hexapdf 0.18.0 → 0.19.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: d95ce1575c017f44b2c0f96e7e5a927b8c4f8c3adf6aa0f3a7dc983c5dfa77a8
-  data.tar.gz: e49a23655e5ce4f4ded50c5ac0c90d7892c41bd526dbdfe72ad85cca4891098b
+  metadata.gz: 85c063a63af9729acc10a54ef53fba69d73b75f1c06ff0e6de246df920e17dd9
+  data.tar.gz: dcea10d0ccfe66282c92e6bb41c1d57927d525e1618753d598a910546c8a8e82
 SHA512:
-  metadata.gz: 37e3b09a059bb7875c797f50f60642b080f20a081fa2540f74837ac51a8cb9c1aa5b93e6502dd9f242d178f68ea654c0619b2e55d4cb934e36badca4a5057d1c
-  data.tar.gz: cd9fc830b8b4f5387478d7e6c32260672a9332bb2faf3aa7afcb4d71bf7c39481616f7f33d731229a6099e44dac84bbecd651c3618593f6c7ba872b9e958a3cb
+  metadata.gz: 4f1d468375c4ce336e55a09d897de9a8c95babcfa1b4441cc04a9dc712435c64906016647baa030f5695f9434d18214c30bca746f245b53a69321139f63256b9
+  data.tar.gz: 1f1895ca7ad46bae1bf33790d4c8daa7f72adfcc00cc26e1611b7ece64d7a5a76e7f1e06be4022bb58f5dbd84154e9846cb4feb94f727c9abe99f5b7c8b87fcf

data/CHANGELOG.md

@@ -1,3 +1,47 @@
+## 0.19.3 - 2021-12-14
+
+### Fixed
+
+* Handling of invalid files where the "startxref" keyword and its value are on
+  the same line
+
+
+## 0.19.2 - 2021-12-14
+
+### Fixed
+
+* Set the trailer's ID field to an array of two empty strings when decrypting in
+  case it is missing
+* Incremental writing when one of the existing revisions contains a
+  cross-reference stream
+
+
+## 0.19.1 - 2021-12-12
+
+### Added
+
+* [HexaPDF::Type::FontType3#bounding_box] to fix content stream processing error
+
+### Fixed
+
+* Calculation of scaled font size for [HexaPDF::Content::GraphicsState] and
+  [HexaPDF::Layout::Style] when Type3 fonts are used
+
+
+## 0.19.0 - 2021-11-24
+
+### Added
+
+* Page resource pruning to the optimization task
+* An option for page resources pruning to the optimization options of the
+  `hexapdf` command
+
+### Fixed
+
+* Handling of invalid date strings with a minute time zone offset greater than
+  59
+
+
 ## 0.18.0 - 2021-11-04
 
 ### Added
@@ -6,7 +50,7 @@
   device colors in parts other than the canvas
 * [HexaPDF::Type::AcroForm::VariableTextField::create_appearance_string] for
   centralized creation of appearance strings
-* [HexaPDF::Object.make_direct] for making objects and all parts of them direct
+* [HexaPDF::Object::make_direct] for making objects and all parts of them direct
   instead of indirect
 
 ### Changed
@@ -26,7 +70,7 @@
   dictionary are indirect objects
 * [HexaPDF::Content::GraphicObject::EndpointArc] to correctly determine the
   start and end points
-* [HexaPDF::Dictionary#perform_validation] to correctly handle objects that
+* HexaPDF::Dictionary#perform_validation to correctly handle objects that
   should not be indirect objects
 
 

data/lib/hexapdf/cli/command.rb

@@ -66,6 +66,7 @@ module HexaPDF
         @out_options.xref_streams = :preserve
         @out_options.streams = :preserve
         @out_options.optimize_fonts = false
+        @out_options.prune_page_resources = false
 
         @out_options.encryption = :preserve
         @out_options.enc_user_pwd = @out_options.enc_owner_pwd = nil
@@ -169,6 +170,10 @@ module HexaPDF
                    "time; default: #{@out_options.compress_pages})") do |c|
           @out_options.compress_pages = c
         end
+        options.on("--[no-]prune-page-resources", "Prunes unused objects from the page resources " \
+                   "(may take a long time; default: #{@out_options.prune_page_resources})") do |c|
+          @out_options.prune_page_resources = c
+        end
         options.on("--[no-]optimize-fonts", "Optimize embedded font files; " \
                    "default: #{@out_options.optimize_fonts})") do |o|
           @out_options.optimize_fonts = o
@@ -236,7 +241,8 @@ module HexaPDF
         doc.task(:optimize, compact: @out_options.compact,
                  object_streams: @out_options.object_streams,
                  xref_streams: @out_options.xref_streams,
-                 compress_pages: @out_options.compress_pages)
+                 compress_pages: @out_options.compress_pages,
+                 prune_page_resources: @out_options.prune_page_resources)
         if @out_options.streams != :preserve || @out_options.optimize_fonts
           doc.each(only_current: false) do |obj|
             optimize_stream(obj)

data/lib/hexapdf/content/canvas.rb

@@ -589,7 +589,7 @@ module HexaPDF
       #
       # The line cap style specifies how the ends of stroked open paths should look like.
       #
-      # The +style+ parameter can be one of:
+      # The +style+ parameter can be one of (also see LineCapStyle):
       #
       # :butt or 0::
       #     Stroke is squared off at the endpoint of a path.
@@ -641,7 +641,7 @@ module HexaPDF
       #
       # The line join style specifies the shape that is used at the corners of stroked paths.
       #
-      # The +style+ parameter can be one of:
+      # The +style+ parameter can be one of (also see LineJoinStyle):
       #
       # :miter or 0::
       #     The outer lines of the two segments continue until the meet at an angle.

