Issues

• [Bug]: long lines/paragraphs are not word-wrapped and thus overflow when converting markdown/html to pdf

  bug
  yfedoseev/pdf_oxide#3088 · Filter by author aryan-79 opened on Nov 6, 2026

• [Bug]: Space inserted between adjacent text objects for a gap that inside one TJ is not a space

  bug
  yfedoseev/pdf_oxide#3087 · Filter by author Tsafaras opened on Nov 5, 2026 · v0.3.79

• [Feature]: Support External Signing Providers for PAdES Signatures

  enhancement
  yfedoseev/pdf_oxide#3086 · Filter by author buonhobo opened on Nov 5, 2026 · v0.4.0

• [Perf]: Each output format redoes the page's layout: only spans are reused across extract_text, to_markdown and to_html

  performance
  yfedoseev/pdf_oxide#3083 · Filter by author yfedoseev opened on Nov 5, 2026 · v0.3.79
  1

• [Perf]: A Type0 glyph's width allocates a Vec and re-tests the encoding name on every glyph

  performance
  yfedoseev/pdf_oxide#3079 · Filter by author yfedoseev opened on Nov 5, 2026 · v0.3.79
  1

• [Perf]: is_multi_column_page is evaluated four times per page, cloning the page's spans each time

  performance
  yfedoseev/pdf_oxide#3074 · Filter by author yfedoseev opened on Nov 5, 2026 · v0.3.79
  1

• [Perf]: postprocess_spans computes an XY-cut reading order on every page that assemble_text_from_spans then mostly replaces: 19% of a paper's extraction

  performance
  yfedoseev/pdf_oxide#3073 · Filter by author yfedoseev opened on Nov 5, 2026 · v0.3.79

• [Perf]: Pages served from the per-page pre-read are walked a second time in character mode for glyph origins: 22% of a book's extraction

  performance
  yfedoseev/pdf_oxide#3072 · Filter by author yfedoseev opened on Nov 5, 2026 · v0.3.79
  1

• [Perf][Regression]: Text extraction compiles the same ICC colour transform again for every page and pass: 91% of a book's extraction time

  bugperformance
  yfedoseev/pdf_oxide#3071 · Filter by author yfedoseev opened on Nov 5, 2026 · v0.3.79
  5

• [Perf]: Row snapping and the XY-cut dominate text-heavy and OCR pages: B-tree structures, repeated calls and re-sorting at every level

  performance
  yfedoseev/pdf_oxide#3070 · Filter by author yfedoseev opened on Nov 5, 2026 · v0.3.79
  2

• [Perf]: Plain text extraction parses the whole structure tree, cloning every element, to decide whether one page's order can be trusted

  performance
  yfedoseev/pdf_oxide#3069 · Filter by author yfedoseev opened on Nov 5, 2026 · v0.3.79
  3

• [Perf]: About 10% of a cold text extraction is allocation and copying: cloned span vectors, strings and objects in the assembly passes

  performance
  yfedoseev/pdf_oxide#3067 · Filter by author yfedoseev opened on Nov 5, 2026 · v0.3.79
  3

• [Perf]: Font loading decodes and parses every embedded font program even when the font dictionary already fixes every code

  performance
  yfedoseev/pdf_oxide#3066 · Filter by author yfedoseev opened on Nov 5, 2026 · v0.3.79
  1

• [Perf]: extract_text runs table detection on every page, re-parsing the content stream for paths

  performance
  yfedoseev/pdf_oxide#3063 · Filter by author yfedoseev opened on Nov 5, 2026 · v0.3.79
  3

• [Perf]: Glyph widths and code maps hash small integer keys with SipHash: about 4% of a cold text extraction

  performance
  yfedoseev/pdf_oxide#3062 · Filter by author yfedoseev opened on Nov 5, 2026 · v0.3.79
  1

• [Bug]: On a /Rotate 180 page, extract_words and extract_text_lines turn their boxes to the displayed page, but extract_chars and extract_spans don't

  bug
  yfedoseev/pdf_oxide#3060 · Filter by author ankursri494 opened on Nov 5, 2026 · v0.3.79

• [Bug]: panic within bimodal_gap_split with "comparison function does not correctly implement a total order"

  bug
  yfedoseev/pdf_oxide#3059 · Filter by author MattDavis00 opened on Nov 2, 2026 · v0.3.79

• [Python]: Error hierarchy with code, kind and op, not rooted in RuntimeError (follow-up to #2491)

  ai-assistedenhancement
  yfedoseev/pdf_oxide#3057 · Filter by author yfedoseev opened on Oct 30, 2026 · v0.4.0

• [JS]: Dual ESM and CommonJS build, optional platform dependencies, and a pack-time prebuild check (follow-up to #2499)

  ai-assistedenhancement
  yfedoseev/pdf_oxide#3056 · Filter by author yfedoseev opened on Oct 30, 2026 · v0.4.0

• [Bug]: Separately placed drawing labels are glued on a schematic page

  ai-assistedbug
  yfedoseev/pdf_oxide#3051 · Filter by author yfedoseev opened on Oct 30, 2026 · v0.3.79

• [Feature]: html_css layout: floats, multi-column, z-index, border-collapse, inline-block and sticky

  ai-assistedenhancement
  yfedoseev/pdf_oxide#3040 · Filter by author yfedoseev opened on Oct 30, 2026 · v0.3.79

• [Feature]: html_css inline layout: inline images, inline box decorations and soft hyphens

  ai-assistedenhancement
  yfedoseev/pdf_oxide#3039 · Filter by author yfedoseev opened on Oct 30, 2026 · v0.3.79

• [CI]: Release gate for the speed claim against the leading Python PDF library

  ai-assistedciperformance
  yfedoseev/pdf_oxide#3036 · Filter by author yfedoseev opened on Oct 30, 2026 · v0.3.79
  1

• [Compliance]: 52 declared PDF/A, PDF/X and PDF/UA rules are never checked

  ai-assistedenhancement
  yfedoseev/pdf_oxide#3035 · Filter by author yfedoseev opened on Oct 30, 2026 · v0.3.79

• [Packaging]: Swift Package.swift links the native library through unsafeFlags with relative paths

  ai-assistedenhancement
  yfedoseev/pdf_oxide#3031 · Filter by author yfedoseev opened on Oct 30, 2026 · v0.3.79
  1