PDFium VCL Component Features

Core Components

• TPdf - Main PDF document handling component for opening, navigating, extracting text, and editing PDF documents
• TPdfView - Visual component for displaying and interacting with PDF pages
• TPdfReadingSession - Read-aloud orchestration component for host TTS engines and TPdfView word tracking
• TPdfThumbnailView - Scrollable thumbnail side panel for page navigation
• Native LCL platforms - Linux Qt6 and macOS Cocoa run native PDFium with correct LP64 bindings, platform font/file services and actual bitmap channel layouts; see native installation for qualified Free Pascal/LCL configurations and capability limits

Document Operations

• Document Loading - Load PDF files from disk, memory streams, complete byte arrays, zero-based Index/Count byte-array windows, or on-demand TStream adapters via LoadCustomDocument. Compatible, strict, and recovery policies add bounded classic, hybrid, and xref-stream diagnostics without replacing PDFium's semantic parser
• Verified Repaired Output - SaveAsRepaired creates a full single-revision copy and independently strict-loads its xref and page structure before returning bytes or atomically replacing a file
• Typed Object Graph - ExportObjectGraph emits bounded typed JSON for direct and object-stream values, while SaveAsObjectGraphPatched applies validated top-level dictionary changes as a staged incremental revision and strict-loads the result before delivery
• Password Protection - Open password-protected PDF documents and report effective UserPermissions for the authenticated session
• Document Creation - Build new PDF documents from scratch with CreateDocument, AddPage, AddText, AddImage, AddPicture, AddJpegImage, and CreatePath
• Document Saving - Save modified documents with various options (incremental, remove security) and explicit PDF target versions from PDF 1.0 through PDF 1.7
• PDF/A-1b Output - SaveAsPdfA / SaveAsPdfAToStream post-process the base save with an incremental update that injects an XMP metadata stream, an sRGB ICC OutputIntent, and an updated document catalog so the result is PDF/A-1b conformant. Custom ICC profiles are supported through TPdfASaveOptions.IccProfileData
• PDF/A Validation - ValidatePdfA and the read-only PdfAConformance property scan an open document for XMP pdfaid markers, OutputIntent with embedded ICC, encryption presence, and report the conformance level (pac1a, pac1b, pac2b, pac3b, pacNone, or pacUnknown) plus a list of TPdfAValidationIssue entries describing each problem found
• PDF Version Compliance Validator - When SaveAs targets PDF 1.0 through PDF 1.7 the writer rejects any dictionary entry, filter, annotation subtype, action type, font, colour space, or encryption parameter introduced after the chosen target version (e.g. JBIG2Decode, MarkInfo, OutputIntents, AES / AES-256, Polygon / Caret / Watermark / Redact annotations, JavaScript / Trans / SetState actions)
• Stream Compression - Automatic FlateDecode compression ensures all PDF streams are optimally compressed when saving; set Compressed := False to opt out for debug pipelines
• Scalable Batch Merge - MergeFiles imports file-backed sources with per-source page ranges and passwords, keeps at most one source document open, and exposes progress, cancellation, stop or skip handling, rollback, form, outline, resource, and viewer-preference policies
• Page Subset Import - ImportPagesByIndex copies an explicit zero-based index array from a source document, with InsertAt = 0 to insert before page 1 or InsertAt = PageCount to append. Pass an empty index array to import every source page
• N-up Composition - ImportNPagesToOne returns a brand-new TPdf whose pages are NumX×NumY composites of the source document's pages, sized in PDF user units (1 unit = 1/72 in)
• Page Reordering - MovePages reorders pages in place using a zero-based index array and a destination index for the first moved page
• Page Plan Validation - ValidatePagePlan previews delete, move, insert, split, and merge mappings, inventories navigation, label, form, structure-tree, imported-feature, and signature effects, and rejects plans that would leave dangling targets or modify signed content without explicit authorisation
