ()));
+ warning (_f ("can't open file: `%s'", filename.to_str0 ()));
return 0;
}
- int ret = fseek (f, 0, SEEK_END);
-
- *len = ftell (f);
+ fseek (f, 0, SEEK_END);
+ *filesize = ftell (f);
rewind (f);
- char * str = new char[*len+1];
- str[*len] = 0;
- ret = fread (str, sizeof (char), *len, f);
- if (ret!=*len)
- warning (_f ("Huh? Got %d, expected %d characters", ret, *len));
+ char *str = new char[*filesize + 1];
+ str[*filesize] = 0;
+ int bytes_read = fread (str, sizeof (char), *filesize, f);
+ if (bytes_read != *filesize)
+ warning (_f ("expected to read %d characters, got %d", bytes_read,
+ *filesize));
fclose (f);
-
return str;
}
-
Source_file::Source_file (String filename, String data)
{
-#if 1
- #warning FILENAME junked. If intentional, add comment here.
- name_string_ = "";
-#else
- name_string_ = filename;
-#endif
+ name_ = filename;
istream_ = 0;
- contents_str0_ = data.get_copy_str0();
- length_ = data.length();
+ contents_str0_ = data.get_copy_str0 ();
+ length_ = data.length ();
pos_str0_ = to_str0 ();
- init_port();
+ init_port ();
+
+ for (int i = 0; i < length_; i++)
+ if (contents_str0_[i] == '\n')
+ newline_locations_.push (contents_str0_ + i);
}
Source_file::Source_file (String filename_string)
{
- name_string_ = filename_string;
+ name_ = filename_string;
istream_ = 0;
contents_str0_ = 0;
load_stdin ();
else
contents_str0_ = gulp_file (filename_string, &length_);
-
+
pos_str0_ = to_str0 ();
- init_port();
+ init_port ();
for (int i = 0; i < length_; i++)
if (contents_str0_[i] == '\n')
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_string_.get_str0()));
+ 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 ()));
}
int
Source_file::tell () const
{
- return pos_str0_ - contents_str0_;
+ return pos_str0_ - contents_str0_;
}
std::istream*
if (!to_str0 ())
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
+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.left_string (ch)
+ + to_string ('\n')
+ + to_string (' ', col)
+ + line.cut_string (ch, INT_MAX);
+ return context;
}
String
Source_file::name_string () const
{
- return name_string_;
+ return map_file_name (name_);
}
Source_file::~Source_file ()
}
Slice
-Source_file::line_slice (char const* pos_str0) const
+Source_file::line_slice (char const *pos_str0) const
{
- if (!in_b (pos_str0))
- return Slice (0,0);
+ if (!contains (pos_str0))
+ return Slice (0, 0);
- char const* data_str0 = to_str0 ();
- char const * eof_C_ = data_str0 + length ();
+ char const *data_str0 = to_str0 ();
+ 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')
{
String
Source_file::line_string (char const* pos_str0) const
{
- if (!in_b (pos_str0))
+ 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 = to_str0 ();
+ return String ((Byte const *)data_str0 + line[LEFT], line.length ());
}
-int
-Source_file::get_char (char const* pos_str0) const
-{
- if (!in_b (pos_str0))
- return 0;
-
- char const* data_str0 = to_str0 ();
- return pos_str0 - (line_slice (pos_str0)[SMALLER] + data_str0);
-}
-int
-Source_file::get_column (char const* pos_str0) const
+void
+Source_file::get_counts (char const *pos_str0,
+ int *line_number,
+ int *line_char,
+ int *column) const
{
- if (!in_b (pos_str0))
- return 0;
-
- int ch_i = get_char (pos_str0);
- String line = line_string (pos_str0);
+ if (!contains (pos_str0))
+ return;
- 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++;
+ *line_number = get_line (pos_str0);
+
+ Slice line = line_slice (pos_str0);
+ char const *data = to_str0 ();
+ Byte const *line_start = (Byte const *)data + line[LEFT];
- return col_i;
-}
+ int left = (Byte const*) pos_str0 - line_start;
+ String line_begin (line_start, left);
+ char const *line_chars = line_begin.to_str0();
+
+ *column = 0;
+ *line_char = 0;
+
+ mbstate_t state;
-String
-Source_file::error_string (char const* pos_str0) const
-{
- if (!in_b (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)
+ {
+ wchar_t multibyte[2];
+
+ /*
+ FIXME, this is apparently locale dependent.
+ */
+ size_t thislen = mbrtowc (multibyte, line_chars, left, &state);
+
+ /* 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) ++;
- return context;
+ (*line_char) ++;
+ /* Advance past this character. */
+ line_chars += thislen;
+ left -= thislen;
+ }
}
bool
-Source_file::in_b (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 ()));
}
int
Source_file::get_line (char const* pos_str0) const
{
- if (!in_b (pos_str0))
+ if (!contains (pos_str0))
return 0;
- if (!newline_locations_.size())
+ if (!newline_locations_.size ())
return 1;
- int lo=0;
- int hi = newline_locations_.size();
+ int lo = 0;
+ int hi = newline_locations_.size ();
if (newline_locations_[lo] > pos_str0)
return 1;
&lo, &hi);
if (*pos_str0 == '\n')
- lo --;
+ lo--;
return lo + 2;
}
void
Source_file::set_pos (char const * pos_str0)
{
- if (in_b (pos_str0))
+ if (contains (pos_str0))
pos_str0_ = pos_str0;
else
- error (error_string (pos_str0) + "invalid pos");
+ error (quote_input (pos_str0) + "invalid pos");
}
-char const*
+char const *
Source_file::seek_str0 (int n)
{
- char const* new_str0 = to_str0 () + n;
+ char const *new_str0 = to_str0 () + n;
if (n < 0)
new_str0 += length ();
- if (in_b (new_str0))
+ if (contains (new_str0))
pos_str0_ = new_str0;
else
- error (error_string (new_str0) + "seek past eof");
+ error (quote_input (new_str0) + "seek past eof");
return pos_str0_;
}
-char const*
+char const *
Source_file::forward_str0 (int n)
{
- char const* old_pos = pos_str0_;
- char const* new_str0 = pos_str0_ + n;
- if (in_b (new_str0))
+ 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");
+ error (quote_input (new_str0) + "forward past eof");
return old_pos;
}
String
Source_file::get_string (int n)
{
- String str = String ((Byte const*)forward_str0 (n), n);
+ String str = String ((Byte const *)forward_str0 (n), n);
return str;
}
+
+SCM
+Source_file::get_port () const
+{
+ return str_port_;
+}