lily/source-file.cc

   1 /*
   2   This file is part of LilyPond, the GNU music typesetter.
   3
   4   Copyright (C) 1997--2011 Jan Nieuwenhuizen <janneke@gnu.org>
   5   Han-Wen Nienhuys <hanwen@xs4all.nl>
   6
   7   LilyPond is free software: you can redistribute it and/or modify
   8   it under the terms of the GNU General Public License as published by
   9   the Free Software Foundation, either version 3 of the License, or
  10   (at your option) any later version.
  11
  12   LilyPond is distributed in the hope that it will be useful,
  13   but WITHOUT ANY WARRANTY; without even the implied warranty of
  14   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15   GNU General Public License for more details.
  16
  17   You should have received a copy of the GNU General Public License
  18   along with LilyPond.  If not, see <http://www.gnu.org/licenses/>.
  19 */
  20
  21 #if GCC_MAJOR < 4
  22 #define _GLIBCXX_HAVE_MBSTATE_T
  23 #include <wchar.h>
  24 #endif /* GCC_MAJOR < 4 */
  25
  26 #include "source-file.hh"
  27
  28 #include "config.hh"
  29
  30 #include <cstdio>
  31
  32 #if HAVE_SSTREAM
  33 #include <sstream>
  34 #else
  35 #include <strstream>
  36 #define istringstream(x) istrstream (x, length ())
  37 #endif
  38 using namespace std;
  39
  40 #include "file-name-map.hh"
  41 #include "international.hh"
  42 #include "misc.hh"
  43 #include "warn.hh"
  44
  45 void
  46 Source_file::load_stdin ()
  47 {
  48   characters_.clear ();
  49   int c;
  50   while ((c = fgetc (stdin)) != EOF)
  51     characters_.push_back ((char)c);
  52 }
  53
  54 /*
  55   return contents of FILENAME. *Not 0-terminated!*
  56  */
  57 vector<char>
  58 gulp_file (string filename, int desired_size)
  59 {
  60   /* "b" must ensure to open literally, avoiding text (CR/LF)
  61      conversions.  */
  62   FILE *f = fopen (filename.c_str (), "rb");
  63   if (!f)
  64     {
  65       warning (_f ("cannot open file: `%s'", filename.c_str ()));
  66
  67       vector<char> cxx_arr;
  68       return cxx_arr;
  69     }
  70
  71   fseek (f, 0, SEEK_END);
  72   int real_size = ftell (f);
  73   int read_count = real_size;
  74
  75   if (desired_size > 0)
  76     read_count = min (read_count, desired_size);
  77
  78   rewind (f);
  79
  80   char *str = new char[read_count + 1];
  81   str[read_count] = 0;
  82
  83   int bytes_read = fread (str, sizeof (char), read_count, f);
  84   if (bytes_read != read_count)
  85     warning (_f ("expected to read %d characters, got %d", bytes_read,
  86                  read_count));
  87   fclose (f);
  88   int filesize = bytes_read;
  89
  90   vector<char> cxx_arr;
  91   cxx_arr.resize (filesize);
  92
  93   copy (str, str + filesize, cxx_arr.begin ());
  94
  95   delete[] str;
  96   return cxx_arr;
  97 }
  98
  99 void
 100 Source_file::init ()
 101 {
 102   istream_ = 0;
 103   line_offset_ = 0;
 104   str_port_ = SCM_EOL;
 105   self_scm_ = SCM_EOL;
 106   smobify_self ();
 107 }
 108
 109 Source_file::Source_file (string filename, string data)
 110 {
 111   init ();
 112
 113   name_ = filename;
 114
 115   characters_.resize (data.length ());
 116   copy (data.begin (), data.end (), characters_.begin ());
 117
 118   characters_.push_back (0);
 119
 120   init_port ();
 121
 122   for (vsize i = 0; i < characters_.size (); i++)
 123     if (characters_[i] == '\n')
 124       newline_locations_.push_back (&characters_[0] + i);
 125 }
 126
 127 Source_file::Source_file (string filename_string)
 128 {
 129   init ();
 130
 131   name_ = filename_string;
 132
 133   if (filename_string == "-")
 134     load_stdin ();
 135   else
 136     {
 137       characters_ = gulp_file (filename_string, -1);
 138     }
 139
 140   characters_.push_back (0);
 141
 142   init_port ();
 143
 144   for (vsize i = 0; i < characters_.size (); i++)
 145     if (characters_[i] == '\n')
 146       newline_locations_.push_back (&characters_[0] + i);
 147 }
 148
 149 void
 150 Source_file::init_port ()
 151 {
 152   SCM str = scm_from_locale_string (c_str ());
 153   str_port_ = scm_mkstrport (SCM_INUM0, str, SCM_OPN | SCM_RDNG, __FUNCTION__);
 154   scm_set_port_filename_x (str_port_, ly_string2scm (name_));
 155 }
 156
 157
 158 istream *
 159 Source_file::get_istream ()
 160 {
 161   if (!istream_)
 162     {
 163       if (length ()) // can-t this be done without such a hack?