• Two-pass Layout Generation - GenerateLayout measures and paginates wrapped text, rules, images, page XObjects, headers, and footers before atomically emitting native page objects, while stable {page} and {pages} placeholders preserve the measured page count
• Disk-backed Paged Layout - GeneratePagedLayout writes commands to caller-selected seekable storage, restores each page from compact non-contiguous segments, bounds retained page arrays with an LRU cache, and appends PDFium pages sequentially
• Cross-page Table Layout - TPdfLayoutTable resolves fixed, fractional, and content-sized columns once, repeats contiguous header rows, splits eligible body rows, keeps row-spanning groups intact, and emits captions, cell borders, links, and one continuous Table / TR / TH / TD semantic hierarchy across pages
• Cached Layout Nodes - Layout plans are cached by complete constraint, content, font, style, source, and measurer identities, while TPdfLayoutCachedNode builds repeated templates lazily and emits each placement from one reusable native Form XObject
• Automatic Tagged PDF Layout - GenerateTaggedLayout emits native MCIDs, artifacts, a Document structure tree, ParentTree and IDTree mappings, headings, lists, figures with alternate text, table headers, and Link annotations with OBJR ownership from the same measured layout plan; item constraints keep blocks and headings together, enforce widow / orphan minima, and force physical page breaks
• Marked-content Editing - PageObjectMarks, AddPageObjectMark, SetPageObjectMarkParam, RemovePageObjectMarkParam, and RemovePageObjectMark inspect and mutate BDC property lists through stable native marks
• Bidi and Complex-script Text - AddShapedText converts directional runs, contextual forms, ligatures, advances, and offsets into positioned CID Type 2 glyphs, while generated ToUnicode mappings and ActualText preserve logical Unicode semantics
• CID Type 2 Font Authoring - AddCidType2Text embeds caller-supplied TrueType data with explicit CID-to-GID and ToUnicode mappings, safe optional subsetting, controlled full-font fallback, and extraction integrity checks
• Unified Cache Byte Budget - TPdf.CacheBudgetBytes coordinates parsed page scenes, thumbnails, structure data, text snapshots, and page-object snapshots through one measurable global LRU, while TPdfTileCache and the system-font cache can join a caller-selected governor
• Host System-font Provider - ConfigureSystemFontProvider routes native missing-font requests through an application provider with byte validation, bounded LRU caching, optional shared byte governance, native fallback, table extraction, and lifecycle statistics
• Form XObject Page Reuse - CreateXObjectFromPage manufactures a reusable TPdfXObject handle from any page of another document; InsertFormObjectFromXObject inserts ONE copy of the wrapped page into the current page. Ideal for watermarks, template overlays, and "stamp one page onto many" workflows. The destination page object can be positioned, scaled, and rotated through FPDFPageObj_SetMatrix and the TPdfMatrix helpers
• Page Management - Add, delete, and manipulate individual pages; AddPage(0, ...) inserts the very first page and activates it immediately
• Page Boxes - Read and write MediaBox, CropBox, BleedBox, TrimBox, and ArtBox on the current page
• Catalog Language - Read and write the document language stored in the catalog /Lang entry
• Trailer File IDs - Read permanent and changing trailer /ID entries as binary bytes
• Document JavaScript Actions - Enumerate document-level JavaScript name-tree action names and scripts
• Tagged PDF Structure Tree - Inspect logical structure elements, accessibility text, language, marked-content IDs, and hierarchy for tagged PDF pages
• Viewer Preferences - Read PrintCopies, PrintPageRanges, PrintScaling, and PrintPaperHandling from the document's viewer-preference dictionary; saved PDFs default to a Fit Height first-page open action

