Optional module · pdftoolkit/pdf-codecs 0.1.0
English | Français
Unreleased
0.1.0 - 2026-10-28
- Introduce the independent optional
pdftoolkit/pdf-codecspackage for 64-bit PHP 8.3+. Production dependencies contain neither PDFToolkit core/Reader nor native programs or image extensions. - Add explicit decoder/request/limit objects and immutable component planes retaining precision, signed samples, sampling geometry, image origin, colour alternatives, ICC data, alpha/channel associations and chroma-key metadata.
- Decode PDF eight-bit sequential/progressive JPEG with exact quantized coefficients, restart handling, component sampling and explicit PDF colour-transform precedence.
- Decode PDF JPEG 2000 component streams and containers, including packet organizations, wavelet transforms, declared precision, palette/channel mapping and colour metadata. The bilingual JPEG 2000 matrix records supported Part 1/Part 2 mechanisms and exclusions.
- Decode PDF bilevel JBIG2 embedded/global streams, arithmetic/Huffman/MMR regions, symbol dictionaries, refinement, text, patterns, halftones, retained contexts and the extended adaptive template.
- Apply per-request input/output, allocation, work, geometry, metadata, segment and nesting budgets. Keep malformed data, unsupported mechanisms and resource exhaustion as separate typed errors; do not silently reduce samples or relax limits.
- Retain Apache-2.0 PDF.js and MIT Reader adaptation notices, exact source identities and fixture provenance. Provide independent JPEG/OpenJPEG/JBIG2/Poppler comparisons, explicitly labelled normative vectors, malformed-input regressions and PHP 8.3/8.4/8.5 qualification.
- Include a local-file example, English/French codec and architecture guides, security reporting and an independent release workflow. Native discovery and native-first/PHP-fallback selection belong to Reader's adapter.