Skip to content

CHANGELOG.md

Latest commit

 

History

History
140 lines (75 loc) · 3.65 KB

CHANGELOG.md

File metadata and controls

140 lines (75 loc) · 3.65 KB
Raw

Changelog

All notable changes to this project are documented here. Versions follow Semantic Versioning. Format follows Keep a Changelog.

[0.5.0] — 2026-10-15

Added

  • dom: Keep mixed content and read CDATA sections literally ⚠ BREAKING CHANGE

  • parser: Report malformed input and file errors through iostat/iomsg

  • parser: Keep the prolog and write it back in stringify

  • dom: Save the document to a file with xml_file%save

  • parser: Keep comments and PIs inside the document

  • parser: Optionally keep leading/trailing spaces of attribute values

  • index: Scan files and read tag content on demand

  • writer: Stream XML to a file, a unit or memory

  • writer: Flush the pending start tag for direct writes to the unit

  • values: Read attributes and contents as numbers and logicals

  • dom: Add tags under a parent with add_tag(parent=, id=)

  • tree: Resolve namespaces and find tags by namespace

Changed

  • dom: Share tree queries between DOM and index

  • parser: Drive the tokenizer through a resumable feed and a builder

  • tokenizer: Make the tokenizer and the builders pure

  • tag: Parse single tags and attribute streams with the tokenizer ⚠ BREAKING CHANGE

Documentation

  • guide: Document the tokenizer, mixed content, CDATA and references

  • guide: Document error reporting, the prolog and xml_file%save

  • guide: Document xml_index, chunked parsing and tag accessors

  • guide: Document the streaming writer

  • guide: Document typed values for attributes and contents

  • guide: Document single-tag parsing and nested add_tag

  • guide: Document compiler support and the fobis fetch setup

  • guide: Document namespace resolution and queries

Fixed

  • parser: Replace string-search parsing with a single-pass tokenizer

  • parser: Stop reading past the source end in xml_tag tag search

  • dom: Normalize literal blanks in attribute values on read

  • parser: Support sources larger than 2 GiB ⚠ BREAKING CHANGE

  • portability: Make FoXy work with ifx and nvfortran

Performance

  • parser: Parse files in chunks instead of loading them whole

[0.4.0] — 2026-10-14

Fixed

  • dom: Decode character references in content and attribute values ⚠ BREAKING CHANGE

Performance

  • bench: Add parse/stringify/lookup benchmark with fobos modes

  • dom: Make parse, stringify and lookups linear in the tag count

[0.3.0] — 2026-10-14

Added

  • dom: Find tags by name, path or parent and read them by ID

Fixed

  • docs: Untrack package-lock and pin esbuild for lock-free vite build

  • docs: Set vite build target to es2022 for version-robust doc builds

  • ci: Run install smoke test from the release workflow

  • build: Exclude third-party docs/ and scripts/ sources from builds

  • parser: Match nested same-name end tags in source order

  • parser: Stop self-closing tags inheriting the previous tag content

  • dom: Keep tag IDs consistent across delete_tag and add_tag

[0.2.8] — 2026-03-02

Fixed

  • install: Run FoBiS.py fetch after wget tarball extraction

[0.2.6] — 2026-03-02

Fixed

  • install: Derive extracted dir from tarball name to avoid broken pipe

[0.2.5] — 2026-03-02

Added

  • install: Add general-purpose install script

[0.2.2] — 2026-03-02

Documentation

  • Refactor README style and update install instructions

[0.2.1] — 2026-02-20

Documentation

  • Add VitePress site and overhaul project infrastructure

[0.2.0] — 2025-06-25

Fixed

  • Fix parse support of nested and repeated tags