Text Operations

• Text Extraction - Extract text from PDF pages with character-level precision via Text, TextInRectangle, indexed Character, CharacterRectangle, and CharacterOrigin properties
• Structured Text Model - GetStructuredText builds a bounded page, block, line, and styled-span hierarchy from native character geometry, font metadata, text-object identity, Tagged PDF semantics, and physical-layout fallback ordering
• Table Detection and Extraction - ExtractTables and ExtractDocumentTables combine native word boxes with transformed path rulings to detect bordered or whitespace-aligned tables, preserve merged spans, score confidence, mark page continuations, and emit CSV or JSON
• Semantic Reflow and Responsive HTML - BuildReflowDocument combines Tagged PDF semantics, physical-layout inference, and table extraction into a bounded reusable node tree whose safe HTML output preserves headings, paragraphs, lists, figures, page sections, and table spans
• Searchable OCR Layer - ApplyOcrSearchLayer passes bounded top-down BGRA page renders to an application provider, maps confidence quads through native page geometry, and adds invisible searchable text without re-encoding scan images
• Sentence-Level Reading Units - TPdf.ReadingUnits, TPdf.DocumentReadingUnits, TPdfView.ReadingUnits and BuildReadingUnits expose sentence-sized TPdfReadingUnit records for synchronised TTS listen-reading
• Visual Line Selection - TPdfView.SelectLineAt selects the visual text line containing a one-based page and zero-based character index through the same line-boundary model used by triple-click selection and Shift+Home or Shift+End
• Word-Level Reading Highlighting - TPdf.PageWordBoxes, TPdfView.PageWordBoxes, BuildWordBoxes, WordBoxIndexAtCharIndex, TPdfView.SelectWordBox, TPdfView.SelectWordSpan, TPdfView.TrackReadingWordAt, TPdfView.SetReadingWord, TPdfView.ClearReadingWord, ReadingWordColor, ReadingWordFollow Maps a page character offset to the containing TPdfWordBox, highlights it, and returns the word index; Highlights one TPdfWordBox as the current spoken word on TPdfView; Clears the current word-level reading cursor overlay on TPdfView; Controls the overlay colour used for the current word-level reading cursor; Controls whether TPdfView scrolls the current spoken word into view while tracking reading
• Read-Aloud Session Orchestration - TPdfReadingSession, TPdfSpokenText, SpokenTextAppend, BuildSpokenTextFromWords, SpokenToPageChar, and EstimateSpokenWordIndex coordinate host TTS engines, spoken chunks, word-position callbacks, estimated pacing, pause/resume handling, and TPdfView word-cursor updates
• Text Search - Find text with case-sensitive, whole word, and consecutive search options using FindFirst, FindNext, and FindPrevious, or stream forward and backward matches across the complete document with TPdf.CreateDocumentSearch, page-space result rectangles, cooperative cancellation, and constant page-handle memory
• Search Highlight Overlay - TPdfView.HighlightSearchText replaces the current page's match overlay, while AddSearchTextHighlight appends unique ranges for additional terms so several queries can remain visible together. ClearHighlight removes the complete set; navigating to another page clears it automatically because character ranges are page-specific. HighlightColor defaults to clYellow and is published so applications can theme the overlay
• Character Properties - Access font size, weight, angle, and positioning for individual characters with the CharacterAngle, CharacterFontSize, CharacterFontWeight, CharacterGenerated, CharacterIsHyphen, and CharacterMapError indexed properties
• European Character Support - AddText automatically loads and caches the matching system TrueType font before creating a text object, so accented Czech, French, German, Spanish, Portuguese, Polish, and Turkish characters that PDFium's built-in standard fonts cannot represent now render correctly without any code change
• Text Addition - Add new text to PDF pages with custom fonts, colours, transparency, and rotation
• Text Object Positioning - Set per-character positions for existing text objects through SetTextPositions and TPdfSingleArray
• Font Information - Retrieve font properties including family name, embedding status, italic angle, ascent / descent, and raw font data on a per-page-object basis

Annotation Support

• Annotation Types - Support for text, link, highlight, underline, strikeout, and many other annotation types
• Annotation Creation - Create new annotations programmatically
• Annotation Editing - Modify existing annotation properties including colour, transparency, and content
• Link Annotations - Handle internal and external links with action support

Form Support

• Voice Form-Filling Accessors - FocusedFormFieldValue, FocusedFormFieldName and FocusedFormFieldDetails on TPdf and TPdfView expose the current logical form field value, name and TPdfFormFieldInfo for keyboard or voice-driven form filling
• Form Types - Support for AcroForm and XFA forms
• Form Fill - Interactive form filling with JavaScript support
• V8 DLL Selection - Set EnableV8Engine before loading to use pdfium.v8.dll from the matching DLLs\Win32 or DLLs\Win64 folder and query V8/XFA helper availability via V8FeaturesAvailable, XfaBStrHelpersAvailable, and GetRecommendedV8Flags
• Field Access - Read and write form field values by index or position
• Field Information - Access field type, flags, name, and other properties
• Field Flags - Update AcroForm field flags by index or position through SetFormFieldFlags / SetFormFieldFlagsAt, including read-only, required, password, and multi-select flags
• Form Flattening - GenerateFormAppearances regenerates the /AP appearance stream for every form widget on every page, and FlattenPage / FlattenAllPages bake annotations and form fields into static page content (non-editable)
• Form Widget Text Selection, Clipboard & Undo / Redo - On the focused AcroForm widget the view exposes GetSelectedFormText, CopySelectedFormTextToClipboard, CutSelectedFormTextToClipboard, PasteFormTextFromClipboard, SelectAllFormText, and FormUndo / FormRedo / FormCanUndo / FormCanRedo. The editor handles Ctrl+A, Ctrl+C, Ctrl+Insert, Ctrl+X, Shift+Delete, Ctrl+V, and Shift+Insert; cut and paste reject read-only fields or sessions without form-fill or modify permission
• Document Edit Transactions and Undo / Redo - BeginEditTransaction, CommitEditTransaction, and RollbackEditTransaction group arbitrary page, object, annotation, attachment, and metadata edits, while UndoEdit, RedoEdit, and portable edit journals use bounded SHA-256-verified bidirectional deltas
• Crash Recovery and Commit-Driven Autosave - checksummed recovery packages bind source and recovered fingerprints to a bounded edit journal, while atomic autosave refreshes an explicit sidecar after commit, undo, or redo without rewriting the source PDF

