use URI::file for (de)serialising paths in XML

MANIFEST

@@ -987,6 +987,9 @@ t/complex/aastex631_deluxetable.xml
 t/complex/aliceblog.pdf
 t/complex/aliceblog.tex
 t/complex/aliceblog.xml
+t/complex/candidates.pdf
+t/complex/candidates.tex
+t/complex/candidates.xml
 t/complex/cleveref_minimal.pdf
 t/complex/cleveref_minimal.tex
 t/complex/cleveref_minimal.xml
@@ -1016,6 +1019,8 @@ t/complex/si.pdf
 t/complex/si.tex
 t/complex/si.xml
 t/complex/sunset.jpg
+t/complex/sunset,encoded.gif
+t/complex/sunset,encoded.jpg
 t/complex/tcilatex_minimal.pdf
 t/complex/tcilatex_minimal.tex
 t/complex/tcilatex_minimal.xml

lib/LaTeXML/Common/Model/DTD.pm

@@ -16,7 +16,6 @@ use LaTeXML::Util::Pathname;
 use LaTeXML::Global;
 use LaTeXML::Common::Error;
 use LaTeXML::Common::XML;
-use URI::file;
 
 #**********************************************************************
 # NOTE: Arglist is DTD specific.
@@ -118,7 +117,7 @@ sub readDTD {
       paths               => $STATE->lookupValue('SEARCHPATHS'),
       installation_subdir => 'resources/DTD');
     if ($dtdfile) {
-      $dtd = XML::LibXML::Dtd->new($$self{public_id}, URI::file->new($dtdfile));
+      $dtd = XML::LibXML::Dtd->new($$self{public_id}, pathname_to_file_url($dtdfile));
       $how = " from $dtdfile ";
       Error('misdefined', $$self{system_id}, undef,
         "Parsing of DTD \"$$self{public_id}\" \"$$self{system_id}\" failed")

lib/LaTeXML/Core.pm

@@ -263,7 +263,7 @@ sub convertDocument {
           my @dedup = ();
           while (my $check = shift(@copy)) {
             unshift(@dedup, $check) if !(grep { $_ eq $check } @dedup); }
-          $document->insertPI('latexml', searchpaths => join(',', @dedup)); } }
+          $document->insertPI('latexml', searchpaths => pathname_to_urls(@dedup)); } }
       foreach my $preload_by_reference (@{ $$self{preload} }) {
         my $preload = $preload_by_reference; # copy preload value, as we want to preserve the hash as-is, for (potential) future daemon calls
         next if $preload =~ /\.pool$/;

lib/LaTeXML/Engine/TeX_FileIO.pool.ltxml

@@ -243,7 +243,7 @@ DefConstructor('\lx@special@graphics OptionalKeyVals:SpecialPS Semiverbatim',
         my $bottom = -$voffset;
         $options .= "," if $options;
         $options .= "trim=$left $bottom 0 0,clip=true"; } }
-    (options => $options, path => $path, candidates => join(',', @candidates)); },
+    (options => $options, path => pathname_to_url($path), candidates => pathname_to_urls(@candidates)); },
   mode => 'restricted_horizontal');
 # Since these ultimately generate external resources, it can be useful to have a handle on them.
 Tag('ltx:graphics', afterOpen => sub { GenerateID(@_, 'g'); });

lib/LaTeXML/Package/epsf.sty.ltxml

@@ -57,8 +57,8 @@ DefConstructor('\epsfbox[] Semiverbatim',
       $options .= ($options ? ',' : '') . 'width=' . ToString($w); }
     if ($h->valueOf > 0) {
       $options .= ($options ? ',' : '') . 'height=' . ToString($h); }
-    (graphic => $file,
-      candidates => join(',', @candidates),
+    (graphic => pathname_to_url($file),
+      candidates => pathname_to_urls(@candidates),
       options    => $options); });
 
