Why can this converter open XPS and OXPS files on Mac?
To open an XPS file on Mac, iPhone, or Linux without booting Windows, this converter unpacks .xps and .oxps ZIP containers locally inside a Comlink Web Worker and translates XAML FixedPage markup into standard multi-page PDF files. Confidential tax statements, invoices, and payroll printouts stay strictly inside your browser tab.
macOS Preview and standard browsers omit native viewers for Windows Presentation Foundation fixed-layout documents defined in the Library of Congress MS-XPS specification notes. Dropping a .xps or .oxps container locally avoids uploading corporate reports, HR forms, or vendor printouts to third-party Windows rendering queues.
Load any local document via Choose XPS / OXPS (or test Load sample XPS, which generates two-page dispatch-note-sample.xps), swap files anytime with Change file, and start conversion with Convert XPS to PDF or cancel mid-run via Cancel. LibreOffice packages can also route to ODT to PDF.
Comparing XPS vs OXPS namespaces and ZIP package discovery
Windows Vista and 7 generated legacy .xps files using Microsoft’s 2005 XML namespaces, whereas Windows 8, 10, and 11 shifted the Microsoft XPS Document Writer toward .oxps archives governed by the ECMA-388 OpenXPS specification. Because XML element names remain consistent across both versions, the parser normalizes tags using localName() so .xps and .oxps packages run through the same renderer.
| Container aspect | Legacy MS-XPS (.xps) |
ECMA-388 OpenXPS (.oxps) |
Implemented converter handling |
|---|---|---|---|
| Primary XML namespace | schemas.microsoft.com/xps/2005/06 |
schemas.openxps.org/oxps/v1.0 |
Normalized transparently via localName() |
| Entry sequence pointer | FixedDocumentSequence.fdseq |
FixedDocSeq.fdseq / *.fdseq |
Read via _rels/.rels fixedrepresentation first |
| Package entry ceiling | ZIP Open Packaging container | ZIP Open Packaging container | Max 5,000 entries and 250 MB per entry |
| Cumulative unpack budget | Uncompressed XAML + artwork | Uncompressed XAML + artwork | 500 MB uncompressed archive limit |
| Non-XPS office archives | Misnamed DOCX or ODT ZIPs | Misnamed DOCX or ODT ZIPs | Halts and identifies actual file format |
Traversing _rels/.rels, FixedDocumentSequence, and PageContent
Package discovery begins by inspecting _rels/.rels for a fixedrepresentation relationship target, falling back to FixedDocumentSequence.fdseq, FixedDocSeq.fdseq, or any .fdseq file at package root. From there, the worker resolves every DocumentReference to its FixedDocument parts and collects ordered PageContent Source paths, falling back to numerically sorted *.fpage files if relationships are incomplete.
Archive size limits and DOCX or ODT signature detection
Unpacking delegates to extractZipArchive, enforcing a 5,000-entry count cap, 250 MB single-file ceiling, and 500 MB cumulative uncompressed threshold to guard memory on client hardware. Before throwing a missing-page error on mismatched ZIP files, the inspector checks for word/document.xml or content.xml plus mimetype and advises users to convert DOCX documents to PDF directly instead of forcing an office container through an XPS parser.
Mapping 96 DPI FixedPage geometry to 72 pt PDF pages
Microsoft’s FixedPage coordinate grid defines layout at 96 logical pixels per inch, whereas the Library of Congress PDF 1.7 format specification defines PDF user space at 72 points per inch. Applying XPS_TO_PT = 0.75 converts page widths, heights, strokes, and FontRenderingEmSize values cleanly, turning a fallback 793.76 x 1122.52 px page into exact 595.32 x 841.89 pt A4 output.
Because XPS positions origin (0, 0) at the top-left corner while PDF anchors (0, 0) at bottom-left, the builder flips Y coordinates while multiplying nested Canvas and element RenderTransform 2D affine matrices [a, b, c, d, e, f] alongside inherited Opacity levels.
Vector paths, ARGB colors, and SVG Data normalization
Form borders, underlines, dividers, and tinted panels reside in <Path> nodes using #RRGGBB or #AARRGGBB hex colors for Fill and Stroke attributes. Before invoking drawSvgPath, the geometry converter parses StrokeThickness and removes F0 or F1 winding-rule prefixes from Data strings so segments follow the W3C SVG 2 path data grammar accurately.
Embedding PNG and JPEG ImageBrush fills accurately
Embedded raster signatures, QR codes, and product logos inside Path.Fill or ImageBrush declarations are matched by ImageSource path and inspected via magic-byte headers (0x89504E47 for PNG and 0xFFD8 for JPEG). Computed extractRectBounds rectangles scale each asset into transformed PDF bounding coordinates, while multi-page scanned bitmaps stored outside XPS envelopes convert faster as raster TIFF print-scan pages as PDF.
Glyphs typography, StyleSimulations, and honest format limits
Text runs live inside <Glyphs> elements carrying UnicodeString, OriginX, OriginY, FontRenderingEmSize, and StyleSimulations attributes. Rather than executing untrusted obfuscated .odttf font subsets directly inside a PDF engine, this client renderer maps simulated weights to predictable standard PDF fonts and surfaces exact diagnostic counts whenever an asset cannot be embedded faithfully.
StyleSimulations font fallback and character replacement counts
When <Glyphs> specifies BoldSimulation, ItalicSimulation, or both, the text emitter switches between StandardFonts.Helvetica, HelveticaBold, HelveticaOblique, and HelveticaBoldOblique. Strings pass through cleanPdfText, folding accents when possible and counting characters outside the standard Latin encoding that are replaced with ? so a yellow warning banner reports the exact replacement tally on completion.
Skipped non-PNG or non-JPEG assets like JPEG XR bitmaps
Enterprise Windows print drivers occasionally serialize scanned backgrounds as JPEG XR (HD Photo) or TIFF streams rather than PNG or JPEG images inside ImageBrush parts. Rather than crashing page assembly mid-document, unsupported image blobs increment skippedImages gracefully so surrounding Glyphs text and vector tables still export cleanly.
Conversion boundary check: Verify non-Latin text reports (
?replacement banner) and heavily stylized CAD exports visually after downloading. Fonts relying on custom glyph indices (Indicesoffsets without completeUnicodeStringattributes) or JPEG XR bitmaps fall outside client standard-font rendering without desktop WPF drivers.
What happens after converting an XPS document to PDF?
Once Convert XPS to PDF finishes inside the Comlink worker, the result panel presents rendered page totals, original .xps or .oxps byte sizes versus output .pdf sizes, and a direct Download control. Built-in workflow shortcuts (Compress PDF, Merge PDFs, and Split PDF) let you optimize or combine outputs without re-uploading files.
- Shrink image-heavy printouts: Route large invoice or catalog PDFs containing high-resolution PNG artwork straight through Compress PDF to reduce archive size before sharing or long-term storage.
- Assemble multi-source dossiers: Combine converted XPS pages with scanned forms via Merge PDFs, break multi-invoice spools with Split PDF, or pair print archives with offline MBOX email archiving without mail clients when organizing client correspondence.
- Verify flagged conversion warnings: Review any non-Latin
?replacement notice or skipped-image tally reported in the summary banner, retaining source.xpscopies whenever Windows print drivers embed unsupported JPEG XR images.