Image Operations

• Image Extraction - Extract bitmaps from PDF pages with the indexed Bitmap property; iterate BitmapCount to save every image to disk
• JPEG Insertion - AddJpegImage takes a TStream of raw JPEG bytes and embeds the image as a /DCTDecode XObject without re-encoding
• Picture Support - AddPicture accepts a TPicture object with automatic format conversion through the registered VCL/LCL graphic classes
• Direct Image Insertion - AddImage(FileName) recognises any format registered with the VCL or LCL graphics units (BMP, PNG via PngImage, JPG, ...) and shares the AddPicture pipeline. AddImage(TBitmap) skips the TPicture intermediary, which simplifies callers that already hold a TBitmap from rendering or batch generation work
• Image ICC Profile - Image page objects expose their embedded ICC profile data when present (binding for FPDFImageObj_GetIccProfileDataDecoded)
• Thumbnail Generation - Generate current-page thumbnail bitmaps via the Thumbnail property or render any one-based page through TPdf.RenderPageThumbnail without changing PageNumber, with a revision-aware LRU bounded by entry and pixel-byte ceilings
• Parsed Page Scene Cache - TPdf.RenderPageByIndex, RenderPageThumbnail, and ExtractThumbnail reuse bounded LRU-cached PDFium page scenes while page edits, page-tree changes, form state, and document teardown invalidate stale entries

Helper Units

• Async Infrastructure - FPdfAsync combines structured cooperative cancellation with removable callbacks and deadline sources, configurable TPdfFuture<T> result dispatch, and a reusable TPdfAsyncExecutor with fixed workers, four FIFO priority lanes, bounded-queue backpressure, per-task handles, timing diagnostics, aggregate statistics, and graceful shutdown
• Stream Adapter - TPdfStreamAdapter bridges a Pascal TStream to PDFium's FPDF_FILEACCESS callbacks so LoadCustomDocument reads blocks on demand instead of buffering the entire PDF in memory
• Progressive Download Sessions - TPdfProgressiveDocument keeps PDFium's native availability provider alive, feeds document, form, and page hints into TPdfRangeRequestScheduler, coalesces missing ranges, preempts background transfers for page-critical bytes, bounds request sizes and retries, cancels redundant transports, and opens a linearised page before the complete PDF arrives
• Bounded Stream Inspection - TPdfStreamInspector inventories physical stream objects, exposes safe raw windows and ordered general-filter decoding, reports image and attachment native-access paths, and enforces source, output, dictionary, count, and expansion-ratio budgets
• Content Stream Normalisation - TPdf.SaveAsNormalizedContent tokenises logical page content, removes comments, validates state and container balance, preserves unsupported operators and inline-image data, and coalesces safe /Contents arrays
• Fast Web View Output - TPdf.SaveAsLinearized writes first-page objects ahead of later pages, emits two cross-reference sections, and builds page-offset plus shared-object hint tables so an HTTP Range client can open a requested page without downloading the complete file
• Object Stream Output - TPdf.SaveAsObjectStream packs eligible values into bounded PDF 1.5+ ObjStm groups, keeps opening-critical objects direct, and writes adaptive type-2 xref entries for smaller files without sacrificing random access
• Dirty-object Incremental Save - TPdf.BeginIncrementalSave and CommitIncrementalSave compare stable semantic snapshots, preserve every previous encrypted or unencrypted byte, inherit R5, R6, or R7 security parameters, append only changed objects, strictly reopen each result, and use a durable side file plus atomic replacement for file commits
• Optional Content Control - TPdf.InspectOptionalContent inventories layers, configurations, usage metadata, radio groups, order, locks, OCMD policies, and visibility expressions, while SaveAsOptionalContentConfigured writes deterministic layer states that PDFium applies through its native rendering context
• Bounded Working Storage - TPdfRandomAccessStore keeps small data in memory, spills once to a private delete-on-close temporary file at a caller-selected ceiling, and supplies positioned reads without changing the shared cursor; TPdf.SaveMemoryLimit applies the same policy to core save post-processing, while default open-action and page-label catalog updates use fixed-size windows and a sparse lazy object index
• Native Version Query - TPdf.PdfVersion reuses FPDF_GetFileVersion parser state for header and catalog semantics, with one eight-byte header guard to reject malformed catalog downgrades
• Matrix Geometry Toolkit - FPdfMatrix wraps PDFium's FS_MATRIX with post- and pre-composition, pivot-aware transforms, stretch / contain / cover rectangle fitting, scale-aware invertibility diagnostics, decomposition, non-mutating inverse helpers, vector and quadrilateral mapping, one-allocation point or rectangle batches, in-place variants, and one-pass transformed bounds
• Cancellable Progressive Rendering - TPdf.RenderPageProgressive drives PDFium's FPDF_RenderPageBitmap_Start + RenderPage_Continue loop and polls an IPdfCancellationToken between iterations via the IFSDK_PAUSE callback so long high-DPI renders can abort mid-page instead of blocking until completion. Returns a TPdfProgressiveStatus of prsDone, prsCancelled, or prsFailed
• Zero-Copy Render Path - RenderPage / RenderTile / RenderPageProgressive write PDFium output directly into the destination TBitmap's DIB buffer instead of staging through an internal buffer, saving one width × height × 4 bytes memory copy per render - most noticeable at high DPI (thumbnail panels, print-quality export, batch export). A per-scanline fallback path is preserved when the destination DIB cannot be wrapped safely
• Low-Vision Bitmap Transforms - InvertPdfBitmap, GrayscalePdfBitmap and DuotonePdfBitmap transform rendered 32-bit TBitmap pixels in place for inverted, greyscale, or two-colour low-vision views
• Live Low-Vision Viewing - TPdfView.Options supports reInvert and reNightDuotone, with NightDuotoneInk and NightDuotonePaper controlling the live night palette before overlays are painted