data/lib/hexapdf/content/graphics_state.rb

@@ -73,7 +73,7 @@ module HexaPDF
     end
 
     # Defines all available line cap styles as constants. Each line cap style is an instance of
-    # NamedValue. For use with Content::GraphicsState#line_cap_style.
+    # NamedValue, see ::normalize. For use with e.g. Content::Canvas#line_cap_style.
     #
     # See: PDF1.7 s8.4.3.3
     module LineCapStyle
@@ -95,18 +95,39 @@ module HexaPDF
       end
 
       # Stroke is squared off at the endpoint of a path.
+      #
+      # Specify as 0 or :butt.
+      #
+      #   #>pdf-small-hide
+      #   canvas.line_cap_style(:butt)
+      #   canvas.line_width(10).line(50, 20, 50, 80).stroke
+      #   canvas.stroke_color("white").line_width(1).line(50, 20, 50, 80).stroke
       BUTT_CAP = NamedValue.new(:butt, 0)
 
       # A semicircular arc is drawn at the endpoint of a path.
+      #
+      # Specify as 1 or :round.
+      #
+      #   #>pdf-small-hide
+      #   canvas.line_cap_style(:round)
+      #   canvas.line_width(10).line(50, 20, 50, 80).stroke
+      #   canvas.stroke_color("white").line_width(1).line(50, 20, 50, 80).stroke
       ROUND_CAP = NamedValue.new(:round, 1)
 
       # The stroke continues half the line width beyond the endpoint of a path.
+      #
+      # Specify as 2 or :projecting_square.
+      #
+      #   #>pdf-small-hide
+      #   canvas.line_cap_style(:projecting_square)
+      #   canvas.line_width(10).line(50, 20, 50, 80).stroke
+      #   canvas.stroke_color("white").line_width(1).line(50, 20, 50, 80).stroke
       PROJECTING_SQUARE_CAP = NamedValue.new(:projecting_square, 2)
 
     end
 
     # Defines all available line join styles as constants. Each line join style is an instance of
-    # NamedValue. For use with Content::GraphicsState#line_join_style.
+    # NamedValue, see ::normalize For use with e.g. Content::Canvas#line_join_style.
     #
     # See: PDF1.7 s8.4.3.4
     module LineJoinStyle
@@ -127,20 +148,47 @@ module HexaPDF
         end
       end
 
-      # The outer lines of the two segments continue until the meet at an angle.
+      # The outer lines of the two segments continue until they meet at an angle.
+      #
+      # Specify as 0 or :miter.
+      #
+      #   #>pdf-small-hide
+      #   canvas.line_join_style(:miter)
+      #   canvas.line_width(10).
+      #     polyline(20, 20, 50, 80, 80, 20).stroke
+      #   canvas.stroke_color("white").line_width(1).line_join_style(:bevel).
+      #     polyline(20, 20, 50, 80, 80, 20).stroke
       MITER_JOIN = NamedValue.new(:miter, 0)
 
       # An arc of a circle is drawn around the point where the segments meet.
+      #
+      # Specify as 1 or :round.
+      #
+      #   #>pdf-small-hide
+      #   canvas.line_join_style(:round)
+      #   canvas.line_width(10).
+      #     polyline(20, 20, 50, 80, 80, 20).stroke
+      #   canvas.stroke_color("white").line_width(1).line_join_style(:bevel).
+      #     polyline(20, 20, 50, 80, 80, 20).stroke
       ROUND_JOIN = NamedValue.new(:round, 1)
 
-      # The two segments are finished with butt caps and the space between the ends is filled with
-      # a triangle.
+      # The two segments are finished with butt caps and the space between the ends is filled with a
+      # triangle.
+      #
+      # Specify as 2 or :bevel.
+      #
+      #   #>pdf-small-hide
+      #   canvas.line_join_style(:bevel)
+      #   canvas.line_width(10).
+      #     polyline(20, 20, 50, 80, 80, 20).stroke
+      #   canvas.stroke_color("white").line_width(1).line_join_style(:bevel).
+      #     polyline(20, 20, 50, 80, 80, 20).stroke
       BEVEL_JOIN = NamedValue.new(:bevel, 2)
 
     end
 
-    # The line dash pattern defines how a line should be dashed. For use with
-    # Content::GraphicsState#line_dash_pattern.
+    # The line dash pattern defines how a line should be dashed. For use with e.g.
+    # Content::Canvas#line_dash_pattern.
     #
     # A dash pattern consists of two parts: the dash array and the dash phase. The dash array
     # defines the length of alternating dashes and gaps (important: starting with dashes). And the
@@ -159,6 +207,12 @@ module HexaPDF
     # See: PDF1.7 s8.4.3.6
     class LineDashPattern
 
+      # :call-seq:
+      #   LineDashPattern.normalize(line_dash_pattern)         -> line_dash_pattern
+      #   LineDashPattern.normalize(array, phase = 0)          -> LineDashPattern.new(array, phase)
+      #   LineDashPattern.normalize(number, phase = 0)         -> LineDashPattern.new([number], phase)
+      #   LineDashPattern.normalize(0)                         -> LineDashPattern.new
+      #
       # Returns the arguments normalized to a valid LineDashPattern instance.
       #
       # If +array+ is 0, the default line dash pattern representing a solid line will be used. If it
@@ -206,8 +260,8 @@ module HexaPDF
 
     end
 
-    # Defines all available rendering intents as constants. For use with
-    # Content::GraphicsState#rendering_intent.
+    # Defines all available rendering intents as constants. For use with e.g.
+    # Content::Canvas#rendering_intent.
     #
     # See: PDF1.7 s8.6.5.8
     module RenderingIntent
@@ -241,7 +295,7 @@ module HexaPDF
     end
 
