Skip to content

CHANGELOG.md

Latest commit

 

History

History
121 lines (98 loc) · 5.77 KB

CHANGELOG.md

File metadata and controls

121 lines (98 loc) · 5.77 KB
Raw

Changelog

[v0.4.0] - 2026-01-01

Features

  • Add GitHub Actions workflow, retire Bitbucket notifications (#90)
  • Wire gcov coverage into the build (#82)

Bug Fixes

  • Honor json from config file and MMTREADER_JSON environment variable (#98)
  • Report product version distinctly from SDK version in --version banner (#93)
  • Rename shadowed rc variable in online capture loop (#89)
  • Restore .deb fallback via compgen glob expansion (#86)
  • Remove eval-built command strings and add adversarial smoke proof (#85)
  • Graceful SIGTERM stop and extraction-failure summary (#84)
  • Fix config precedence for env vars and prevent buffer overflow (#83)
  • Restore Jenkins test stage running make test (#79)
  • Groundwork: Jenkins make, SDK pin, test wiring (#78)

Refactoring

  • Named dispatch modes, PROTO_META, PRIu64 (#92)
  • Split parse_options into layered helpers (#91)
  • Dead-code sweep across engine, cli, config, colors (#88)
  • Extract shared assert header tests/test_util.h (#80)

Documentation

  • Regenerate user-guide and architecture samples from real output (#99)
  • Align README and DEVELOPMENT docs with the real build (#95)
  • Document cap_net_raw capability tradeoff (#81)
  • Add CLAUDE.md and AGENTS.md agent config files (#77)
  • Document agent install/run notes for the C toolchain (#76)

Build / CI

  • Enable -Wall -Wextra and verify the tree is warning-free (#94)
  • Pin base image and bump toolchain to current GCC (#87)

[Unreleased]

Fixed

  • Honor the json config-file key and the MMTREADER_JSON environment variable (#96). cli_options_t carried two fields meaning one thing — output_format, which every output decision reads, and json, whose only reader was a verbose debug printf. The config and environment paths wrote json alone, so json = 1 and MMTREADER_JSON=1 were silently ignored; only -j/--json, which wrote both, had any effect. json is gone and output_format is now the single carrier.

  • Apply -c/--config files before the CLI flags are parsed (#96). The named config file used to be re-read after the option loop and re-copied unconditionally, so it beat explicit flags (-b 100 --config <buffer=777> used 777) and skipped the environment entirely — the opposite precedence from ~/.mmtreader.conf. Both files now load before the flags, under one rule: compiled defaults < ~/.mmtreader.conf < --config file < environment < CLI flags. See docs/CONFIG.md.

  • Report mmtReader's own product version distinctly from the MMT-DPI SDK version (#70, F-BUG-005). --version and the startup banner now print two labeled lines, and the product version is injected at build time via make MMTREADER_VERSION=... (-DMMTREADER_VERSION).

Changed

  • Breaking (JSON): the --json "version" field changed from a bare string holding the SDK version to an object: {"mmtreader": "0.4.0", "mmt_dpi": "1.8.0 (42cac8b7)"}. Consumers reading .version as a string must read .version.mmt_dpi instead.

Documentation

  • Align README, docs/DEVELOPMENT.md, docs/TESTING.md, AGENT_ENVIRONMENT.md and the mmt-reader skill with the real build (#72, F-DOCS-001, F-DOCS-002): the manual compile command now names flows.c and links, libconfuse is gone as a prerequisite (config.c has its own INI parser), the -Wall -Wextra warning gate and -DMMTREADER_VERSION injection are documented, and every test total is re-measured from an actual run (14 numbered groups, 299 unit asserts, 52/52 CLI checks, 6/6 SDK checks).
  • Replace the invented README sample output and JSON sample with output captured from ./mmtReader analyze -t smallFlows.pcap -a, and drop a duplicated Usage Examples block.
  • Regenerate every sample block in docs/USER_GUIDE.md from real output (#97). The guide was left out of #72 and still showed a TCP.HTTP.Google protocol path format, Input:/Sessions: labels that cli/output.c never prints, a JSON sample predating both the protocol_paths[] split and the version object from #70, and a whole "PCAP Statistics" section for a feature that does not exist — pcap_stats() is never called. Every block is now spliced from an actual ./mmtReader analyze -t smallFlows.pcap run; every command shown carries its required analyze/capture subcommand and points at the bundled smallFlows.pcap, so it runs verbatim and exits 0. The config and environment sections now agree with docs/CONFIG.md on which keys are live and on the defaults < ~/.mmtreader.conf < --config < environment < CLI precedence fixed in #96.
  • Correct two identifiers in docs/ARCHITECTURE.md (#97): the sorted-list helper is proto_info_insert() (cli/output.c:104), not insert_proto_info(), and engine_print_stats_ex() lives in core/engine.c:281, not cli/output.c. Its output sequence also no longer lists the non-existent PCAP-statistics section.

[v0.3.0] - 2026-01-01

Added

  • Add mmtReader AI agent skill for network traffic analysis (#34) (a154152)

Fixed

  • Stop engine_destroy() from printing the statistics summary (#39) (#42) (089be40)
  • Take traffic statistics from the MMT-DPI API instead of recomputing them (#40) (c2927d4)
  • Route live capture through the engine's packet accounting (#37) (ae0c0fb)
  • Remove double quotes from Mermaid node labels and message text to fix parser errors (72ffba0)
  • Use <br/> instead of \n for line breaks in Mermaid sequence diagrams (f32e946)

Testing

  • Pin input_stats to MMT-DPI's per-protocol accounting (#38) (#43) (045bce9)

Documentation

  • Fix README and PLAYBOOK to match actual binary output (3f503c0)
  • Add Use Case 6 — AI Agent-Assisted Analysis (5a91a04)
  • Add MMT-Reader AI Agent Playbook (b646712)
  • Rewrite README as AI Agent-focused landing page with visual workflow (0277af4)