Drawing Operations

• Path Drawing - Create paths with lines, Bezier curves, and closed shapes
• Object Removal - Remove page objects and regenerate page content from TPdf or TPdfView
• Fill and Stroke - Configure fill mode, colours, transparency, and stroke properties
• Blend Modes - Support for various PDF blend modes
• Line Styles - Configure line cap, line join, and stroke width

Digital Signatures

• Signature Detection - Detect and enumerate digital signatures in PDF documents
• Signature Information - Access signature contents, byte range, sub-filter, reason, and time
• MDP Permissions - Query document modification detection and prevention permissions
• Post-signature revision analysis - AnalyzeSignatureRevisions classifies original-byte incremental changes against DocMDP and FieldMDP, identifies validation-material updates, and reports shadow definitions or unresolved compressed objects without resaving the document
• PDF 2.0 elliptic-curve profiles - PAdES validation identifies NIST, Brainpool, Ed25519, and Ed448 signatures, checks curve-to-digest congruence and AlgorithmIdentifier parameters, and reports unsupported native cryptography independently from invalid combinations

Attachment Support

• Attachment Access - Read, create, modify, and delete embedded file attachments
• Bounded Batch Extraction - ReadAttachmentMetadata inventories attachments once and ExtractAttachmentsToStreams writes one decoded payload at a time under per-file, aggregate, chunk, and cancellation controls
• Attachment Names - Access attachment filenames
• Attachment Types - Read embedded file MIME subtype values when present

Navigation

• Bookmarks - Navigate document bookmarks with hierarchical support via the Bookmarks array, BookmarkChildren, HasBookmarkChildren, and BookmarkFrom traversal helpers
• Named Destinations - Access named destinations by index or name through Destination, DestinationByName, and DestinationCount
• Page Labels - Retrieve page labels for each page via PageLabel; useful when documents use Roman numerals for front matter
• Web Links - Detect and navigate web links on pages with WebLink, WebLinkCount, and TPdfView.WebLinkAtPos
• Link Annotations - Access link annotations with action and destination information through LinkAnnotation, LinkAnnotationCount, and TPdfView.LinkAnnotationAtPos
• Safe Link Click Handling - TPdfView raises OnAnnotationLinkClick / OnWebLinkClick events before automatic goto / URI handling, with launch actions disabled by default. Link click detection runs only when the gesture is a click rather than a drag, using a fresh MouseDown anchor so panning gestures still flow through the existing scroll path unchanged

Rendering

• High-Quality Rendering - Render pages to device context, bitmap, or tile through RenderPage / RenderTile, with zero-copy direct writes into the destination DIB buffer
• Cancellable Progressive Rendering - RenderPageProgressive drives PDFium's progressive loop and polls an IPdfCancellationToken between iterations, returning prsDone, prsCancelled, or prsFailed. The bitmap holds the partial work done so far when cancelled
• Parallel Page Rendering - RenderPagesParallel captures the current document once, gives each worker an independently loaded PDFium module and document, preserves request order, and bounds per-page plus aggregate retained output memory
• Secure Content Redaction - SaveAsRedacted applies /Redact rectangles to an isolated snapshot, conservatively removes intersecting top-level page objects and overlapping annotations, paints opaque replacements, removes embedded attachments, and writes a sanitised single-revision result without changing the loaded document
• Render Options - LCD text optimisation (reLcd), annotations (reAnnotations), greyscale, and printing mode flags
• Skia Hooks - Low-level FPDF_RenderPageSkia and FPDF_FFLDrawSkia bindings are available when callers provide an external SkCanvas (depends on bundled DLL exports)
• Print Mode Selection - Set EMF, text-only, PostScript, pass-through, image-mask, and Type 42 PDFium print modes via SetPdfPrintMode with TPdfPrintMode
• Async Overlay - TPdfView can display an opt-in loading overlay while application-owned workers render with separate TPdf instances
• Page Rotation - Support for 0, 90, 180, and 270 degree rotation (ro0, ro90, ro180, ro270)
• Transparency Detection - Transparent property reports whether any object on the current page has transparency, including blend modes and partial alpha