     # Defines all available text rendering modes as constants. Each text rendering mode is an
-    # instance of NamedValue. For use with Content::GraphicsState#text_rendering_mode.
+    # instance of NamedValue. For use with e.g. Content::Canvas#text_rendering_mode.
     #
     # See: PDF1.7 s9.3.6
     module TextRenderingMode
@@ -272,28 +326,97 @@ module HexaPDF
         end
       end
 
-      # Fill text
+      # Fill text.
+      #
+      # Specify as 0 or :fill.
+      #
+      #   #>pdf-small-hide
+      #   canvas.font("Helvetica", size: 13)
+      #   canvas.stroke_color("green").line_width(0.5)
+      #   canvas.text_rendering_mode(:fill)
+      #   canvas.text("#{canvas.text_rendering_mode.name}", at: [10, 50])
       FILL = NamedValue.new(:fill, 0)
 
-      # Stroke text
+      # Stroke text.
+      #
+      # Specify as 1 or :stroke.
+      #
+      #   #>pdf-small-hide
+      #   canvas.font("Helvetica", size: 13)
+      #   canvas.stroke_color("green").line_width(0.5)
+      #   canvas.text_rendering_mode(:stroke)
+      #   canvas.text("#{canvas.text_rendering_mode.name}", at: [10, 50])
       STROKE = NamedValue.new(:stroke, 1)
 
-      # Fill, then stroke text
+      # Fill, then stroke text.
+      #
+      # Specify as 2 or :fill_stroke.
+      #
+      #   #>pdf-small-hide
+      #   canvas.font("Helvetica", size: 13)
+      #   canvas.stroke_color("green").line_width(0.5)
+      #   canvas.text_rendering_mode(:fill_stroke)
+      #   canvas.text("#{canvas.text_rendering_mode.name}", at: [10, 50])
       FILL_STROKE = NamedValue.new(:fill_stroke, 2)
 
-      # Neither fill nor stroke text (invisible)
+      # Neither fill nor stroke text (invisible).
+      #
+      # Specify as 3 or :invisible.
+      #
+      #   #>pdf-small-hide
+      #   canvas.font("Helvetica", size: 13)
+      #   canvas.stroke_color("green").line_width(0.5)
+      #   canvas.text_rendering_mode(:invisible)
+      #   canvas.text("#{canvas.text_rendering_mode.name}", at: [10, 50])
+      #   canvas.stroke_color("red").line_width(20).line(30, 20, 30, 80).stroke
       INVISIBLE = NamedValue.new(:invisible, 3)
 
-      # Fill text and add to path for clipping
+      # Fill text and add to path for clipping.
+      #
+      # Specify as 4 or :fill_clip.
+      #
+      #   #>pdf-small-hide
+      #   canvas.font("Helvetica", size: 13)
+      #   canvas.stroke_color("green").line_width(0.5)
+      #   canvas.text_rendering_mode(:fill_clip)
+      #   canvas.text("#{canvas.text_rendering_mode.name}", at: [10, 50])
+      #   canvas.stroke_color("red").line_width(20).line(30, 20, 30, 80).stroke
       FILL_CLIP = NamedValue.new(:fill_clip, 4)
 
-      # Stroke text and add to path for clipping
+      # Stroke text and add to path for clipping.
+      #
+      # Specify as 5 or :stroke_clip.
+      #
+      #   #>pdf-small-hide
+      #   canvas.font("Helvetica", size: 13)
+      #   canvas.stroke_color("green").line_width(0.5)
+      #   canvas.text_rendering_mode(:stroke_clip)
+      #   canvas.text("#{canvas.text_rendering_mode.name}", at: [10, 50])
+      #   canvas.stroke_color("red").line_width(20).line(30, 20, 30, 80).stroke
       STROKE_CLIP = NamedValue.new(:stroke_clip, 5)
 
-      # Fill, then stroke text and add to path for clipping
+      # Fill, then stroke text and add to path for clipping.
+      #
+      # Specify as 6 or :fill_stroke_clip.
+      #
+      #   #>pdf-small-hide
+      #   canvas.font("Helvetica", size: 13)
+      #   canvas.stroke_color("green").line_width(0.5)
+      #   canvas.text_rendering_mode(:fill_stroke_clip)
+      #   canvas.text("#{canvas.text_rendering_mode.name}", at: [10, 50])
+      #   canvas.stroke_color("red").line_width(20).line(30, 20, 30, 80).stroke
       FILL_STROKE_CLIP = NamedValue.new(:fill_stroke_clip, 6)
 
-      # Add text to path for clipping
+      # Add text to path for clipping.
+      #
+      # Specify as 7 or :clip.
+      #
+      #   #>pdf-small-hide
+      #   canvas.font("Helvetica", size: 13)
+      #   canvas.stroke_color("green").line_width(0.5)
+      #   canvas.text_rendering_mode(:clip)
+      #   canvas.text("#{canvas.text_rendering_mode.name}", at: [10, 50])
+      #   canvas.stroke_color("red").line_width(20).line(30, 20, 30, 80).stroke
       CLIP = NamedValue.new(:clip, 7)
 
     end
@@ -389,7 +512,7 @@ module HexaPDF
       attr_accessor :leading
 
       # The font for the text.
-      attr_accessor :font
+      attr_reader :font
 
       # The font size.
       attr_reader :font_size
@@ -415,23 +538,25 @@ module HexaPDF
 
       # The scaled character spacing used in glyph displacement calculations.
       #
-      # This returns the value T_c multiplied by #scaled_horizontal_scaling.
+      # This returns the character spacing multiplied by #scaled_horizontal_scaling.
       #
       # See PDF1.7 s9.4.4
       attr_reader :scaled_character_spacing
 
