############################################################################### # # Package: NaturalDocs::File # ############################################################################### # # A package to manage file access across platforms. Incorporates functions from various standard File:: packages, but more # importantly, works around the glorious suckage present in File::Spec, at least in version 0.82 and earlier. Read the "Why oh # why?" sections for why this package was necessary. # # Usage and Dependencies: # # - The package doesn't depend on any other Natural Docs packages and is ready to use immediately. # # - All functions except assume that all parameters are canonized. # ############################################################################### # This file is part of Natural Docs, which is Copyright (C) 2003-2005 Greg Valure # Natural Docs is licensed under the GPL use File::Spec (); use File::Path (); use File::Copy (); use strict; use integer; package NaturalDocs::File; # # Function: CheckCompatibility # # Checks if the standard packages required by this one are up to snuff and dies if they aren't. This is done because I can't # tell which versions of File::Spec have splitpath just by the version numbers. # sub CheckCompatibility { my ($self) = @_; eval { File::Spec->splitpath(''); }; if ($@) { NaturalDocs::Error->SoftDeath("Natural Docs requires a newer version of File::Spec than you have. " . "You must either upgrade it or upgrade Perl."); }; }; ############################################################################### # Group: Path String Functions # # Function: CanonizePath # # Takes a path and returns a logically simplified version of it. # # Why oh why?: # # Because File::Spec->canonpath doesn't strip quotes on Windows. So if you pass in "a b\c" or "a b"\c, they still end up as # different strings even though they're logically the same. # # It also doesn't remove things like "..", so "a/b/../c" doesn't simplify to "a/c" like it should. # sub CanonizePath #(path) { my ($self, $path) = @_; if ($::OSNAME eq 'MSWin32') { # We don't have to use a smarter algorithm for dropping quotes because they're invalid characters for actual file and # directory names. $path =~ s/\"//g; }; $path = File::Spec->canonpath($path); # Condense a/b/../c into a/c. my $upDir = File::Spec->updir(); if (index($path, $upDir) != -1) { my ($volume, $directoryString, $file) = $self->SplitPath($path); my @directories = $self->SplitDirectories($directoryString); my $i = 1; while ($i < scalar @directories) { if ($i > 0 && $directories[$i] eq $upDir) { splice(@directories, $i - 1, 2); $i--; } else { $i++; }; }; $directoryString = $self->JoinDirectories(@directories); $path = $self->JoinPath($volume, $directoryString, $file); }; return $path; }; # # Function: PathIsAbsolute # # Returns whether the passed path is absolute. # sub PathIsAbsolute #(path) { my ($self, $path) = @_; return File::Spec->file_name_is_absolute($path); }; # # Function: JoinPath # # Creates a path from its elements. # # Parameters: # # volume - The volume, such as the drive letter on Windows. Undef if none. # dirString - The directory string. Create with if necessary. # file - The file name, or undef if none. # # Returns: # # The joined path. # sub JoinPath #(volume, dirString, $file) { my ($self, $volume, $dirString, $file) = @_; return File::Spec->catpath($volume, $dirString, $file); }; # # Function: JoinPaths # # Joins two paths. # # Parameters: # # basePath - May be a relative path, an absolute path, or undef. # extraPath - May be a relative path, a file, a relative path and file together, or undef. # noFileInExtra - Set this to true if extraPath is a relative path only, and doesn't have a file. # # Returns: # # The joined path. # # Why oh why?: # # Because nothing in File::Spec will simply slap two paths together. They have to be split up for catpath/file, and rel2abs # requires the base to be absolute. # sub JoinPaths #(basePath, extraPath, noFileInExtra) { my ($self, $basePath, $extraPath, $noFileInExtra) = @_; # If both are undef, it will return undef, which is what we want. if (!defined $basePath) { return $extraPath; } elsif (!defined $extraPath) { return $basePath; }; my ($baseVolume, $baseDirString, $baseFile) = File::Spec->splitpath($basePath, 1); my ($extraVolume, $extraDirString, $extraFile) = File::Spec->splitpath($extraPath, $noFileInExtra); my @baseDirectories = $self->SplitDirectories($baseDirString); my @extraDirectories = $self->SplitDirectories($extraDirString); my $fullDirString = $self->JoinDirectories(@baseDirectories, @extraDirectories); my $fullPath = File::Spec->catpath($baseVolume, $fullDirString, $extraFile); return $self->CanonizePath($fullPath); }; # # Function: SplitPath # # Takes a path and returns its elements. # # Parameters: # # path - The path to split. # noFile - Set to true if the path doesn't have a file at the end. # # Returns: # # The array ( volume, directoryString, file ). If any don't apply, they will be undef. Use to split the # directory string if desired. # # Why oh Why?: # # Because File::Spec->splitpath may leave a trailing slash/backslash/whatever on the directory string, which makes # it a bit hard to match it with results from File::Spec->catdir. # sub SplitPath #(path, noFile) { my ($self, $path, $noFile) = @_; my @segments = File::Spec->splitpath($path, $noFile); if (!length $segments[0]) { $segments[0] = undef; }; if (!length $segments[2]) { $segments[2] = undef; }; $segments[1] = File::Spec->catdir( File::Spec->splitdir($segments[1]) ); return @segments; }; # # Function: JoinDirectories # # Creates a directory string from an array of directory names. # # Parameters: # # directory - A directory name. There may be as many of these as desired. # sub JoinDirectories #(directory, directory, ...) { my ($self, @directories) = @_; return File::Spec->catdir(@directories); }; # # Function: SplitDirectories # # Takes a string of directories and returns an array of its elements. # # Why oh why?: # # Because File::Spec->splitdir might leave an empty element at the end of the array, which screws up both joining in # and navigation in . Morons. # sub SplitDirectories #(directoryString) { my ($self, $directoryString) = @_; my @directories = File::Spec->splitdir($directoryString); if (!length $directories[-1]) { pop @directories; }; return @directories; }; # # Function: MakeRelativePath # # Takes two paths and returns a relative path between them. # # Parameters: # # basePath - The starting path. May be relative or absolute, so long as the target path is as well. # targetPath - The target path. May be relative or absolute, so long as the base path is as well. # # If both paths are relative, they are assumed to be relative to the same base. # # Returns: # # The target path relative to base. # # Why oh why?: # # Wow, where to begin? First of all, there's nothing that gives a relative path between two relative paths. # # Second of all, if target and base are absolute but on different volumes, File::Spec->abs2rel creates a totally non-functional # relative path. It should return the target as is, since there is no relative path. # # Third of all, File::Spec->abs2rel between absolute paths on the same volume, at least on Windows, leaves the drive letter # on. So abs2rel('a:\b\c\d', 'a:\b') returns 'a:c\d' instead of the expected 'c\d'. That makes no fucking sense whatsoever. It's # not like it was designed to handle only directory names, either; the documentation says 'path' and the code seems to # explicitly handle it. There's just an 'unless' in there that tacks on the volume, defeating the purpose of a *relative* path and # making the function worthless. Morons. # # Update: This last one appears to be fixed in File::Spec 0.83, but that version isn't even listed on CPAN. Lovely. Apparently # it just comes with ActivePerl. Somehow I don't think most Linux users are using that. # sub MakeRelativePath #(basePath, targetPath) { my ($self, $basePath, $targetPath) = @_; my ($baseVolume, $baseDirString, $baseFile) = $self->SplitPath($basePath, 1); my ($targetVolume, $targetDirString, $targetFile) = $self->SplitPath($targetPath); # If the volumes are different, there is no possible relative path. if ($targetVolume ne $baseVolume) { return $targetPath; }; my @baseDirectories = $self->SplitDirectories($baseDirString); my @targetDirectories = $self->SplitDirectories($targetDirString); # Skip the parts of the path that are the same. while (scalar @baseDirectories && @targetDirectories && $baseDirectories[0] eq $targetDirectories[0]) { shift @baseDirectories; shift @targetDirectories; }; # Back out of the base path until it reaches where they were similar. for (my $i = 0; $i < scalar @baseDirectories; $i++) { unshift @targetDirectories, File::Spec->updir(); }; $targetDirString = $self->JoinDirectories(@targetDirectories); return File::Spec->catpath(undef, $targetDirString, $targetFile); }; # # Function: IsSubPathOf # # Returns whether the path is a descendant of another path. # # Parameters: # # base - The base path to test against. # path - The possible subpath to test. # # Returns: # # Whether path is a descendant of base. # sub IsSubPathOf #(base, path) { my ($self, $base, $path) = @_; # This is a quick test that should find a false quickly. if ($base eq substr($path, 0, length($base))) { # This doesn't guarantee true, because it could be "C:\A B" and "C:\A B C\File". So we test for it by seeing if the last # directory in base is the same as the equivalent directory in path. my ($baseVolume, $baseDirString, $baseFile) = NaturalDocs::File->SplitPath($base, 1); my @baseDirectories = NaturalDocs::File->SplitDirectories($baseDirString); my ($pathVolume, $pathDirString, $pathFile) = NaturalDocs::File->SplitPath($path); my @pathDirectories = NaturalDocs::File->SplitDirectories($pathDirString); return ( $baseDirectories[-1] eq $pathDirectories[ scalar @baseDirectories - 1 ] ); } else { return undef; }; }; # # Function: ConvertToURL # # Takes a relative path and converts it from the native format to a relative URL. Note that it _doesn't_ convert special characters # to amp chars. # sub ConvertToURL #(path) { my ($self, $path) = @_; my ($pathVolume, $pathDirString, $pathFile) = $self->SplitPath($path); my @pathDirectories = $self->SplitDirectories($pathDirString); my $i = 0; while ($i < scalar @pathDirectories && $pathDirectories[$i] eq File::Spec->updir()) { $pathDirectories[$i] = '..'; $i++; }; return join('/', @pathDirectories, $pathFile); }; # # Function: NoUpwards # # Takes an array of directory entries and returns one without all the entries that refer to the parent directory, such as '.' and '..'. # sub NoUpwards #(array) { my ($self, @array) = @_; return File::Spec->no_upwards(@array); }; # # Function: NoFileName # # Takes a path and returns a version without the file name. Useful for sending paths to . # sub NoFileName #(path) { my ($self, $path) = @_; my ($pathVolume, $pathDirString, $pathFile) = File::Spec->splitpath($path); return File::Spec->catpath($pathVolume, $pathDirString, undef); }; # # Function: ExtensionOf # # Returns the extension of the passed path, or undef if none. # sub ExtensionOf #(path) { my ($self, $path) = @_; my ($pathVolume, $pathDirString, $pathFile) = File::Spec->splitpath($path); # We need the leading dot in the regex so files that start with a dot but don't have an extension count as extensionless files. if ($pathFile =~ /.\.([^\.]+)$/) { return $1; } else { return undef; }; }; # # Function: IsCaseSensitive # # Returns whether the current platform has case-sensitive paths. # sub IsCaseSensitive { return !(File::Spec->case_tolerant()); }; ############################################################################### # Group: Disk Functions # # Function: CreatePath # # Creates a directory tree corresponding to the passed path, regardless of how many directories do or do not already exist. # Do _not_ include a file name in the path. Use first if you need to. # sub CreatePath #(path) { my ($self, $path) = @_; File::Path::mkpath($path); }; # # Function: RemoveEmptyTree # # Removes an empty directory tree. The passed directory will be removed if it's empty, and it will keep removing its parents # until it reaches one that's not empty or a set limit. # # Parameters: # # path - The path to start from. It will try to remove this directory and work it's way down. # limit - The path to stop at if it doesn't find any non-empty directories first. This path will *not* be removed. # sub RemoveEmptyTree #(path, limit) { my ($self, $path, $limit) = @_; my ($volume, $directoryString) = $self->SplitPath($path, 1); my @directories = $self->SplitDirectories($directoryString); my $directory = $path; while (-d $directory && $directory ne $limit) { opendir FH_ND_FILE, $directory; my @entries = readdir FH_ND_FILE; closedir FH_ND_FILE; @entries = $self->NoUpwards(@entries); if (scalar @entries || !rmdir($directory)) { last; }; pop @directories; $directoryString = $self->JoinDirectories(@directories); $directory = $self->JoinPath($volume, $directoryString); }; }; # # Function: Copy # # Copies a file from one path to another. If the destination file exists, it is overwritten. # # Parameters: # # source - The file to copy. # destination - The destination to copy to. # sub Copy #(source, destination) { my ($self, $source, $destination) = @_; File::Copy::copy($source, $destination); }; 1;