Viewing (TPdfView)

• Display Modes - dmSingleContinuous, dmTwoPageContinuous, and dmTwoPageContinuousWithCover (the cover page stands alone, then continues with two-page spreads). The unit-level ComputeSpreadPages helper exposes the spread-page maths without instantiating a viewer
• Fit Mode - FitMode chooses how the page sits in the viewport: pfmNone (manual zoom), pfmActualSize (100%), pfmFitPage (whole page fits in the viewport), pfmFitWidth (page width matches viewport width), and pfmFitHeight (page height matches viewport height). The active fit mode is re-applied automatically on Resize and when the current page changes; setting Zoom directly cancels the active fit mode (FitMode reverts to pfmNone), matching the standard PDF viewer convention
• Zoom Control - Variable zoom with PageZoom, PageWidthZoom, PageHeightZoom, and ActualSizeZoom per-page helpers
• Page Colour - PageColor sets the PDF page background independently of the surrounding control's Color. It defaults to clWhite, keeping pages paper-white whatever theme the host application uses. A dark-mode viewer can keep its scroll area dark whilst displaying the PDF page in white, or another colour chosen by the application
• Interactive Text Selection - End-user drag-to-select text on PDF pages, with hit-test auto-switching between selection and pan: clicking on a character starts a selection drag, clicking on whitespace pans. Double-click selects the word, triple-click selects the line, Ctrl+A selects the page (both pages in spread mode), Ctrl+C copies as CF_UNICODETEXT, Shift+arrows extend selection, hover over text switches the cursor to crIBeam, and dragging into a viewer edge auto-scrolls the page. New published SelectionColor and OnSelectionChange; new public HasSelection, SelectedText, SelectionCharCount, ClearSelection, SelectAll, CopySelectionToClipboard. Spread display modes select across both visible pages with CR/LF joining. A default right-click Copy popup is auto-attached when the host has not assigned its own PopupMenu. Gated by the v1.17.0 AllowUserTextSelection flag for kiosk hosts
• Coordinate Mapping - Convert between viewer device coordinates and PDF page coordinates with DeviceToPage / PageToDevice, including overloads that take a TRect and return a TPdfRectangle
• Text Selection and Hit Testing - Resolve characters under viewer points with CharacterIndexAtPos and paint text selection highlights via PaintSelection
• Mouse & Keyboard - Full input device support with mouse drag scrolling, PgUp / PgDn / Ctrl+Home / Ctrl+End shortcuts, and wheel-up / wheel-down page navigation. Ctrl+wheel is reserved for application zoom and skipped by the navigation handler
• User Input Lock-Down - AllowUserPageChange controls keyboard page shortcuts, ChangePageOnMouseScrolling controls wheel page jumps, and AllowUserTextSelection controls mouse / keyboard text-selection input. All three default to True; flip them False to lock down kiosks, preview panes, or read-only embeddings without subclassing
• Async Render Overlay - AsyncRendering, LoadingText, and LoadingColor draw a small "Loading..." indicator over the rendered page so applications using their own worker for long renders share a consistent visual affordance. Defaults preserve existing behaviour. TPdf now serialises render and page-unload calls per instance; applications still own worker dispatch and must avoid mutating a document while a render is in flight
• Safe Link Click Handling - LinkOptions is a published set that controls automatic behaviour for the four PDF link action types: loAutoGoto (page jump), loAutoOpenURI (ShellExecute), loAutoLaunch (ShellExecute), and loAutoEmbeddedGoto (reserved). Defaults enable loAutoGoto and loAutoOpenURI only; loAutoLaunch and loAutoEmbeddedGoto stay off so a stray click cannot run an arbitrary program. OnAnnotationLinkClick and OnWebLinkClick fire before any automatic action and let the application mark the event Handled or redirect it
• Form Interaction - Interactive form editing within the viewer, plus the v1.23.0 selection / Undo / Redo helpers
• Visual Styling - PageShadowSize, PageShadowColor, and PageBorderColor paint an optional drop shadow under the page rectangle and a one-pixel border around it; SmoothScroll now animates line/page scrollbar commands in timer-driven segments while thumb tracking remains immediate
• Thumbnail Navigation - TPdfThumbnailView (unit FPdfThumbnail) renders one page thumbnail per row in a scrollable side panel with hover and selection highlighting plus click-to-go navigation. Designed to sit next to a TPdfView. Properties: Pdf, CurrentPageIndex, ThumbnailWidth, ThumbnailHeight, TextColor, HoverColor, SelectionColor, BorderColor; event: OnPageClick