       # The scaled word spacing used in glyph displacement calculations.
       #
-      # This returns the value T_w  multiplied by #scaled_horizontal_scaling.
+      # This returns the word spacing  multiplied by #scaled_horizontal_scaling.
       #
       # See PDF1.7 s9.4.4
       attr_reader :scaled_word_spacing
 
       # The scaled font size used in glyph displacement calculations.
       #
-      # This returns the value T_fs / 1000 multiplied by #scaled_horizontal_scaling.
+      # This returns the font size multiplied by the scaling factor from glyph space to text space
+      # (0.001 for all fonts except Type3 fonts or the scaling specified in /FontMatrix for Type3
+      # fonts) and multiplied by #scaled_horizontal_scaling.
       #
-      # See PDF1.7 s9.4.4
+      # See PDF1.7 s9.4.4, HexaPDF::Type::FontType3
       attr_reader :scaled_font_size
 
       # The scaled horizontal scaling used in glyph displacement calculations.
@@ -542,6 +667,15 @@ module HexaPDF
         self.fill_color = color_space.default_color
       end
 
+      ##
+      # :attr_writer: font
+      #
+      # Sets the font and updates the glyph space to text space scaling.
+      def font=(font)
+        @font = font
+        update_scaled_font_size
+      end
+
       ##
       # :attr_writer: character_spacing
       #
@@ -566,7 +700,7 @@ module HexaPDF
       # Sets the font size and updates the scaled font size.
       def font_size=(size)
         @font_size = size
-        @scaled_font_size = size / 1000.0 * @scaled_horizontal_scaling
+        update_scaled_font_size
       end
 
       ##
@@ -579,7 +713,15 @@ module HexaPDF
         @scaled_horizontal_scaling = scaling / 100.0
         @scaled_character_spacing = @character_spacing * @scaled_horizontal_scaling
         @scaled_word_spacing = @word_spacing * @scaled_horizontal_scaling
-        @scaled_font_size = @font_size / 1000.0 * @scaled_horizontal_scaling
+        update_scaled_font_size
+      end
+
+      private
+
+      # Updates the cached value for the scaled font size.
+      def update_scaled_font_size
+        @scaled_font_size = @font_size * (@font&.glyph_scaling_factor || 0.001) *
+          @scaled_horizontal_scaling
       end
 
     end

data/lib/hexapdf/dictionary_fields.rb

@@ -293,7 +293,7 @@ module HexaPDF
       end
 
       # :nodoc:
-      DATE_RE = /\AD:(\d{4})(\d\d)?(\d\d)?(\d\d)?(\d\d)?(\d\d)?([Z+-])?(?:(\d\d)(?:'|'(\d\d)'?|\z)?)?\z/n
+      DATE_RE = /\AD:(\d{4})(\d\d)?(\d\d)?(\d\d)?(\d\d)?(\d\d)?([Z+-])?(?:(\d\d)(?:'|'([0-5]\d)'?|\z)?)?\z/n
 
       # Checks if the given object is a string and converts into a Time object if possible.
       # Otherwise returns +nil+.

data/lib/hexapdf/encryption/standard_security_handler.rb

@@ -328,8 +328,7 @@ module HexaPDF
           raise(HexaPDF::UnsupportedEncryptionError,
                 "Invalid /R value for standard security handler")
         elsif dict[:R] <= 4 && !document.trailer[:ID].kind_of?(PDFArray)
-          raise(HexaPDF::EncryptionError,
-                "Document ID for needed for decryption")
+          document.trailer[:ID] = ['', '']
         end
         @trailer_id_hash = trailer_id_hash
 

data/lib/hexapdf/layout/style.rb

@@ -1069,7 +1069,8 @@ module HexaPDF
 
       # The font size scaled appropriately.
       def scaled_font_size
-        @scaled_font_size ||= calculated_font_size / 1000.0 * scaled_horizontal_scaling
+        @scaled_font_size ||= calculated_font_size * font.pdf_object.glyph_scaling_factor *
+          scaled_horizontal_scaling
       end
 
       # The character spacing scaled appropriately.

data/lib/hexapdf/parser.rb

@@ -62,9 +62,15 @@ module HexaPDF
       @object_stream_data = {}
       @reconstructed_revision = nil
       @in_reconstruct_revision = false
+      @contains_xref_streams = false
       retrieve_pdf_header_offset_and_version
     end
 
+    # Returns +true+ if the PDF file contains cross-reference streams.
+    def contains_xref_streams?
+      @contains_xref_streams
+    end
+
     # Loads the indirect (potentially compressed) object specified by the given cross-reference
     # entry.
     #
@@ -230,6 +236,7 @@ module HexaPDF
           maybe_raise("Cross-reference stream doesn't contain entry for itself", pos: pos)
           xref_section.add_in_use_entry(obj.oid, obj.gen, pos)
         end
+        @contains_xref_streams = true
       end
       xref_section.delete(0)
       [xref_section, trailer]
@@ -335,7 +342,8 @@ module HexaPDF
       step_size = 1024
       pos = @io.pos
       eof_not_found = pos == 0
-      startxref_missing = false
+      startxref_missing = startxref_mangled = false
+      startxref_offset = nil
 
