pdfrb 0.6.0 → 0.7.10

data/lib/pdfrb/source/recovery.rb

@@ -6,9 +6,23 @@ module Pdfrb
     # patterns and synthesise an XrefSection from the offsets found.
     # Triggered when the xref table is missing or points to a wrong
     # offset.
+    #
+    # Extended recovery capabilities:
+    #   * Cross-reference table reconstruction (the original mode).
+    #   * Trailer recovery: find /Root and /Encrypt by scanning for
+    #     the keys directly when the trailer is corrupt.
+    #   * Hybrid xref detection: read both xref table and xref stream
+    #     when both are present.
+    #   * Object-stream reconstruction: rebuild an XrefSection from a
+    #     /Type /ObjStm stream when its header is intact but the
+    #     xref stream is missing.
     module Recovery
       OBJ_PATTERN = /(\d+)\s+(\d+)\s+obj\b/.freeze
-      private_constant :OBJ_PATTERN
+      TRAILER_ROOT_PATTERN = %r{/Root\s+(\d+)\s+(\d+)\s+R}.freeze
+      TRAILER_ENCRYPT_PATTERN = %r{/Encrypt\s+(\d+)\s+(\d+)\s+R}.freeze
+      TRAILER_INFO_PATTERN = %r{/Info\s+(\d+)\s+(\d+)\s+R}.freeze
+      private_constant :OBJ_PATTERN, :TRAILER_ROOT_PATTERN,
+                       :TRAILER_ENCRYPT_PATTERN, :TRAILER_INFO_PATTERN
 
       module_function
 
@@ -30,6 +44,56 @@ module Pdfrb
         end
         section
       end
+
+      # Scan the file for /Root, /Info, /Encrypt references when the
+      # trailer dict is unreadable. Returns a Hash with :Root, :Info,
+      # :Encrypt keys as References (or nil per key).
+      def recover_trailer_references(io)
+        io.seek(0, IO::SEEK_SET)
+        data = io.read.to_s
+        data.force_encoding(Encoding::BINARY)
+
+        root_match = data.match(TRAILER_ROOT_PATTERN)
+        info_match = data.match(TRAILER_INFO_PATTERN)
+        encrypt_match = data.match(TRAILER_ENCRYPT_PATTERN)
+
+        {
+          Root: root_match ? Pdfrb::Model::Reference.new(root_match[1].to_i, root_match[2].to_i) : nil,
+          Info: info_match ? Pdfrb::Model::Reference.new(info_match[1].to_i, info_match[2].to_i) : nil,
+          Encrypt: encrypt_match ? Pdfrb::Model::Reference.new(encrypt_match[1].to_i, encrypt_match[2].to_i) : nil,
+        }
+      end
+
+      # Detect hybrid xref: PDF 1.7+ allows both a classical xref
+      # table AND an xref stream in the same revision, referenced
+      # from the table via /XRefStm. Returns true if the file has
+      # both. Pure-Ruby: scans for `startxref` followed by both
+      # `xref` and a Type=XRef stream object.
+      def hybrid_xref?(io)
+        io.seek(0, IO::SEEK_SET)
+        data = io.read.to_s
+        data.force_encoding(Encoding::BINARY)
+
+        has_table = data.include?("\nxref\n")
+        has_stream = data.match?(/\/Type\s*\/XRef\b/)
+        has_table && has_stream
+      end
+
+      # Reconstruct an XrefSection from a /Type /ObjStm stream by
+      # reading its /N + /First + decompressed body. Used when the
+      # outer xref stream is corrupt but a known ObjStm is intact.
+      def rebuild_from_object_stream(objstm, document)
+        return nil unless objstm
+
+        pairs = ObjectStreamReader.read(objstm, document)
+        section = XrefSection.new
+        pairs.each_key do |oid|
+          # Mark as compressed in objstm 0 (placeholder; caller
+          # fills in the real objstm_oid).
+          section.add_compressed(oid, 0, 0, objstm.oid)
+        end
+        section
+      end
     end
   end
 end

data/lib/pdfrb/source/tokenizer.rb

@@ -52,6 +52,27 @@ module Pdfrb
         self
       end
 
+      # Whether the underlying IO is at end of stream.
+      def eof?
+        @io.eof?
+      end
+
+      # Read one byte from the underlying IO. Returns nil at EOF.
+      def read_byte
+        b = @io.getbyte
+        @pos += 1 if b
+        b
+      end
+
+      # Skip whitespace bytes (NUL, HT, LF, FF, CR, SP).
+      def skip_whitespace
+        while (b = peek_byte)
+          break unless WHITESPACE_BYTES.include?(b)
+
+          advance_byte
+        end
+      end
+
       private
 