Compression

• Automatic Compression - When the Compressed property is True (default), all PDF streams are ensured to be FlateDecode compressed when saving
• Post-Processing - The compression engine scans saved PDF data and compresses any uncompressed streams using zlib maximum compression
• Cross-Reference Rebuild - Automatically rebuilds the xref table when stream compression changes byte offsets
• Zero Overhead - If all streams are already compressed, the data passes through without modification

Security

• Password Support - Open password-protected PDF documents
• Permissions - Read document and user permissions; the SecurityAudit demo combines permissions, attachments, document JavaScript, URI / Launch links, signatures, XFA, and V8 availability into a read-only risk surface with TXT / JSON report save and TXT copy actions
• Bounded DER Toolkit - FPdfAsn1 encodes low and high tag numbers, typed values, sorted SET OF collections, and canonical time or string forms; its zero-copy reader windows, typed decoders, and resource-limited validator support CMS, PAdES, certificate, and timestamp workflows
• Signature-specific long-term validation - TPadesDssOptions.Vris writes hash-keyed indirect VRI dictionaries while shared certificate, OCSP, and CRL streams are content-hash deduplicated and PAdES B-LT or B-LTA signing creates the new signature's VRI automatically
• Remote and interrupted signing - PreparePadesRemoteSignature persists the SHA-256 document digest, exact ByteRange, placeholder coordinates, and prepared-PDF fingerprint for HSM or cloud signing, while CompletePadesRemoteSignature rejects modified, reused, oversized, or digest-mismatched sessions before writing output
• Signature revision policy analysis - AnalyzeSignatureRevisions reconstructs classic, hybrid, and xref-stream revision chains, applies DocMDP and FieldMDP per signature, and fails closed on malformed, shadow, compressed, or over-budget changes
• Ambiguity-resistant signature profiles - PAdES validation counts required keys only at the top signature-dictionary level, rejects duplicates and inconsistent Sig or DocTimeStamp profiles, and requires ByteRange to exclude exactly the direct hexadecimal Contents token
• Multi-signature evidence export - BuildPadesEvidenceReport correlates PDFium facts, CMS policy, trust and timestamp status, certificate-path hashes, DSS/VRI coverage, and revision decisions with bounded JSON and CSV output
• Archival timestamp renewal - RenewPadesArchiveTimestamp appends an RFC 3161 Document Time-stamp over the exact existing signature history, optionally refreshes the complete DSS first, and validates the resulting B-LTA chain before output
• Asynchronous trust validation - ValidatePadesTrust and ValidatePadesTrustAsync verify original-byte detached signatures, signer and TSA chains, optional online revocation evidence, and RFC 3161 timestamp tokens with staged progress, bounded retrieval, and cooperative cancellation
• Qualified trust - TPdfEuropeanTrustedList parses only caller-authenticated ETSI trusted lists, reauthenticates cached XML, accepts caller-owned update sources, and supplies offline qualified-service evaluation through QualifiedTrustEvaluator
• Security Removal - Save documents with security removed (when authorised)

Printing

• Print Modes - Choose between EMF, text-only, PostScript Level 2 / 3, PostScript pass-through, EMF image-mask, and Type 42 print modes through the global SetPdfPrintMode function
• Viewer Preferences - Surface document author intent through PrintCopies, PrintPageRanges (one-based page ranges), PrintScaling, and PrintPaperHandling; SetPdfPrintPaperHandlingDevMode maps the duplex hint into Windows DEVMODE.dmDuplex
• N-up Printing - The Delphi, C++Builder, and Lazarus PrintPreferences demos show N-up sheet composition (1 / 2 / 4 / 9 pages per sheet) with odd / even page filtering, greyscale rendering, optional 1-pixel cell borders, PDF author print preferences, and duplex intent mapping via SetPdfPrintPaperHandlingDevMode

Shipped Demos

The Demo\Delphi, Demo\CBuilder, and Demo\Lazarus trees ship runnable sample projects that exercise the live API surface across all three IDE families. Highlights:

