#!/usr/bin/perl

###############################################################################
#                                                                             #
#  Name   : filename_helper                                                   #
#  Author : Chris Koeritz                                                     #
#  Rights : Copyright (C) 1996-$now by Author                                 #
#                                                                             #
#  Purpose:                                                                   #
#                                                                             #
#    Support for manipulating filenames.                                      #
#                                                                             #
###############################################################################
#  This program is free software; you can redistribute it and/or modify it    #
#  under the terms of the GNU General Public License as published by the Free #
#  Software Foundation; either version 2 of the License or (at your option)   #
#  any later version.  See: "http://www.gruntose.com/Info/GNU/GPL.txt" for a  #
#  version of the License.  Please send any updates to "fred@gruntose.com".   #
###############################################################################

require "importenv.pl";

############################################################################

#hmmm: make this lower-level, a script that is inherited by all perl shells.

sub yeti_interrupt_handler {
  die "caught an interrupt; exiting.\n";
}

# hook in a ctrl-c catcher, since that seems to be universally needed.
sub install_interrupt_catcher {
  $SIG{INT} = 'yeti_interrupt_handler';
  $SIG{QUIT} = 'yeti_interrupt_handler';
#print "mapped int and quit signals\n";
  return 0
}

############################################################################

# takes an array of filenames (each possibly containing spaces and/or
# wildcards) and resolves it to a useful list of actual files.

sub glob_list {
  local(@temp_list) = @_;  # the names we're given.
  local(@to_return) = ();  # the final form of the name list.
#print "temp list is @temp_list\n";

  # scan through the list we're given.
  foreach $entry (@temp_list) {
#print "entry is $entry\n";
    local(@chopped_filename) = &split_filename($entry);
#print "chopped 0=$chopped_filename[0]\n";
#print "chopped 1=$chopped_filename[1]\n";
    if ( (@chopped_filename[0] eq ".") || (@chopped_filename[0] eq "..") ) {
      # add the simple directory name into the list.
      push @to_return, $chopped_filename[0];
      next;
    }
    if (@chopped_filename[1] eq ".") {
      # add a directory that didn't have more pattern attached.
      push @to_return, $chopped_filename[0];
      next;
    }
    opendir WHERE, $chopped_filename[0];  # open the directory up.
    local(@files_found) = readdir(WHERE);
    closedir WHERE;
    foreach $possible_name (@files_found) {
      # zoom through the list and see if we need to add it to the ones
      # matching the passed in patterns.
#      if ( ($possible_name eq ".") || ($possible_name eq "..") ) { 
#        # skip the directory entries.
#        print "skipping dir entries\n";
#        next;
#      }
      # we need to process this a bit; directory patterns are different.
      local($match) = $chopped_filename[1];
      $match =~ s/\./\\./g;  # replace periods with escaped ones.
      $match =~ s/\*/.*/g;  # replace asterisks with dot star.
      $match =~ s/\+/\\+/g;  # escape plusses.
      $match = "^" . $match . "\$";  # make sure the whole thing matches.
#print "possibname is $possible_name\n";
      if ($possible_name =~ /$match/) {
        # this one matches so add it.
        push @to_return, $chopped_filename[0] . $possible_name;
      }
    }
  }
  return @to_return;
}

############################################################################

# reports if two file names are the same file.

sub same_file {
  local($file1, $file2) = @_;
 
  ($dev1, $ino1, $junk1) = stat $file1;
  ($dev2, $ino2, $junk2) = stat $file2;

  return ($dev1 == $dev2) && ($ino1 == $ino2);
}

############################################################################

# splits a filename into a directory and file specification.

sub split_filename {
  local($chewed_name) = &remove_trailing_slashes(@_);
  $chewed_name = &canonicalize($chewed_name);
  $chewed_name = &patch_name_for_pc($chewed_name);
  if ($chewed_name =~ /\//) {
    # there's a slash in there.
    local($directory_part) = $chewed_name;
    $directory_part =~ s/^(.*\/)[^\/]*$/\1/;
    local($file_part) = $chewed_name;
    $file_part =~ s/^.*\/([^\/]*)$/\1/;
    if ($file_part eq "") {
      # if there was no file specification, just add a non-matching spec.
      $file_part = ".";
    }
    return ($directory_part, $file_part);
  } elsif ($chewed_name eq ".") {
    return (".", "");
  } elsif ($chewed_name eq "..") {
    return ("..", "");
  } else {
    # no slash in this name, so we fix that and also make sure we match
    # the whole name.
    return ("./", $chewed_name);
  }
}