 164         istream_ = new istringstream (c_str ());
 165       else
 166         {
 167           istream_ = new istringstream ("");
 168           istream_->setstate (ios::eofbit);
 169           //      istream_->set (ios::eofbit);
 170         }
 171     }
 172   return istream_;
 173 }
 174
 175 string
 176 Source_file::file_line_column_string (char const *context_str0) const
 177 {
 178   if (!c_str ())
 179     return " (" + _ ("position unknown") + ")";
 180   else
 181     {
 182       int l, ch, col, offset;
 183       get_counts (context_str0, &l, &ch, &col, &offset);
 184
 185       return name_string () + ":" + to_string (l)
 186         + ":" + to_string (col);
 187     }
 188 }
 189
 190 string
 191 Source_file::quote_input (char const *pos_str0) const
 192 {
 193   if (!contains (pos_str0))
 194     return " (" + _ ("position unknown") + ")";
 195
 196   int l, ch, col, offset;
 197   get_counts (pos_str0, &l, &ch, &col, &offset);
 198   string line = line_string (pos_str0);
 199   string context = line.substr (0, offset)
 200     + to_string ('\n')
 201     + to_string (' ', col)
 202     + line.substr (offset, line.length () - offset);
 203   return context;
 204 }
 205
 206 string
 207 Source_file::name_string () const
 208 {
 209   return map_file_name (name_);
 210 }
 211
 212 Source_file::~Source_file ()
 213 {
 214   delete istream_;
 215 }
 216
 217 Slice
 218 Source_file::line_slice (char const *pos_str0) const
 219 {
 220   if (!contains (pos_str0))
 221     return Slice (0, 0);
 222
 223   char const *data_str0 = c_str ();
 224   char const *eof_C_ = data_str0 + length ();
 225
 226   if (pos_str0 == eof_C_)
 227     pos_str0--;
 228   char const *begin_str0 = pos_str0;
 229   while (begin_str0 > data_str0)
 230     if (*--begin_str0 == '\n')
 231       {
 232         begin_str0++;
 233         break;
 234       }
 235
 236   char const *end_str0 = pos_str0;
 237   while (end_str0 < eof_C_)
 238     if (*end_str0++ == '\n')
 239       {
 240         end_str0--;
 241         break;
 242       }
 243
 244   return Slice (begin_str0 - data_str0, end_str0 - data_str0);
 245 }
 246
 247 string
 248 Source_file::line_string (char const *pos_str0) const
 249 {
 250   if (!contains (pos_str0))
 251     return "";
 252
 253   Slice line = line_slice (pos_str0);
 254   char const *data_str0 = c_str ();
 255   return string (data_str0 + line[LEFT], line.length ());
 256 }
 257
 258 void
 259 Source_file::get_counts (char const *pos_str0,
 260                          int *line_number,
 261                          int *line_char,
 262                          int *column,
 263                          int *byte_offset) const
 264 {
 265   *line_number = 0;
 266
 267   if (!contains (pos_str0))
 268     return;
 269
 270   *line_number = get_line (pos_str0);
 271
 272   Slice line = line_slice (pos_str0);
 273   char const *data = c_str ();
 274   char const *line_start = (char const *)data + line[LEFT];
 275
 276   ssize left = (char const *) pos_str0 - line_start;
 277   string line_begin (line_start, left);
 278   char const *line_chars = line_begin.c_str ();
 279
 280   *line_char = 0;
 281   *column = 0;
 282   *byte_offset = 0;
 283
 284   while (left > 0)
 285     {
 286       size_t thislen = utf8_char_len (*line_chars);
 287
 288       if (thislen == 1 && line_chars[0] == '\t')
 289         (*column) = (*column / 8 + 1) * 8;
 290       else
 291         (*column)++;
 292
 293       (*line_char)++;
 294
 295       /*
 296         To have decent output in UTF-8 aware terminals,
 297         we must keep track of the number of bytes from
 298         the left edge of the terminal.
 299       */
 300       *byte_offset += thislen;
 301
 302       /* Advance past this character. */
 303       line_chars += thislen;
 304       left -= thislen;
 305     }
 306 }
 307
 308 bool
 309 Source_file::contains (char const *pos_str0) const
 310 {
 311   return (pos_str0 && (pos_str0 >= c_str ()) && (pos_str0 <= c_str () + length ()));
 312 }
 313
 314 int
 315 Source_file::get_line (char const *pos_str0) const
 316 {
 317   if (!contains (pos_str0))
 318     return 0;
 319
 320   if (!newline_locations_.size ())
 321     return 1;
 322
 323   /* this will find the '\n' character at the end of our line */
 324   vsize lo = lower_bound (newline_locations_,
 325                           pos_str0,
 326                           less<char const*> ());
 327
 328   /* the return value will be indexed from 1 */
 329   return lo + 1 + line_offset_;
 330 }
 331
 332 void
 333 Source_file::set_line (char const *pos_str0, int line)
 334 {
 335   int current_line = get_line (pos_str0);
 336   line_offset_ += line - current_line;
 337
 338   assert (line == get_line (pos_str0));
 339 }
 340
 341 int
 342 Source_file::length () const
 343 {
 344   return characters_.size ();
 345 }
 346
 347 char const *
 348 Source_file::c_str () const
 349 {
 350   return &characters_[0];
 351 }
 352
 353 SCM
 354 Source_file::get_port () const
 355 {
 356   return str_port_;
 357 }
 358
 359 /****************************************************************/
 360
 361 #include "ly-smobs.icc"
 362
 363 IMPLEMENT_SMOBS (Source_file);
 364 IMPLEMENT_DEFAULT_EQUAL_P (Source_file);
 365 IMPLEMENT_TYPE_P (Source_file, "ly:source-file?");
 366
 367 SCM
 368 Source_file::mark_smob (SCM smob)
 369 {
 370   Source_file *sc = (Source_file *) SCM_CELL_WORD_1 (smob);
 371
 372   return sc->str_port_;
 373 }
 374
 375
 376 int
 377 Source_file::print_smob (SCM smob, SCM port, scm_print_state *)
 378 {
 379   Source_file *sc = (Source_file *) SCM_CELL_WORD_1 (smob);
 380
 381   scm_puts ("#<Source_file ", port);
 382   scm_puts (sc->name_.c_str (), port);
 383
 384   /* Do not print properties, that is too much hassle.  */
 385   scm_puts (" >", port);
 386   return 1;
 387 }
 388