Skip to content

CHANGELOG.md

Latest commit

 

History

History
136 lines (90 loc) · 5.12 KB

CHANGELOG.md

File metadata and controls

136 lines (90 loc) · 5.12 KB

Changelog

v0.1.7

New features

  • --json — emit JSON Lines output for machine parsing. Each record has a type field (match, context, path, summary, warning) with structured data including submatches with byte offsets, truncation metadata, and file size warnings

Performance

  • Optimized search hot path and walker scheduling for faster parallel traversal

Build

  • Build script always recomputes git SHA for accurate --version output

Testing

  • 207 integration tests (up from 204)

v0.1.6

New flags

  • -q / --quiet / --silent — suppress all output, exit with status 0 on match
  • -h / --no-filename — suppress filename prefix in multi-file output
  • -H / --with-filename — force filename prefix even for single file
  • --exclude-dir=GLOB — skip directories matching the glob pattern
  • -L / --files-without-match — print names of files with no matches
  • -m NUM / --max-count=NUM — stop reading a file after NUM matches
  • --colour — alias for --color
  • -s / --no-messages — suppress error messages about nonexistent or unreadable files
  • --group-separator=SEP / --no-group-separator — customize or disable the context group separator
  • -b / --byte-offset — print byte offset of each matching line
  • -I — equivalent to --binary-files=without-match
  • -f FILE / --file=FILE — read patterns from a file (one per line)
  • --no-ignore-case — cancel a preceding -i
  • -x / --line-regexp — match only whole lines
  • --label=LABEL — use LABEL as filename for stdin
  • -T / --initial-tab — align content after prefix with a tab (GNU grep compatible field widths)
  • -Z / --null — print NUL byte after filenames
  • --exclude-from=FILE — read exclude globs from a file
  • -a / --text — process binary files as text
  • -U / --binary — do not strip CR characters (no-op on Unix, accepted for compatibility)

Improvements

  • Baseline benchmark separated from default cargo bench
  • README updated with -o/-A/-B/-C examples and corrected unsupported flags list

Testing

  • 204 integration tests (up from 125)

v0.1.5

Improvements

  • Trigram index now carries a format version; incompatible or corrupted indexes are automatically deleted and rebuilt
  • --version output shows the index format version

Testing

  • Integration tests split into 9 focused modules (basic matching, flags, files/recursive, stdin, edge cases, binary, only-matching, context, regression)
  • Added rstest for parametrized test cases, reducing boilerplate

v0.1.4

New features

  • -o / --only-matching — print only the matched parts of a line, each on its own line
  • -A NUM / --after-context — print NUM lines of trailing context after each match
  • -B NUM / --before-context — print NUM lines of leading context before each match
  • -C NUM / --context — print NUM lines of context before and after each match
  • Group separators (--) between non-contiguous context blocks (GNU grep compatible)
  • Context lines use - separator instead of : (GNU grep compatible)

Improvements

  • --version now shows git commit hash (or release for crates.io installs), crates.io link, and feature list
  • --help updated with AI-agent-optimized messaging

Testing

  • 24 new integration tests (125 total)

v0.1.3

Packaging

  • Release binaries are now distributed as .tar.gz archives
  • Added one-line install commands for Linux (musl) and macOS in README

v0.1.2

CI/CD

  • Reusable workflow architecture (_lint.yml, _test.yml, _build.yml)
  • Release pipeline: lint → test → build → smoke-test → publish (crates.io + GitHub Releases)
  • Nightly toolchain for rustfmt and clippy in CI
  • Release tests run under --release profile
  • Added x86_64-unknown-linux-musl and aarch64-unknown-linux-musl build targets
  • Parallel lint jobs (format, clippy, docs run concurrently)
  • GitHub Releases with platform binaries and changelog

v0.1.1

New features

  • File size limit — files larger than 100 MiB are skipped by default with a machine-readable warning to stderr. Override with FASTGREP_NO_LIMIT=1 or --max-file-size=<BYTES>. Prevents grep from hanging on accidental large binaries in the repo.

Performance

  • New parallel directory walker with cooperative work-stealing across 2–4 threads
  • Performance optimizations in search pipeline

Testing

  • Expanded integration test suite (101 tests covering flags, edge cases, binary detection, color output)

Documentation

  • Added install instructions and usage examples to README
  • Added ENVIRONMENT.md with full reference of env vars and CLI flags
  • Added AI_AGENT_GREP_USECASES.md documenting common grep patterns used by AI coding agents
  • Documented all differences from GNU grep behaviour

v0.1.0

Initial release.

  • GNU grep-compatible interface (-r, -i, -n, -l, -c, -v, -w, -E, -F, --include, --exclude, --color)
  • Parallel search across all CPU threads
  • SIMD-accelerated literal and prefix search via memchr::memmem
  • Lazy trigram index with automatic invalidation
  • Line truncation (--max-line-len, default 15000)