       while pos != 0
         @io.pos = [pos - step_size, 0].max
@@ -343,27 +351,31 @@ module HexaPDF
         lines = @io.read(step_size + 40).split(/[\r\n]+/)
 
         eof_index = lines.rindex {|l| l.strip == '%%EOF' }
-        unless eof_index
+        if !eof_index
           eof_not_found = true
-          next
-        end
-        unless eof_index >= 2 && lines[eof_index - 2].strip == "startxref"
+        elsif lines[eof_index - 1].strip =~ /\Astartxref\s(\d+)\z/
+          startxref_offset = $1.to_i
+          startxref_mangled = true
+          break # we found it even if it the syntax is not entirely correct
+        elsif eof_index < 2 || lines[eof_index - 2].strip != "startxref"
           startxref_missing = true
-          next
+        else
+          startxref_offset = lines[eof_index - 1].to_i
+          break # we found it
         end
-
-        break # we found the startxref offset
       end
 
       if eof_not_found
         maybe_raise("PDF file trailer with end-of-file marker not found", pos: pos,
                     force: !eof_index)
+      elsif startxref_mangled
+        maybe_raise("PDF file trailer keyword startxref on same line as value", pos: pos)
       elsif startxref_missing
         maybe_raise("PDF file trailer is missing startxref keyword", pos: pos,
                     force: eof_index < 2 || lines[eof_index - 2].strip != "startxref")
       end
 
-      @startxref_offset = lines[eof_index - 1].to_i
+      @startxref_offset = startxref_offset
     end
 
     # Returns the reconstructed revision.

data/lib/hexapdf/task/optimize.rb

@@ -72,8 +72,19 @@ module HexaPDF
       #   Compresses the content streams of all pages if set to +true+. Note that this can take a
       #   *very* long time because each content stream has to be unfiltered, parsed, serialized
       #   and then filtered again.
+      #
+      # prune_page_resources::
+      #   Removes all unused XObjects from the resources dictionaries of all pages. It is
+      #   recommended to also set the +compact+ argument because otherwise the unused XObjects won't
+      #   be deleted from the document.
+      #
+      #   This is sometimes necessary after importing pages from other PDF files that use a single
+      #   resources dictionary for all pages.
       def self.call(doc, compact: false, object_streams: :preserve, xref_streams: :preserve,
-                    compress_pages: false)
+                    compress_pages: false, prune_page_resources: false)
+        used_refs = compress_pages(doc) if compress_pages
+        prune_page_resources(doc, used_refs) if prune_page_resources
+
         if compact
           compact(doc, object_streams, xref_streams)
         elsif object_streams != :preserve
@@ -83,8 +94,6 @@ module HexaPDF
         else
           doc.each(only_current: false, &method(:delete_fields_with_defaults))
         end
-
-        compress_pages(doc) if compress_pages
       end
 
       # Compacts the document by merging all revisions into one, deleting null and unused entries
@@ -214,12 +223,41 @@ module HexaPDF
 
       # Compresses the contents of all pages by parsing and then serializing again. The HexaPDF
       # serializer is already optimized for small output size so nothing else needs to be done.
+      #
+      # Returns a hash of the form key=>true where the keys are the used XObjects (for use with
+      # #prune_page_resources).
       def self.compress_pages(doc)
+        used_refs = {}
         doc.pages.each do |page|
           processor = SerializationProcessor.new
           HexaPDF::Content::Parser.parse(page.contents, processor)
           page.contents = processor.result
           page[:Contents].set_filter(:FlateDecode)
+          xobjects = page.resources[:XObject]
+          processor.used_references.each {|ref| used_refs[xobjects[ref]] = true }
+        end
+        used_refs
+      end
+
+      # Deletes all XObject entries from the resources dictionaries of all pages whose names do not
+      # match the keys in +used_refs+.
+      def self.prune_page_resources(doc, used_refs)
+        unless used_refs
+          used_refs = {}
+          doc.pages.each do |page|
+            xobjects = page.resources[:XObject]
+            HexaPDF::Content::Parser.parse(page.contents) do |op, operands|
+              used_refs[xobjects[operands[0]]] = true if op == :Do
+            end
+          end
+        end
+
+        doc.pages.each do |page|
+          xobjects = page.resources[:XObject]
+          xobjects.each do |key, obj|
+            next if used_refs[obj]
+            xobjects.delete(key)
+          end
         end
       end
 
@@ -228,14 +266,19 @@ module HexaPDF
 
         attr_reader :result #:nodoc:
 
+        # Contains all found references
+        attr_reader :used_references
+
         def initialize #:nodoc:
           @result = ''.b
           @serializer = HexaPDF::Serializer.new
+          @used_references = []
         end
 
         def process(op, operands) #:nodoc:
           @result << HexaPDF::Content::Operator::DEFAULT_OPERATORS[op].
             serialize(@serializer, *operands)
+          @used_references << operands[0] if op == :Do
         end
 
       end

data/lib/hexapdf/type/font.rb

@@ -98,6 +98,11 @@ module HexaPDF
         embedded?
       end
 
+      # Returns the glyph scaling factor for transforming from glyph space to text space.
+      def glyph_scaling_factor
+        0.001
+      end
+
       private
 
       # Parses and caches the ToUnicode CMap.

data/lib/hexapdf/type/font_type3.rb

@@ -41,6 +41,10 @@ module HexaPDF
 
     # Represents a Type 3 font.
     #
+    # Note: We assume the /FontMatrix is only used for scaling, i.e. of the form [x 0 0 +/-x 0 0].
+    # If it is of a different form, things won't work correctly. This will be handled once such a
+    # case is found.
+    #
     # See: PDF1.7 s9.6.5
     class FontType3 < FontSimple
 