       def fill_lookahead(n)

data/lib/pdfrb/source.rb

@@ -20,5 +20,6 @@ module Pdfrb
     autoload :ObjectStreamReader, "pdfrb/source/object_stream_reader"
     autoload :Recovery, "pdfrb/source/recovery"
     autoload :LinearizationDetection, "pdfrb/source/linearization_detection"
+    autoload :LinearizationReader, "pdfrb/source/linearization_reader"
   end
 end

data/lib/pdfrb/task/regenerate_appearances.rb

@@ -0,0 +1,98 @@
+# frozen_string_literal: true
+
+module Pdfrb
+  module Task
+    # Walks every widget annotation in the document and regenerates
+    # the /AP /N appearance stream via Appearance::Generator. Use
+    # after mutating field values directly (e.g. via Form#set_value)
+    # or to repair documents with /NeedAppearances set so viewers
+    # don't have to synthesise appearances on the fly.
+    #
+    # Per ISO 32000-2 §12.5.5, an annotation's /AP /N is the normal
+    # appearance; PDF/A requires it for every widget. This task
+    # also clears /NeedAppearances on /AcroForm since by definition
+    # all fields now have a computed appearance.
+    module RegenerateAppearances
+      module_function
+
+      # @param document [Pdfrb::Document]
+      # @param only: [Array<Symbol>, nil] limit to the given FT values
+      #   (e.g. [:Tx, :Btn]). Default: all field types.
+      # @return [Integer] count of appearances regenerated.
+      def call(document, only: nil)
+        generator = Pdfrb::Appearance::Generator.new(document)
+        widgets = each_widget(document).to_a
+        count = 0
+        widgets.each do |widget|
+          next unless matching?(widget, only)
+
+          regenerated = regenerate(generator, widget)
+          count += 1 if regenerated
+        end
+        clear_need_appearances(document)
+        count
+      end
+
+      def each_widget(document)
+        return enum_for(:each_widget, document) unless block_given?
+
+        document.each_indirect_object do |obj|
+          next unless obj.value.is_a?(::Hash)
+
+          subtype = obj.value[:Subtype]
+          yield obj if subtype == :Widget
+        end
+      end
+
+      def matching?(widget, only)
+        return true if only.nil?
+
+        ft = widget.value[:FT]
+        ft && only.include?(ft.to_sym)
+      end
+
+      def regenerate(generator, widget)
+        ft = widget.value[:FT]&.to_sym
+        value = widget.value[:V]
+        case ft
+        when :Tx
+          generator.text_field(widget, value: value.to_s)
+          true
+        when :Btn
+          if value.is_a?(::Hash)
+            # Multi-state radio; regenerate each state's appearance.
+            value.each_key do |state|
+              generator.checkbox(widget, checked: state != :Off)
+            end
+          else
+            generator.checkbox(widget, checked: [:Yes, true].include?(value))
+          end
+          true
+        when :Ch
+          generator.combo(widget, value: value.to_s)
+          true
+        else
+          false
+        end
+      rescue StandardError
+        # Appearance generation is best-effort; if a single field
+        # raises, skip it rather than aborting the whole sweep.
+        false
+      end
+
+      def clear_need_appearances(document)
+        catalog = document.catalog
+        return unless catalog
+
+        acroform = catalog.value[:AcroForm]
+        return unless acroform
+
+        if acroform.is_a?(Pdfrb::Model::Cos::Dictionary)
+          acroform.value.delete(:NeedAppearances)
+        elsif acroform.is_a?(::Hash)
+          acroform.delete(:NeedAppearances)
+        end
+      end
+    end
+  end
+end