############################################################################

# returns the base part of the filename; this omits any directories.

sub basename {
  local(@parts) = &split_filename(@_);
  return $parts[1];
}

# returns the directory part of the filename.

sub dirname {
  local(@parts) = &split_filename(@_);
  return $parts[0];
}

# returns the extension found on the filename, if any.
sub extension {
  local($base) = &basename(@_);
#printf "base is $base\n";
  local($found) = -1;
  for (local($i) = length($base) - 1; $i >= 0; $i--) {
#printf "char is " . substr($base, $i, 1) . "\n";
    if (substr($base, $i, 1) eq '.') {
      $found = $i;
#printf "got period found is $found\n";
      last;
    }
  }
  if ($found >=0) {
    return substr($base, $found, length($base) - $found);
  }
  return "";  # no extension seen.
}

# returns the portion of the filename without the extension.
sub non_extension {
  local($full) = &remove_trailing_slashes(@_);
  $full = &canonicalize($full);
  $full = &patch_name_for_pc($full);
  local($ext) = &extension($full);
  local($to_remove) = length($ext);
  return substr($full, 0, length($full) - $to_remove);
}

############################################################################

# removes all directory slashes (either '/' or '\') from the end of a string.

sub remove_trailing_slashes {
  local($directory_name) = @_;
  # start looking at the end of the string.
  local($inspection_point) = length($directory_name) - 1;
  while ($inspection_point > 0) {
    # examine the last character in the string to see if it's a slash.
    local($final_char) = substr($directory_name, $inspection_point, 1);
    # leave the loop if it's not a slash.
    if ( ($final_char ne "/") && ($final_char ne "\\") ) { last; }
    chop($directory_name);  # remove the slash.
    $inspection_point--;  # check the new last character.
  }

  return $directory_name;
}

############################################################################

# returns the proper directory separator for this platform.  this requires
# an environment variable called "OS" for non-Unix operating systems.  the
# valid values for that are shown below.

sub directory_separator {
  if ( ($OS eq "Windows_NT") || ($OS eq "Windows_95") 
      || ($OS eq "DOS") || ($OS eq "OS2") ) { return "\\"; }
  else { return "/"; }
}

############################################################################

# these mutate the directory slashes in a directory name.

# the one we use most frequently; it uses the unix slash.
sub canonicalize {
  return &canonicalizer(@_, "/");
}

# one that turns names into the style native on the current platform.
sub native_canonicalize {
  return &canonicalizer(@_, &directory_separator());
}

# one that explicitly uses pc style back-slashes.
sub pc_canonicalize {
  return &canonicalizer(@_, "\\");
}

# one that explicitly does unix style forward slashes.
sub unix_canonicalize {
  return &canonicalizer(@_, "/");
}

# similar to the normal unix version, but supports msys's weird paths.
sub msys_canonicalize {
  return &canonicalizer(@_, "/", 1);
}

# similar to the msys version, but services cygwin.
sub cygwin_canonicalize {
  return &canonicalizer(@_, "/", 2);
}


# this more general routine gets a directory separator passed in.  it then
# replaces all the separators with that one.

sub canonicalizer {
  local($directory_name) = $_[0];
  local($dirsep) = $_[1];
  local($do_funky) = $_[2];
#print "do funky is set to \"$do_funky\"\n";

#print "old dir name is \"$directory_name\"\n";
  
  if ($OS =~ /win/i) {
#somewhat abbreviated check; only catches windoze systems, not dos or os2.
    if ($do_funky eq "1") {
      # msys utilities version (http://www.mingw.org)
      $directory_name =~ s/^(.):[\\\/](.*)$/\/\1\/\2/;
    } elsif ($do_funky eq "2") {
      # cygwin utilities version (http://www.cygwin.com)
      $directory_name =~ s/^(.):[\\\/](.*)$/\/cygdrive\/\1\/\2/;
    }
#print "new dir name is \"$directory_name\"\n";
  }