@@ -51,6 +55,22 @@ module HexaPDF
       define_field :CharProcs,  type: Dictionary, required: true
       define_field :Resources,  type: Dictionary, version: '1.2'
 
+      # Returns the bounding box of the font.
+      def bounding_box
+        matrix = self[:FontMatrix]
+        bbox = self[:FontBBox].value
+        if matrix[3] < 0    # Some writers invert the y-axis
+          bbox = bbox.dup
+          bbox[1], bbox[3] = -bbox[3], -bbox[1]
+        end
+        bbox
+      end
+
+      # Returns the glyph scaling factor for transforming from glyph space to text space.
+      def glyph_scaling_factor
+        self[:FontMatrix][0]
+      end
+
       private
 
       def perform_validation

data/lib/hexapdf/version.rb

@@ -37,6 +37,6 @@
 module HexaPDF
 
   # The version of HexaPDF.
-  VERSION = '0.18.0'
+  VERSION = '0.19.3'
 
 end

data/lib/hexapdf/writer.rb

@@ -66,6 +66,8 @@ module HexaPDF
       @serializer = Serializer.new
       @serializer.encrypter = @document.encrypted? ? @document.security_handler : nil
       @rev_size = 0
+
+      @use_xref_streams = false
     end
 
     # Writes the document to the IO object.
@@ -87,6 +89,7 @@ module HexaPDF
       IO.copy_stream(@document.revisions.parser.io, @io)
 
       @rev_size = @document.revisions.current.next_free_oid
+      @use_xref_streams = @document.revisions.parser.contains_xref_streams?
 
       revision = Revision.new(@document.revisions.current.trailer)
       @document.revisions.each do |rev|
@@ -170,10 +173,13 @@ module HexaPDF
         end
       end
 
-      if !object_streams.empty? && xref_stream.nil?
-        raise HexaPDF::Error, "Cannot use object streams when there is no xref stream"
+      if (!object_streams.empty? || @use_xref_streams) && xref_stream.nil?
+        xref_stream = @document.wrap({Type: :XRef}, oid: rev.next_free_oid)
+        rev.add(xref_stream)
       end
 
+      @use_xref_streams = true if xref_stream
+
       [xref_stream, object_streams]
     end
 

data/test/hexapdf/content/test_graphics_state.rb

@@ -2,6 +2,7 @@
 
 require 'test_helper'
 require 'hexapdf/content/graphics_state'
+require 'ostruct'
 
 # Dummy class used as wrapper so that constant lookup works correctly
 class GraphicsStateWrapper < Minitest::Spec
@@ -146,6 +147,13 @@ class GraphicsStateWrapper < Minitest::Spec
     it "fails when restoring the graphics state if the stack is empty" do
       assert_raises(HexaPDF::Error) { @gs.restore }
     end
-  end
 
+    it "uses the correct glyph to text space scaling" do
+      font = OpenStruct.new
+      font.glyph_scaling_factor = 0.002
+      @gs.font = font
+      @gs.font_size = 10
+      assert_equal(0.02, @gs.scaled_font_size)
+    end
+  end
 end

data/test/hexapdf/content/test_operator.rb

@@ -4,6 +4,7 @@ require 'test_helper'
 require 'hexapdf/content/operator'
 require 'hexapdf/content/processor'
 require 'hexapdf/serializer'
+require 'ostruct'
 
 describe HexaPDF::Content::Operator::BaseOperator do
   before do
@@ -190,9 +191,11 @@ end
 
 describe_operator :SetGraphicsStateParameters, :gs do
   it "applies parameters from an ExtGState dictionary" do
+    font = OpenStruct.new
+    font.glyph_scaling_factor = 0.01
     @processor.resources[:ExtGState] = {Name: {LW: 10, LC: 2, LJ: 2, ML: 2, D: [[3, 5], 2],
                                                RI: 2, SA: true, BM: :Multiply, CA: 0.5, ca: 0.5,
-                                               AIS: true, TK: false, Font: [:Test, 10]}}
+                                               AIS: true, TK: false, Font: [font, 10]}}
     @processor.resources.define_singleton_method(:document) do
       Object.new.tap {|obj| obj.define_singleton_method(:deref) {|o| o } }
     end
@@ -210,7 +213,7 @@ describe_operator :SetGraphicsStateParameters, :gs do
     assert_equal(0.5, gs.stroke_alpha)
     assert_equal(0.5, gs.fill_alpha)
     assert(gs.alpha_source)
-    assert_equal(:Test, gs.font)
+    assert_equal(font, gs.font)
     assert_equal(10, gs.font_size)
     refute(gs.text_knockout)
   end
@@ -448,7 +451,9 @@ describe_operator :SetFontAndSize, :Tf do
       self[:Font] && self[:Font][name]
     end
 
-    @processor.resources[:Font] = {F1: :test}
+    font = OpenStruct.new
+    font.glyph_scaling_factor = 0.01
+    @processor.resources[:Font] = {F1: font}
     invoke(:F1, 10)
     assert_equal(@processor.resources.font(:F1), @processor.graphics_state.font)
     assert_equal(10, @processor.graphics_state.font_size)

data/test/hexapdf/encryption/test_standard_security_handler.rb

@@ -229,19 +229,21 @@ describe HexaPDF::Encryption::StandardSecurityHandler do
       assert_match(/Invalid \/R/i, exp.message)
     end
 