 Let('\epsfgetlitbb', '\epsfbox');

lib/LaTeXML/Package/epsfig.sty.ltxml

@@ -47,9 +47,9 @@ DefConstructor('\psfig RequiredKeyVals:epsGin',
     if (my $base = LookupValue('SOURCEDIRECTORY')) {
       @candidates = map { pathname_relative($_, $base) } @candidates; }
     my $options = ToString($kv);
-    (graphic => $file,
-      candidates => join(',', @candidates),
-      options => $options); });
+    (graphic => pathname_to_url($file),
+      candidates => pathname_to_urls(@candidates),
+      options    => $options); });
 
 Let('\epsfig', '\psfig');
 

lib/LaTeXML/Package/graphics.sty.ltxml

@@ -248,7 +248,7 @@ DefConstructor('\reflectbox {}',
 DefConstructor('\graphicspath DirectoryList', sub {
     my ($document, $paths, %props) = @_;
     foreach my $path (@{ $props{paths} }) {
-      $document->insertPI('latexml', graphicspath => $path); } },
+      $document->insertPI('latexml', graphicspath => pathname_to_url($path)); } },
   properties => sub {
     my ($stomach, $paths) = @_;
     my @paths = ();
@@ -297,8 +297,8 @@ DefConstructor('\@includegraphics OptionalMatch:* [][] Semiverbatim',
     my ($stomach, $starred, $op1, $op2, $graphic) = @_;
     my $options = graphicS_options($starred, $op1, $op2);
     my ($path, @candidates) = image_candidates(ToString($graphic));
-    (graphic => $path,
-      candidates => join(',', @candidates),
+    (graphic => pathname_to_url($path),
+      candidates => pathname_to_urls(@candidates),
       options    => $options); },
   alias => '\includegraphics');
 

lib/LaTeXML/Package/graphicx.sty.ltxml

@@ -65,8 +65,8 @@ DefConstructor('\@includegraphicx OptionalMatch:* OptionalKeyVals:Gin Semiverbat
     my $options = graphicX_options($starred, $kv);
     my ($path, @candidates) = image_candidates(ToString($graphic));
     my $alt = $kv && $kv->getValue('alt');
-    (path => $path,
-      candidates => join(',', @candidates),
+    (path => pathname_to_url($path),
+      candidates => pathname_to_urls(@candidates),
       (defined $alt ? (alt => $alt) : ()),
       options => $options); },
   mode => 'restricted_horizontal', enterHorizontal => 1);

lib/LaTeXML/Package/pgfsys-latexml.def.ltxml

@@ -617,8 +617,8 @@ DefConstructor('\lxSVG@includegraphics{}{} Semiverbatim',
       paths => LookupValue('GRAPHICSPATHS'));
     if (my $base = LookupValue('SOURCEDIRECTORY')) {
       @candidates = map { pathname_relative($_, $base) } @candidates; }
-    (graphic => $graphic,
-      candidates => join(',', @candidates),
+    (graphic => pathname_to_url($graphic),
+      candidates => pathname_to_urls(@candidates),
       options    => $options); },
   alias => '\includegraphics');
 

lib/LaTeXML/Package/psfrag.sty.ltxml

@@ -87,8 +87,8 @@ DefConstructor('\lx@psfrag@includegraphics OptionalMatch:* [][] Semiverbatim',
     my $options = graphicS_options($starred, $op1, $op2);
     my ($path, @candidates) = image_candidates(ToString($graphic));
     my $ifpicture = psfrag_requirepicture(@candidates);
-    (graphic => $path,
-      candidates => join(',', @candidates),
+    (path => pathname_to_url($path),
+      candidates => pathname_to_urls(@candidates),
       options    => $options,
       pic        => $ifpicture); },
   afterDigest => sub {
@@ -110,8 +110,8 @@ DefConstructor('\lx@psfrag@includegraphicx OptionalMatch:* OptionalKeyVals:Gin S
     my $options = graphicX_options($starred, $kv);
     my ($path, @candidates) = image_candidates(ToString($graphic));
     my $ifpicture = psfrag_requirepicture(@candidates);
-    (path => $path,
-      candidates => join(',', @candidates),
+    (path => pathname_to_url($path),
+      candidates => pathname_to_urls(@candidates),
       options    => $options,
       pic        => $ifpicture); },
   afterDigest => sub {
@@ -133,8 +133,8 @@ DefConstructor('\lx@psfrag@epsfbox[] Semiverbatim',
     my $options = ($clip ? ($bb ? "viewport=$bb, clip" : "clip") : '');
     my ($path, @candidates) = image_candidates(ToString($graphic));
     my $ifpicture = psfrag_requirepicture(@candidates);
-    (graphic => $path,
-      candidates => join(',', @candidates),
+    (graphic => pathname_to_url($path),
+      candidates => pathname_to_urls(@candidates),
       options    => $options,
       pic        => $ifpicture); },
   afterDigest => sub {