data/lib/pdfrb/task/thumbnail.rb

@@ -0,0 +1,133 @@
+# frozen_string_literal: true
+
+module Pdfrb
+  module Task
+    # Generates a low-resolution thumbnail for each page (or a single
+    # page) and attaches it via the page-level /Thumb key. PDF
+    # viewers use /Thumb to render page previews quickly without
+    # parsing the full content stream.
+    #
+    # The thumbnail is a Form XObject rendered to a small image
+    # XObject. Pure-Ruby implementation: renders only the page's text
+    # (no image / form XObject content) at the target resolution
+    # using the standard 14 AFM metrics when available, otherwise a
+    # placeholder rectangle.
+    module Thumbnail
+      DEFAULT_DPI = 72
+      DEFAULT_MAX_WIDTH = 200
+
+      module_function
+
+      # Generate thumbnails for every page in +document+. Returns the
+      # count of thumbnails added. Idempotent: existing /Thumb
+      # entries are left alone unless +force:+ is true.
+      def call(document, dpi: DEFAULT_DPI, max_width: DEFAULT_MAX_WIDTH, force: false)
+        count = 0
+        document.pages.each do |page|
+          next if page.value[:Thumb] && !force
+
+          thumb = build_thumbnail(document, page, dpi: dpi, max_width: max_width)
+          next unless thumb
+
+          page.value[:Thumb] = Pdfrb::Model::Reference.new(thumb.oid, thumb.gen)
+          count += 1
+        end
+        count
+      end
+
+      # Build a single thumbnail image XObject for +page+ at the
+      # given DPI, capped at +max_width+ pixels. Returns the image
+      # XObject or nil if the page has no media box.
+      def build_thumbnail(document, page, dpi: DEFAULT_DPI, max_width: DEFAULT_MAX_WIDTH)
+        media_box = page.value[:MediaBox]
+        return nil unless media_box.is_a?(::Array) && media_box.length == 4
+
+        width, height = page_pixel_dimensions(media_box, dpi, max_width)
+        return nil unless width.positive? && height.positive?
+
+        # Single-component grayscale image — 1 byte per pixel.
+        # White background, black "text" rectangle placeholder. Real
+        # text rendering would require running the content processor
+        # against a rasteriser; that's a much larger feature.
+        pixel_data = render_placeholder(width, height, page)
+        compressed = require_zlib.deflate(pixel_data)
+
+        image = document.add(
+          {
+            Type: :XObject, Subtype: :Image,
+            Width: width, Height: height,
+            ColorSpace: :DeviceGray,
+            BitsPerComponent: 8,
+            Filter: :FlateDecode,
+            Length: compressed.bytesize
+          },
+          type: Pdfrb::Model::Cos::Stream
+        )
+        image.stream = compressed
+        image
+      end
+
+      # Compute thumbnail pixel dimensions, capped at +max_width+.
+      def page_pixel_dimensions(media_box, dpi, max_width)
+        _x0, _y0, x1, y1 = media_box
+        pt_w = x1.to_f
+        pt_h = y1.to_f
+        return [0, 0] if pt_w.zero? || pt_h.zero?
+
+        scale = (dpi / 72.0) * (max_width.to_f / pt_w)
+        scale = 1.0 if scale > 1.0
+        [(pt_w * scale).round, (pt_h * scale).round]
+      end
+
+      # Render a 1-bpp-style grayscale placeholder: white background,
+      # a black border, and a few black "text line" rectangles
+      # representing where text content might appear. Pure visual
+      # cue, not a real render.
+      def render_placeholder(width, height, _page)
+        bytes = Array.new(width * height, 255) # white
+
+        draw_rect(bytes, width, height, 0, 0, width, height, 0) # border (black)
+        # Determine text area: roughly 80% of page width, centred
+        margin = (width * 0.1).round
+        text_top = (height * 0.85).round
+        line_height = (height * 0.02).round
+        line_height = 1 if line_height.zero?
+        line_count = [(height * 0.6 / line_height).floor, 3].max
+        line_count.times do |i|
+          y = text_top - (i * line_height * 2)
+          draw_horizontal_line(bytes, width, height, margin, y,
+                               width - margin, 0)
+        end
+        bytes.pack("C*").b
+      end
+
+      def draw_rect(bytes, width, height, x0, y0, x1, y1, value)
+        x0.upto([x1 - 1, width - 1].min) do |x|
+          set_pixel(bytes, width, height, x, y0, value)
+          set_pixel(bytes, width, height, x, y1 - 1, value)
+        end
+        y0.upto([y1 - 1, height - 1].min) do |y|
+          set_pixel(bytes, width, height, x0, y, value)
+          set_pixel(bytes, width, height, x1 - 1, y, value)
+        end
+      end
+
+      def draw_horizontal_line(bytes, width, height, x0, y, x1, value)
+        x0.upto([x1 - 1, width - 1].min) do |x|
+          set_pixel(bytes, width, height, x, y, value)
+        end
+      end
+
+      def set_pixel(bytes, width, height, x, y, value)
+        return if x.negative? || y.negative? || x >= width || y >= height
+
+        bytes[(y * width) + x] = value
+      end
+
+      def require_zlib
+        require "zlib"
+        ::Zlib
+      end
+    end
+  end
+end

