AntidoteDB/antidote

This system is a distributed database platform built on Erlang, centered around the Antidote transactional CRDT engine. It provides core data structures for conflict-free replication, including counters, registers, and sets, alongside a Protocol Buffers-based client library for external communication. The architecture is supported by internal monitoring via Prometheus metrics and vector clocks for distributed state tracking, all managed through a modernized Rebar3 build and release infrastructure.

The \antidote\_crdt\ application is introduced as a standalone Erlang library providing operation-based CRDT implementations. The library exposes a unified API for creating, updating, and reading values for various CRDT types, including PN-Counters, Bounded Counters, Fat Counters, Enable/Disable-wins Flags, Grow-Only/Add/Remove-wins Sets, Last-Writer-Wins Registers, Multi-Value Registers, and Grow-Only/Recursive-Reset Maps. The diff includes the full source code for these implementations, along with build configuration (\rebar.config\, \Makefile\), CI workflows, and licensing/authorship metadata.

_apps/antidote\crdt · high confidence

The \antidotec\_pb\ application is introduced, providing the Erlang client library for AntidoteDB via the Protocol Buffers interface. This includes the core socket management (\antidotec\_pb\_socket\), transactional operations (\antidotec\_pb\), and management functions (\antidotec\_pb\_management\). It also adds support for CRDT data types such as counters (\antidotec\_counter\), registers (\antidotec\_reg\), and sets (\antidotec\_set\) with corresponding EUnit tests. The application is licensed under Apache 2.0 and depends on \antidote\_pb\_codec\.

_apps/antidotec\pb · high confidence

The repository now includes two new Erlang applications: \antidote\_stats\, which provides a stats collector and Prometheus metrics endpoint for monitoring, and \vectorclock\, a library for managing vector clocks in Erlang. These additions establish the foundational components for internal monitoring and distributed state tracking within the monorepo structure.

_apps/antidote\stats, apps/vectorclock · high confidence

The Antidote application is introduced as part of the new mono-repository structure, providing the core transactional CRDT database functionality. This includes the main API for transaction management (start, commit, abort), object operations (read, update), and hooks. The release also establishes the inter-datacenter replication protocol, defines the message types for inter-DC communication, and sets up the application's supervision tree and configuration defaults.

apps/antidote · high confidence

The release directory has been cleaned up by deleting the app.config, erl, floppy, floppy-admin, nodetool, and vm.args files. These files previously provided the infrastructure for managing the Erlang runtime, cluster administration, and node control. Their removal indicates a shift away from these specific legacy deployment and operational scripts.

rel/files · high confidence

The legacy Floppy application and its associated components have been removed from the source tree. This includes the main application module (floppy\_app.erl), the supervisor tree (floppy\_sup.erl), the vnode implementation (floppy\_vnode.erl), and related console and event handler modules. This change eliminates the old implementation, likely as part of a migration to the ClockSI architecture.

src · high confidence

The file rel/vars/dev\_vars.config.src, which previously defined development configuration values such as the web and handoff ports, node name, and cookie, has been removed from the repository.

rel/vars · high confidence

The release build system for the 'floppy' application has been removed. This includes the deletion of the 'rel' directory containing the 'gen\_dev' script, 'reltool.config', and 'vars.config'. These files were responsible for generating development environment configurations and defining the release structure for the 'floppy' application. Users relying on these scripts for local development setups or release generation will no longer have access to this legacy tooling.

rel · high confidence

The project has migrated from Rebar2 to Rebar3, updating the build system to use \rebar3\ and \relx\ for release management. This change includes adding a \rebar.config.script\ for CI integration, configuring the \rebar3\_lint\ and \erlfmt\ plugins for code style enforcement, and updating the \rebar.lock\ with current dependency versions. The \Makefile\ has been updated to use \rebar3\ commands and remove legacy dev cluster targets, while \.gitignore\ and \elvis.config\ are added to support the new tooling.

(repo-wide) · high confidence

Added new configuration files (network.config, sys.config.src, vm.args.src) that define network ports for Riak Core and Antidote, set up the Erlang logger with file and standard IO handlers, and configure the Erlang VM with node names and cookies. These files replace previous configuration methods and provide a structured way to manage Antidote's runtime and network settings.

config · medium confidence

A new automated test script has been added to the test suite. This script builds the Antidote release, starts the node, and executes a transaction to verify that the release is functional. The test includes an Erlang script that connects to the running node, performs a counter operation, and ensures the system responds correctly.

test · high confidence

Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

• CAI 59 → 58 (-1.0)
• Rubric changed (rubric-2026.08.19 → rubric-2026.09.15) — scores are not directly comparable.

• Code Health 100 → 91 (-9.3)
• Architecture 100 → 100 (+0.0)
• Maturity 45 → 44 (-0.4)
• Readiness 64 → 51 (-13.4)
• Security 61 → 81 (+20.3)
• Event Sourcing 100 → 100 (+0.0)

• Coverage not included — suite not readable by the collector
• Dependency hygiene not measured — dependency manifest found but not parsed for hygiene
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• No exposed public API
• Test reliability not included
• complexity unreadable for .erl, .hrl — churn × complexity hotspots could not be measured
• dormant codebase — no living knowledge left to concentrate

• Coverage not measured — no coverage collector is wired up
• Dependency not covered by the lockfile: antidote_crdt
• Dependency not covered by the lockfile: antidote_pb_codec
• Dependency not covered by the lockfile: antidote_pb_codec
• Dependency not covered by the lockfile: antidote_stats
• Dependency not covered by the lockfile: vectorclock
• Duplicated block (10–11 lines × 2) (apps/antidote/src/clocksi_interactive_coord.erl)
• Duplicated block (11–12 lines × 2) (apps/antidotec_pb/src/antidotec_pb.erl)
• Duplicated block (12 lines × 2) (apps/antidote/src/clocksi_vnode.erl)
• Duplicated block (14 lines × 2) (apps/antidote_crdt/src/antidote_crdt_set_aw.erl)
• Duplicated block (5 lines × 2) (apps/antidote/src/logging_vnode.erl)
• Duplicated block (5 lines × 2) (apps/antidote_crdt/src/antidote_crdt_map_go.erl)
• Duplicated block (5 lines × 2) (apps/antidote_crdt/src/antidote_crdt_map_go.erl)
• Duplicated block (6 lines × 2) (apps/antidote_crdt/src/antidote_crdt_map_go.erl)
• Duplicated block (6 lines × 2) (apps/antidote_crdt/src/antidote_crdt_set_aw.erl)
• Duplicated block (7 lines × 2) (apps/antidote/src/antidote_hooks.erl)
• Duplicated block (7 lines × 2) (apps/antidote/src/clocksi_vnode.erl)
• Duplicated block (7 lines × 2) (apps/antidote_crdt/src/antidote_crdt_set_aw.erl)
• Duplicated block (7 lines × 4) (apps/antidote/src/inter_dc_query_dealer.erl)
• Duplicated block (9 lines × 2) (apps/antidote_crdt/src/antidote_crdt_set_aw.erl)
• …and 86 more

Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

AntidoteDB/antidote was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point a surveyor at a repository you know and see whether you agree with it.