• PDFViewer / MultiPageViewer / SplitView - End-to-end TPdfView usage: load, navigate, zoom, rotate, two-page spread, cover-aware layouts, and Delphi / C++Builder / Lazarus side-by-side PDF comparison with synchronised commands; the viewer demos are also practical speed tests for multi-gigabyte PDFs, where 2.x GB documents can reach a usable first view almost immediately while page rendering stays responsive
• ViewerInteractionLab - Delphi, C++Builder, and Lazarus focused TPdfView interaction samples for display modes, fit modes, page colours, rotation, search highlighting, text selection, selected-text preview, clipboard actions, and DeviceToPage coordinate feedback
• SearchAndSelect - Delphi, Lazarus, and C++Builder focused workflow samples for all-page search result listing, previous / next match navigation, double-click result jumping, HighlightSearchText, match-case / whole-word search, AllowUserTextSelection, SelectedText, SelectAll, CopySelectionToClipboard, ClearSelection, and spread display modes
• PreflightReport - Reusable FPdfPreflightReport objects collect PDF/A, PDF/UA, PDF/E, PDF/X, PDF/R, and PDF/VT results with status, Info / Warning / Error severity, lookup, search, filtering, deterministic aggregation, stable SHA-256 fingerprints, TXT / HTML / Markdown / JSON / CSV output, and SARIF 2.1.0 / JUnit XML / NDJSON / XML automation formats; Delphi, Lazarus, and C++Builder demos plus console workflows provide previews, batch routing, summaries, report attachments, and CI exit-code gating
• ContentExtractionLab - Delphi, Lazarus, and C++Builder unified content inventory workflow for document metadata, page text, page objects, images, attachments, links, annotations, bookmarks, and font / character metrics with TXT / JSON export, all / none category selection, source-PDF save, and report copy actions
• WatermarkStamp - Delphi, Lazarus, and C++Builder reusable Form XObject workflow for centred watermarks, corner stamps, and page number labels across a multi-page target PDF
• CreatePDF - Build a multi-page PDF from scratch with text, images, paths, tables, and watermark composition; includes a "European Text" option to write accented Czech / French / German / Spanish / Portuguese / Polish / Turkish text for verifying font selection and Unicode handling
• ExtractText / ExtractImages / ContentExtractionLab - Save extracted text (including per-page separated output) or embedded images to disk; or generate a multi-category content inventory report in one pass
• FormFill - Read AcroForm field metadata, write values, regenerate /AP appearances, and save either the editable or the flattened result
• FontProperties - Inspect indexed character / font metadata through both TPdf and TPdfView in Delphi, C++Builder, and Lazarus
• MergePDF / SplitPDF - Combine multiple source PDFs (with per-file page range selection, optional output path, run log, cancel button, and merge completion summary) or split one PDF into per-page output files
• PDFtoJpg / BatchExport - Delphi, C++Builder, and Lazarus one-shot or batched multi-file PDF to JPG conversion with per-file overall progress, per-page current progress, DPI / quality / password inputs, and a cooperatively-cancellable batch backed by IPdfCancellationToken
• StandardsLab - Delphi, C++Builder, and Lazarus workflow samples for PDF/A, PDF/UA, PDF/E, PDF/X, PDF/R, and PDF/VT validation, issue reporting, and standards-oriented save outputs
• PrintPDF / PrintPreferences - Delphi, C++Builder, and Lazarus standard PDF printing with preview, page ranges, copies, collate, progress, and cancellation (PrintPDF), plus Delphi, C++Builder, and Lazarus N-up sheet composition with odd / even filtering and PDF author print preferences (PrintPreferences)
• SecurityAudit - Delphi, C++Builder, and Lazarus read-only risk-surface samples for permissions, attachments, JavaScript actions, URI / Launch links, web links, signatures, XFA state, V8 / XFA helper availability, and TXT / JSON audit reports
• Attachment - Create PDFs, add / list / extract / delete / save embedded file attachments with MIME subtype display
• ImageToPDF - Delphi, C++Builder, and Lazarus image batch conversion into one PDF, one image per page, with preview and per-image progress

The Delphi, C++Builder, and Lazarus regression suites cover 147 (Pas) / 123 (C++) cases sharing the same fixture PDFs so behaviour stays byte-aligned across all three IDE families

Platform Support

• Delphi - Delphi XE5 (RAD Studio XE5) through Delphi 13.1 Florence (RAD Studio 13.1)
• C++Builder - C++Builder XE5 (RAD Studio XE5) through C++Builder 13.1 Florence
• Lazarus - Free Pascal / Lazarus support
• Win32 / Win64 - Both 32-bit and 64-bit Windows platforms