-    it "fails if the ID in the document's trailer is missing although it is needed" do
+    it "fails if the supplied password is invalid" do
       exp = assert_raises(HexaPDF::EncryptionError) do
-        @handler.set_up_decryption({Filter: :Standard, V: 2, R: 2})
+        @handler.set_up_decryption({Filter: :Standard, V: 2, R: 6, U: 'a' * 48, O: 'a' * 48,
+                                    UE: 'a' * 32, OE: 'a' * 32})
       end
-      assert_match(/Document ID/i, exp.message)
+      assert_match(/Invalid password/i, exp.message)
     end
 
-    it "fails if the supplied password is invalid" do
+    it "assigns empty strings to the trailer's ID field if it is missing" do
+      refute(@document.trailer.key?(:ID))
       exp = assert_raises(HexaPDF::EncryptionError) do
-        @handler.set_up_decryption({Filter: :Standard, V: 2, R: 6, U: 'a' * 48, O: 'a' * 48,
-                                    UE: 'a' * 32, OE: 'a' * 32})
+        @handler.set_up_decryption({Filter: :Standard, V: 1, R: 2, U: 'a' * 48, O: 'a' * 48, P: 15})
       end
       assert_match(/Invalid password/i, exp.message)
+      assert_equal(['', ''], @document.trailer[:ID].value)
     end
 
     describe "/Perms field checking" do

data/test/hexapdf/layout/test_style.rb

@@ -597,6 +597,11 @@ end
 describe HexaPDF::Layout::Style do
   before do
     @style = HexaPDF::Layout::Style.new
+    @style.font = Object.new.tap do |obj|
+      obj.define_singleton_method(:pdf_object) do
+        Object.new.tap {|pdf| pdf.define_singleton_method(:glyph_scaling_factor) { 0.001 } }
+      end
+    end
   end
 
   it "can assign values on initialization" do
@@ -644,6 +649,7 @@ describe HexaPDF::Layout::Style do
   end
 
   it "has several simple and dynamically generated properties with default values" do
+    @style = HexaPDF::Layout::Style.new
     assert_raises(HexaPDF::Error) { @style.font }
     assert_equal(10, @style.font_size)
     assert_equal(0, @style.character_spacing)
@@ -725,6 +731,11 @@ describe HexaPDF::Layout::Style do
       font = Object.new
       font.define_singleton_method(:scaling_factor) { 1 }
       font.define_singleton_method(:wrapped_font) { wrapped_font }
+      font.define_singleton_method(:pdf_object) do
+        obj = Object.new
+        obj.define_singleton_method(:glyph_scaling_factor) { 0.001 }
+        obj
+      end
       @style.font = font
     end
 

data/test/hexapdf/task/test_optimize.rb

@@ -159,4 +159,30 @@ describe HexaPDF::Task::Optimize do
       assert_equal("10 10 m\nq\nQ\nBI\n/Name 5 ID\ndataEI\n", page.contents)
     end
   end
+
+  describe "prune_page_resources" do
+    it "removes all unused XObject references" do
+      [false, true].each do |compress_pages|
+        page1 = @doc.pages.add
+        page1.resources[:XObject] = {}
+        page1.resources[:XObject][:test] = @doc.add({})
+        page1.resources[:XObject][:used_on_page2] = @doc.add({})
+        page1.resources[:XObject][:unused] = @doc.add({})
+        page1.contents = "/test Do"
+        page2 = @doc.pages.add
+        page2.resources[:XObject] = {}
+        page2.resources[:XObject][:used_on2] = page1.resources[:XObject][:used_on_page2]
+        page2.resources[:XObject][:also_unused] = page1.resources[:XObject][:unused]
+        page2.contents = "/used_on2 Do"
+
+        @doc.task(:optimize, prune_page_resources: true, compress_pages: compress_pages)
+
+        assert(page1.resources[:XObject].key?(:test))
+        assert(page1.resources[:XObject].key?(:used_on_page2))
+        refute(page1.resources[:XObject].key?(:unused))
+        assert(page2.resources[:XObject].key?(:used_on2))
+        refute(page2.resources[:XObject].key?(:also_unused))
+      end
+    end
+  end
 end

data/test/hexapdf/test_dictionary_fields.rb

@@ -173,6 +173,7 @@ describe HexaPDF::DictionaryFields do
 
     it "allows conversion to a Time object from a binary string" do
       refute(@field.convert('test'.b, self))
+      refute(@field.convert('D:01211016165909+00\'64'.b, self))
 
       [
         ["D:1998", [1998, 01, 01, 00, 00, 00, "-00:00"]],

data/test/hexapdf/test_parser.rb

@@ -338,6 +338,11 @@ describe HexaPDF::Parser do
       assert_equal(5, @parser.startxref_offset)
     end
 
+    it "handles the case of startxref and its value being on the same line" do
+      create_parser("startxref 5\n%%EOF")
+      assert_equal(5, @parser.startxref_offset)
+    end
+
     it "fails even in big files when nothing is found" do
       create_parser("\nhallo" * 5000)
       exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.startxref_offset }
@@ -366,6 +371,13 @@ describe HexaPDF::Parser do
       exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.startxref_offset }
       assert_match(/end-of-file marker not found/, exp.message)
     end
+
+    it "fails on strict parsing if the startxref is on the same line as its value" do
+      @document.config['parser.on_correctable_error'] = proc { true }
+      create_parser("startxref 5\n%%EOF")
+      exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.startxref_offset }
+      assert_match(/startxref on same line/, exp.message)
+    end
   end
 
   describe "file_header_version" do
@@ -531,12 +543,14 @@ describe HexaPDF::Parser do
       xref_section, trailer = @parser.load_revision(@parser.startxref_offset)
       assert_equal({Test: 'now'}, trailer)
       assert(xref_section[1].in_use?)
+      refute(@parser.contains_xref_streams?)
     end
 