  # turn all the non-default separators into the default.
  for (local($j) = 0; $j < length($directory_name); $j++) {
    if ( (substr($directory_name, $j, 1) eq "\\") 
        || (substr($directory_name, $j, 1) eq "/") ) {
      substr($directory_name, $j, 1) = $dirsep;
    }
  }
  # remove all occurrences of double separators except for the first
  # double set, which could be a UNC filename.
  local($saw_sep) = 0;
  for (local($i) = 1; $i < length($directory_name); $i++) {
    # iterate through the string looking for redundant separators.
    if (substr($directory_name, $i, 1) eq $dirsep) {
      # we found a separator character.
      if ($saw_sep) {
        # we had just seen a separator, so this is two in a row.
        local($head, $tail) = (substr($directory_name, 0, $i - 1),
            substr($directory_name, $i, length($directory_name) - 1));
        $directory_name = $head . $tail;
          # put the name back together without this redundant character.
        $i--;  # skip back one and try again.
      } else {
        # we have now seen a separator.
        $saw_sep = 1;
      }
    } else {
      # this character was not a separator.
      $saw_sep = 0;
    }
  }
  if ($directory_name =~ /^.:$/) {
    # fix a dos style directory that's just X:, since we don't want the
    # current directory to be used on that device.  that's too random.
    # instead, we assume they meant the root of the drive.
    $directory_name = $directory_name . "/";
  }
  return $directory_name;
}

############################################################################

# fixes a PC directory name if it is only a drive letter plus colon.

sub patch_name_for_pc {
  local($name) = @_;
#print "name=$name\n";
  if (length($name) != 2) { return $name; }
  local($colon) = substr($name, 1, 1);
#print "colon=$colon\n";
  # check whether the string needs patching.
  if ($colon eq ":") {
    # name is currently in feeble form of "X:"; fix it.
    $name = $name . '/';
  }
#print "returning=$name\n";
  return $name;
}

############################################################################

# tells whether a filename is important or not.  the unimportant category
# can usually be safely ignored or deleted.

sub important_filename {
  local($name) = &basename($_[0]);
  
  # these are endings that we consider unimportant.  where a caret is used
  # at the front, we will match only the whole string.  double slashes are
  # used before periods to ensure we match a real period character.

  local(@junk_files) = ("~", "^\\.#.*", "^\\._.*", "\\.aps", "\\.clw",
      "^cpdiff_tmp\\.txt", "^Debug", "^\\.ds_store", "^diffs\\.txt",
      "^diff_tmp\\.txt", "\\.dsp", "\\.dsw", "^generated_.*", "\\.gid",
      "gmon\\.out", "^isconfig\\.ini", "\\.log", "^manifest.txt", "\\.ncb",
      "^obj", "\\.obj", "\\.opt", "\\.plg", "^Release", "\\.res", "\\.sbr",
      ".*scc", "\\.stackdump", "^string1033\\.txt", "\\.suo", "\\.swp",
      "^thumbs.db", "\\.tmp", "^trans\\.tbl", "\\.user", "^version\\.h",
      "_version\\.rc", "\\.ws4", "\\.wsm");

  foreach $temp (@junk_files) {
    $temp = $temp . '$';
    if ($name =~ /${temp}/i) { return 0; }
      # we hit a match on it being unimportant.
  }

  return 1;  # anything else is considered important.
}

############################################################################

sub sanitize_name {
  return &patch_name_for_pc
      (&remove_trailing_slashes
          (&canonicalize(@_)));
}

############################################################################

sub get_drive_letter {
  local($path) = @_;
  if (substr($path, 0, 1) =~ /[a-zA-Z]/) {
    if (substr($path, 1, 1) eq ":") { return substr($path, 0, 1); }
  }
  return "";
}

############################################################################

sub remove_drive_letter {
  local($path) = @_;
  if (substr($path, 0, 1) =~ /[a-zA-Z]/) {
    if (substr($path, 1, 1) eq ":") { return substr($path, 2); }
  }
  return $path;
}

############################################################################

# these return their argument with the case flipped to lower or upper case.

sub lower {
  local($name) = @_;
  $name =~ tr/A-Z/a-z/;
  return $name;
}

sub upper {
  local($name) = @_;
  $name =~ tr/a-z/A-Z/;
  return $name;
}

############################################################################

# recursively deletes a directory that is passed as the single parameter.
# from http://developer.novell.com/wiki/index.php/Recursive_Directory_Remove
sub recursive_delete {
  my $dir = shift;
  local *DIR;

  opendir DIR, $dir or die "opendir $dir: $!";
  my $found = 0;
  while ($_ = readdir DIR) {
    next if /^\.{1,2}$/;
    my $path = "$dir/$_";
    unlink $path if -f $path;
    recursive_delete($path) if -d $path;
  }
  closedir DIR;
  rmdir $dir or print "error - $!";
}

############################################################################

1;