data/lib/pdfrb/task.rb

@@ -9,5 +9,7 @@ module Pdfrb
     autoload :GenerateCorpus, "pdfrb/task/generate_corpus"
     autoload :Benchmark, "pdfrb/task/benchmark"
     autoload :MemoryProfile, "pdfrb/task/memory_profile"
+    autoload :RegenerateAppearances, "pdfrb/task/regenerate_appearances"
+    autoload :Thumbnail, "pdfrb/task/thumbnail"
   end
 end

data/lib/pdfrb/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Pdfrb
-  VERSION = "0.6.0"
+  VERSION = "0.7.10"
 end

metadata

@@ -1,13 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: pdfrb
 version: !ruby/object:Gem::Version
-  version: 0.6.0
+  version: 0.7.10
 platform: ruby
 authors:
 - Ribose Inc.
+autorequire:
 bindir: exe
 cert_chain: []
-date: 1980-01-02 00:00:00.000000000 Z
+date: 2026-08-11 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: thor
@@ -698,6 +699,7 @@ files:
 - data/pdfrb/arlington/latest/_UniversalArray.tsv
 - data/pdfrb/arlington/latest/_UniversalDictionary.tsv
 - data/pdfrb/glyphlist.txt
+- data/pdfrb/layout/hyphenation_en.txt
 - data/pdfrb/zapfdingbats.txt
 - exe/pdfrb
 - lib/pdfrb.rb
@@ -745,18 +747,28 @@ files:
 - lib/pdfrb/cli/watermark.rb
 - lib/pdfrb/color.rb
 - lib/pdfrb/color/color_space.rb
+- lib/pdfrb/color/default_profile.rb
 - lib/pdfrb/color/icc_profile.rb
+- lib/pdfrb/color/icc_validator.rb
 - lib/pdfrb/compare.rb
 - lib/pdfrb/compare/comparator.rb
 - lib/pdfrb/compare/report.rb
 - lib/pdfrb/composer.rb
 - lib/pdfrb/configuration.rb
 - lib/pdfrb/conformance.rb
+- lib/pdfrb/conformance/ltv.rb
+- lib/pdfrb/conformance/pades.rb
+- lib/pdfrb/conformance/pdf_2_af.rb
 - lib/pdfrb/conformance/pdf_a.rb
+- lib/pdfrb/conformance/pdf_a4_deep.rb
 - lib/pdfrb/conformance/pdf_ua.rb
+- lib/pdfrb/conformance/pdf_ua2_deep.rb
+- lib/pdfrb/conformance/pdf_ua_tagging_deep.rb
+- lib/pdfrb/conformance/pdf_vt.rb
 - lib/pdfrb/conformance/pdf_x.rb
 - lib/pdfrb/conformance/rule.rb
 - lib/pdfrb/conformance/structure_elements.rb
+- lib/pdfrb/conformance/tagged_pdf.rb
 - lib/pdfrb/conformance/verapdf_bridge.rb
 - lib/pdfrb/content.rb
 - lib/pdfrb/content/canvas.rb
@@ -774,6 +786,7 @@ files:
 - lib/pdfrb/content/operators/color.rb
 - lib/pdfrb/content/operators/general.rb
 - lib/pdfrb/content/operators/graphics_state_params.rb
+- lib/pdfrb/content/operators/inline_image.rb
 - lib/pdfrb/content/operators/marked_content.rb
 - lib/pdfrb/content/operators/painting.rb
 - lib/pdfrb/content/operators/path.rb
@@ -794,6 +807,7 @@ files:
 - lib/pdfrb/digital_signature/cms_handler.rb
 - lib/pdfrb/digital_signature/handler.rb
 - lib/pdfrb/digital_signature/signing.rb
+- lib/pdfrb/digital_signature/timestamp_client.rb
 - lib/pdfrb/digital_signature/timestamp_handler.rb
 - lib/pdfrb/digital_signature/verification.rb
 - lib/pdfrb/digital_signature/verification_result.rb
@@ -824,9 +838,11 @@ files:
 - lib/pdfrb/encryption/aes.rb
 - lib/pdfrb/encryption/identity.rb
 - lib/pdfrb/encryption/password_verification.rb
+- lib/pdfrb/encryption/public_key_security_handler.rb
 - lib/pdfrb/encryption/rc4.rb
 - lib/pdfrb/encryption/security_handler.rb