     it "works for a cross-reference stream" do
       xref_section, trailer = @parser.load_revision(212)
       assert_equal({Size: 2}, trailer)
       assert(xref_section[1].in_use?)
+      assert(@parser.contains_xref_streams?)
     end
 
     it "fails if another object is found instead of a cross-reference stream" do

data/test/hexapdf/test_writer.rb

@@ -40,7 +40,7 @@ describe HexaPDF::Writer do
       219
       %%EOF
       3 0 obj
-      <</Producer(HexaPDF version 0.18.0)>>
+      <</Producer(HexaPDF version 0.19.3)>>
       endobj
       xref
       3 1
@@ -72,7 +72,7 @@ describe HexaPDF::Writer do
       141
       %%EOF
       6 0 obj
-      <</Producer(HexaPDF version 0.18.0)>>
+      <</Producer(HexaPDF version 0.19.3)>>
       endobj
       2 0 obj
       <</Length 10>>stream
@@ -103,21 +103,50 @@ describe HexaPDF::Writer do
     assert_document_conversion(@compressed_input_io)
   end
 
-  it "writes a document in incremental mode" do
-    doc = HexaPDF::Document.new(io: @std_input_io)
-    doc.pages.add
-    output_io = StringIO.new
-    HexaPDF::Writer.write(doc, output_io, incremental: true)
-    assert_equal(output_io.string[0, @std_input_io.string.length], @std_input_io.string)
-    doc = HexaPDF::Document.new(io: output_io)
-    assert_equal(4, doc.revisions.size)
-    assert_equal(2, doc.revisions.current.each.to_a.size)
+  describe "write_incremental" do
+    it "writes a document in incremental mode" do
+      doc = HexaPDF::Document.new(io: @std_input_io)
+      doc.pages.add
+      output_io = StringIO.new
+      HexaPDF::Writer.write(doc, output_io, incremental: true)
+      assert_equal(output_io.string[0, @std_input_io.string.length], @std_input_io.string)
+      doc = HexaPDF::Document.new(io: output_io)
+      assert_equal(4, doc.revisions.size)
+      assert_equal(2, doc.revisions.current.each.to_a.size)
+    end
+
+    it "uses an xref stream if the document already contains at least one" do
+      doc = HexaPDF::Document.new(io: @compressed_input_io)
+      doc.pages.add
+      output_io = StringIO.new
+      HexaPDF::Writer.write(doc, output_io, incremental: true)
+      refute_match(/^trailer/, output_io.string)
+    end
   end
 
-  it "raises an error if no xref stream is in a revision but object streams are" do
+  it "creates an xref stream if no xref stream is in a revision but object streams are" do
     document = HexaPDF::Document.new
     document.add({Type: :ObjStm})
-    assert_raises(HexaPDF::Error) { HexaPDF::Writer.new(document, StringIO.new).write }
+    HexaPDF::Writer.new(document, StringIO.new).write
+    assert(:XRef, document.object(2).type)
+  end
+
+  it "creates an xref stream if a previous revision had one" do
+    document = HexaPDF::Document.new
+    document.pages.add
+    document.revisions.add
+    document.pages.add
+    document.add({Type: :ObjStm})
+    document.revisions.add
+    document.pages.add
+    io = StringIO.new
+    HexaPDF::Writer.new(document, io).write
+
+    document = HexaPDF::Document.new(io: io)
+    assert_equal(3, document.revisions.count)
+    assert(document.revisions[0].none? {|obj| obj.type == :XRef })
+    assert(document.revisions[1].one? {|obj| obj.type == :XRef })
+    assert(document.revisions[2].one? {|obj| obj.type == :XRef })
   end
 
   it "raises an error if the class is misused and an xref section contains invalid entries" do

data/test/hexapdf/type/test_font.rb

@@ -64,4 +64,8 @@ describe HexaPDF::Type::Font do
       assert_equal(5, @font.font_file)
     end
   end
+
+  it "returns the glyph scaling factor" do
+    assert_equal(0.001, @font.glyph_scaling_factor)
+  end
 end

data/test/hexapdf/type/test_font_type3.rb

@@ -9,10 +9,25 @@ describe HexaPDF::Type::FontType3 do
     @doc = HexaPDF::Document.new
     @font = @doc.add({Type: :Font, Subtype: :Type3, Encoding: :WinAnsiEncoding,
                       FirstChar: 32, LastChar: 34, Widths: [600, 0, 700],
-                      FontBBox: [0, 0, 100, 100], FontMatrix: [1, 0, 0, 1, 0, 0],
+                      FontBBox: [0, 100, 100, 0], FontMatrix: [0.002, 0, 0, 0.002, 0, 0],
                       CharProcs: {}})
   end
 
+  describe "bounding_box" do
+    it "returns the font's bounding box" do
+      assert_equal([0, 0, 100, 100], @font.bounding_box)
+    end
+
+    it "inverts the y-values if necessary based on /FontMatrix" do
+      @font[:FontMatrix][3] *= -1
+      assert_equal([0, -100, 100, 0], @font.bounding_box)
+    end
+  end
+
+  it "returns the glyph scaling factor" do
+    assert_equal(0.002, @font.glyph_scaling_factor)
+  end
+
   describe "validation" do
     it "works for valid objects" do
       assert(@font.validate)

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: hexapdf
 version: !ruby/object:Gem::Version
-  version: 0.18.0
+  version: 0.19.3
 platform: ruby
 authors:
 - Thomas Leitner
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2021-11-04 00:00:00.000000000 Z
+date: 2021-12-14 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: cmdparse