lib/LaTeXML/Post.pm

@@ -664,7 +664,6 @@ use DB_File;
 use Unicode::Normalize;
 use LaTeXML::Post;    # to import error handling...
 use LaTeXML::Common::Error;
-use URI::file;
 use base qw(LaTeXML::Common::Object);
 our $NSURI = "http://dlmf.nist.gov/LaTeXML";
 our $XPATH = LaTeXML::Common::XML::XPath->new(ltx => $NSURI);
@@ -821,7 +820,7 @@ sub setDocument_internal {
     @paths = @{ $$self{searchpaths} } if $$self{searchpaths};
     foreach my $pi (@{ $$self{processingInstructions} }) {
       if ($pi =~ /^\s*searchpaths\s*=\s*([\"\'])(.*?)\1\s*$/) {
-        push(@paths, split(',', $2)); } }
+        push(@paths, pathname_from_urls($2)); } }
     ### No, this ultimately can be the xml source, which may be the destination;
     ### adding this gets the wrong graphics (already processed!)
     ### push(@paths, pathname_absolute($$self{sourceDirectory})) if $$self{sourceDirectory};
@@ -857,7 +856,8 @@ sub setDocument_internal {
   else {
     Fatal('unexpected', $root, undef, "Dont know how to use '$root' as document element"); }
   # set URI to destination
-  $$self{document}->setURI(URI::file->new($self->getDestination));
+  my $url = pathname_to_file_url($self->getDestination);
+  $$self{document}->setURI($url) if defined $url;
   return $self; }
 
 our @MonthNames = (qw( January February March April May June

lib/LaTeXML/Post/CrossRef.pm

@@ -650,8 +650,9 @@ sub generateURL {
     if ($location = $object->getValue('location')) {
       my $doclocation = $doc->siteRelativeDestination;
       my $pathdir     = pathname_directory($doclocation);
-      my $url         = pathname_relative(($location =~ m|^/| ? $location : '/' . $location),
+      my $relpath     = pathname_relative(($location =~ m|^/| ? $location : '/' . $location),
         ($pathdir =~ m|^/| ? $pathdir : '/' . $pathdir));
+      my $url       = pathname_to_url($relpath);
       my $extension = $$self{extension} || 'xml';
       my $urlstyle  = $$self{urlstyle}  || 'file';
       if ($urlstyle eq 'server') {
@@ -667,7 +668,7 @@ sub generateURL {
         $url .= '#' . $fragid; }
       elsif ($location eq $doclocation) {
         $url = ''; }
-      return pathname_to_url($url); }
+      return $url; }
     else {
       $self->note_missing('warn', 'File location for ID', $id); } }
   else {

lib/LaTeXML/Post/Graphics.pm

@@ -53,7 +53,7 @@ sub new {
   $$self{trivial_scaling} = $options{trivial_scaling} || 1;
   $$self{graphics_types}  = $options{graphics_types}
     || [qw(svg png gif jpg jpeg
-      eps ps postscript ai pdf)];
+    eps ps postscript ai pdf)];
   $$self{type_properties} = $options{type_properties}
     || {
     ai => { destination_type => 'png',
@@ -123,7 +123,7 @@ sub findGraphicsPaths {
   my @paths = ();
   foreach my $pi ($doc->findnodes('.//processing-instruction("latexml")')) {
     if ($pi->textContent =~ /^\s*graphicspath\s*=\s*([\"\'])(.*?)\1\s*$/) {
-      push(@paths, $2); } }
+      push(@paths, pathname_from_url($2)); } }
   return @paths; }
 