 - lib/pdfrb/encryption/standard_security_handler.rb
+- lib/pdfrb/encryption/v5_writer.rb
 - lib/pdfrb/error.rb
 - lib/pdfrb/filter.rb
 - lib/pdfrb/filter/ascii_85_decode.rb
@@ -879,15 +895,18 @@ files:
 - lib/pdfrb/font_loader/standard14.rb
 - lib/pdfrb/font_loader/true_type.rb
 - lib/pdfrb/font_loader/type1.rb
+- lib/pdfrb/font_loader/type3.rb
 - lib/pdfrb/font_loader/variant_from_name.rb
 - lib/pdfrb/font_resolver.rb
 - lib/pdfrb/image.rb
 - lib/pdfrb/image/audit.rb
 - lib/pdfrb/image/downsampler.rb
 - lib/pdfrb/image_loader.rb
+- lib/pdfrb/image_loader/gif.rb
 - lib/pdfrb/image_loader/jpeg.rb
 - lib/pdfrb/image_loader/pdf.rb
 - lib/pdfrb/image_loader/png.rb
+- lib/pdfrb/image_loader/tiff.rb
 - lib/pdfrb/importer.rb
 - lib/pdfrb/layout.rb
 - lib/pdfrb/layout/bidi.rb
@@ -895,18 +914,25 @@ files:
 - lib/pdfrb/layout/box_fitter.rb
 - lib/pdfrb/layout/column_box.rb
 - lib/pdfrb/layout/container_box.rb
+- lib/pdfrb/layout/font_fallback.rb
 - lib/pdfrb/layout/frame.rb
+- lib/pdfrb/layout/hyphenation.rb
 - lib/pdfrb/layout/image_box.rb
 - lib/pdfrb/layout/inline_box.rb
+- lib/pdfrb/layout/justification_kashidas.rb
 - lib/pdfrb/layout/line.rb
 - lib/pdfrb/layout/list_box.rb
+- lib/pdfrb/layout/multi_cell_text_layout.rb
+- lib/pdfrb/layout/multi_page_table_box.rb
 - lib/pdfrb/layout/numeric_refinements.rb
 - lib/pdfrb/layout/page_style.rb
+- lib/pdfrb/layout/polygon_frame.rb
 - lib/pdfrb/layout/style.rb
 - lib/pdfrb/layout/table_box.rb
 - lib/pdfrb/layout/text_box.rb
 - lib/pdfrb/layout/text_fragment.rb
 - lib/pdfrb/layout/text_layouter.rb
+- lib/pdfrb/layout/text_shaper.rb
 - lib/pdfrb/linearization.rb
 - lib/pdfrb/linearization/hint_stream.rb
 - lib/pdfrb/linearization/writer.rb
@@ -1057,6 +1083,7 @@ files:
 - lib/pdfrb/model/type/output_intent.rb
 - lib/pdfrb/model/type/page.rb
 - lib/pdfrb/model/type/page_label.rb
+- lib/pdfrb/model/type/page_piece_info.rb
 - lib/pdfrb/model/type/page_tree_node.rb
 - lib/pdfrb/model/type/page_tree_node_root.rb
 - lib/pdfrb/model/type/pattern.rb
@@ -1142,6 +1169,7 @@ files:
 - lib/pdfrb/source/header_reader.rb
 - lib/pdfrb/source/helpers.rb
 - lib/pdfrb/source/linearization_detection.rb
+- lib/pdfrb/source/linearization_reader.rb
 - lib/pdfrb/source/object_reader.rb
 - lib/pdfrb/source/object_stream_reader.rb
 - lib/pdfrb/source/parser.rb
@@ -1160,6 +1188,8 @@ files:
 - lib/pdfrb/task/memory_profile.rb
 - lib/pdfrb/task/merge.rb
 - lib/pdfrb/task/optimize.rb
+- lib/pdfrb/task/regenerate_appearances.rb
+- lib/pdfrb/task/thumbnail.rb
 - lib/pdfrb/test_utils.rb
 - lib/pdfrb/validator.rb
 - lib/pdfrb/version.rb
@@ -1181,6 +1211,7 @@ metadata:
   changelog_uri: https://github.com/claricle/pdfrb/blob/main/CHANGELOG.md
   bug_tracker_uri: https://github.com/claricle/pdfrb/issues
   rubygems_mfa_required: 'true'
+post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -1195,7 +1226,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 4.0.16
+rubygems_version: 3.5.22
+signing_key:
 specification_version: 4
 summary: Pure-Ruby PDF parser, Arlington-model-driven domain model, and serializer
 test_files: []