 sub getGraphicsSourceTypes {
@@ -133,9 +133,9 @@ sub getGraphicsSourceTypes {
 # Return the pathname to an appropriate image.
 sub findGraphicFile {
   my ($self, $doc, $node) = @_;
-  if (my $source = $node->getAttribute('graphic')) {
+  if (my $source = pathname_from_url($node->getAttribute('graphic'))) {
     # if we already have a usable candidate, save ourselves the work
-    my @paths = grep { -e $_ } split(',', $node->getAttribute('candidates') || '');
+    my @paths = grep { -e $_ } pathname_from_urls($node->getAttribute('candidates') || '');
     if (!scalar(@paths)) {
       # Find all acceptable image files, in order of search paths
       my ($dir, $name, $reqtype) = pathname_split($source);

lib/LaTeXML/Post/Manifest/Epub.pm

@@ -13,7 +13,7 @@ package LaTeXML::Post::Manifest::Epub;
 use strict;
 use warnings;
 use File::Find qw(find);
-use URI::file;
+use LaTeXML::Util::Pathname;
 
 our $uuid_tiny_installed;
 
@@ -27,7 +27,7 @@ BEGIN {
 use base qw(LaTeXML::Post::Manifest);
 use LaTeXML::Util::Pathname;
 use File::Spec::Functions qw(catdir);
-use POSIX qw(strftime);
+use POSIX                 qw(strftime);
 use LaTeXML::Post;    # for error handling!
 our $container_content = <<'EOL';
 <?xml version="1.0"?>
@@ -171,7 +171,7 @@ sub process {
       # Add to manifest
       my $manifest = $$self{opf_manifest};
       my $item     = $manifest->addNewChild(undef, 'item');
-      my $item_url = URI::file->new($relative_destination);
+      my $item_url = pathname_to_url($relative_destination);
       my $item_id  = url_id($item_url);
       $item->setAttribute('id',         $item_id);
       $item->setAttribute('href',       $item_url);
@@ -215,7 +215,7 @@ sub finalize {
       $file_type = 'application/octet-stream'; }
 
     my $file_item = $manifest->addNewChild(undef, 'item');
-    my $file_url  = URI::file->new($file);
+    my $file_url  = pathname_to_url($file);
     $file_item->setAttribute('id',         url_id($file_url));
     $file_item->setAttribute('href',       $file_url);
     $file_item->setAttribute('media-type', $file_type); }

lib/LaTeXML/Util/Image.pm

@@ -252,7 +252,7 @@ sub image_graphicx_sizer {
   my ($whatsit) = @_;
   if (my $candidates = $whatsit->getProperty('candidates')) {
     my $dpi = ($STATE && $STATE->lookupValue('DPI')) || $DPI;
-    foreach my $source (split(/,/, $candidates)) {
+    foreach my $source (pathname_from_urls($candidates)) {
       if (!pathname_is_absolute($source)) {
         if (my $base = $STATE->lookupValue('SOURCEDIRECTORY')) {
           $source = pathname_concat($base, $source); } }

lib/LaTeXML/Util/Pathname.pm

@@ -31,6 +31,7 @@ use warnings;
 use File::Spec;
 use File::Copy;
 use File::Which;
+use URI::file;
 use Cwd;
 use base qw(Exporter);
 our @EXPORT = qw( &pathname_find &pathname_findall &pathname_kpsewhich
@@ -39,6 +40,7 @@ our @EXPORT = qw( &pathname_find &pathname_findall &pathname_kpsewhich
   &pathname_timestamp
   &pathname_concat
   &pathname_relative &pathname_absolute &pathname_to_url
+  &pathname_to_urls &pathname_from_url &pathname_from_urls &pathname_to_file_url
   &pathname_is_absolute &pathname_is_contained
   &pathname_is_url &pathname_is_literaldata
   &pathname_is_raw
@@ -212,13 +214,34 @@ sub pathname_absolute {
     ? File::Spec->rel2abs($pathname, ($base ? pathname_canonical($base) : pathname_cwd()))
     : $pathname); }
 
+# conversions to/from file URLs
+
+# remove file:// part for backwards compatibility
 sub pathname_to_url {
-  my ($pathname) = @_;
-  return unless defined $pathname;
-  my $relative_pathname = pathname_relative($pathname);
-  if ($SEP ne '/') {
-    $relative_pathname = join('/', split(/\Q$SEP\E/, $relative_pathname)); }
-  return $relative_pathname; }
+  return unless defined $_[0];
+  my $url = URI::file->new(pathname_canonical($_[0]))->path;
+  $url =~ s/,/%2C/g;
+  return $url; }
+
+sub pathname_to_urls {
+  return join(',', map { pathname_to_url($_) } @_); }
+
+sub pathname_from_url {
+  return unless defined $_[0];
+  return URI->new($_[0], 'file')->file; }
+
+sub pathname_from_urls {
+  return unless defined $_[0];
+  my @urls     = split(/,/, $_[0]);
+  my @nonempty = grep { $_ } @urls;
+  return map { pathname_from_url($_) } @nonempty; }
+
+# complete file:// URL
+sub pathname_to_file_url {
+  return unless defined $_[0];
+  my $url = URI::file->new(pathname_canonical($_[0]))->as_string;
+  $url =~ s/,/%2C/g;
+  return $url; }
 
 #======================================================================
 # Actual file system operations.
@@ -377,6 +400,7 @@ sub candidate_pathnames {
     opendir(my $dir_handle, $dir) or next;
     my @dir_files = readdir($dir_handle);
     closedir($dir_handle);
+    my @dir_paths;
     for my $local_file (@dir_files) {
       for my $regex_pair (@regexes) {
         my ($i_regex, $regex) = @$regex_pair;
@@ -386,7 +410,8 @@ sub candidate_pathnames {
           if ($local_file =~ m/$regex/) {
             # if we are only interested in the first match, return it:
             return ($full_file) if $options{findfirst};
-            push(@paths, $full_file); } } } } }
+            push(@dir_paths, $full_file); } } } }
+    push(@paths, sort @dir_paths); }
   # Fallback: if no strict matches were found, return any existing case-insensitive matches
   # Defer the -f check until we are sure we need it, to keep the usual cases fast.
   return @paths ? @paths : @nocase_paths; }
@@ -553,10 +578,26 @@ Returns the absolute pathname resulting from interpretting
 C<$path> relative to the directory C<$base>.  If C<$path>
 is already absolute, it is returned unchanged.
 
-=item C<< $relative_url = pathname_to_url($path); >>
+=item C<< $url = pathname_to_file_url($path); >>
+
+Creates a file URL for a given pathname.
+
+=item C<< $url = pathname_to_url($path); >>
+
+Creates a file URL for a given pathname without the file:// scheme.
+
+=item C<< $url = pathname_to_urls(@paths); >>
+
+Create a comma separated list of file URLs for a given list of pathnames.
+
+=item C<< $path = pathname_from_url($url); >>
+
+Create a pathname, with local OS path separators, from a file URL.
+
+=item C<< @paths = pathname_from_urls($urls); >>
 
-Creates a local, relative URL for a given pathname,
-also ensuring proper path separators on non-Unix systems.
+Create a list of pathnames, with local OS path separator, from a
+comma separated list of file URLs.
 
 =back
 

lib/LaTeXML/resources/RelaxNG/LaTeXML-misc.rnc

@@ -82,12 +82,12 @@ graphics_attributes =
   ## the path to the graphics file. This is the (often minimally specified) path
   ## to a graphics file omitting the type extension.  Once resolved to a specific
   ## image file, the \attr{imagesrc} (from Imageable.attributes) is used.
-  attribute graphic { text }?,
+  attribute graphic { xsd:anyURI }?,
 
   ## a comma separated list of candidate graphics files that could be used to
   ## for \attr{graphic}.  A post-processor or application may choose from these,
   ## or may make its own selection or synthesis to implement the graphic for a given target.
-  attribute candidates { text }?,
+  attribute candidates { xsd:anyURI, (",", xsd:anyURI)* }?,
 
   ## an encoding of the scaling and positioning options
   ## to be used in processing the graphic.