618 commits

This branch

This branch

All branches

Author	SHA1	Message	Date
ppom	285954f7cd	Remove outdated FIXME	2026-02-18 12:00:00 +01:00
ppom	dc51d7d432	Add support for macOS	2026-02-17 12:00:00 +01:00
ppom	488dc6c66f	Update release instructions	2026-02-15 12:00:00 +01:00
ppom	88c99fff0f	Fix install instructions	2026-02-12 12:00:00 +01:00
ppom	645d72ac1e	.gitignore cleanup	2026-02-12 12:00:00 +01:00
ppom	a7e958f248	Update ARCHITECTURE.md	2026-02-12 12:00:00 +01:00
ppom	5577d4f46f	reaction-plugin: Add metadata v2.3.0	2026-02-12 12:00:00 +01:00
ppom	a8cd1af78d	Set CapabiltyBoundingSet again	2026-02-12 12:00:00 +01:00
ppom	2f57f73ac9	Fix systemd functionality ... - Non-absolute WorkingDirectory was refused by systemd - Plugin specific-conf updated Improvements: - ReadOnlyPaths=/ - ProtectHome=true in release builds - SystemCallFilter further restricted Disabled: - DynamicUser: breaks stdio communication, FIXME! - RestrictAddressFamilies: seems impossible to override to default. - CapabilityBoundingSet: too restrictive	2026-02-12 12:00:00 +01:00
ppom	d629d57a7e	Change ipset version option from 4/6/46 to ipv4/ipv6/ip	2026-02-12 12:00:00 +01:00
ppom	3c20d8f008	Fix merging of systemd options	2026-02-12 12:00:00 +01:00
ppom	5a030ffb7e	Make systemd default options more accessible for users by moving them up	2026-02-12 12:00:00 +01:00
ppom	a4ea173c13	Do not permit options key when stream/action is not a plugin	2026-02-12 12:00:00 +01:00
ppom	3a61db9e6f	plugin: shutdown: add function that permit graceful shutdown by signal ... Handling SIGTERM (etc) signals permit graceful shutdown, cleaning of resources etc. Added in ipset and cluster.	2026-02-12 12:00:00 +01:00
ppom	b4313699df	systemd: Let reaction stop its subprocesses before killing them ... systemd by default send SIGTERM to all processes in the cgroup, which doesn't let reaction handle the shutdown of its plugins. This is fixed by adding KillMode=mixed.	2026-02-12 12:00:00 +01:00
ppom	270c6cb969	systemd service: config file must live in /etc/reaction/ ... This is a breaking change, but it unifies config for yaml, json, jsonnet and directory users.	2026-02-12 12:00:00 +01:00
ppom	15f923ef64	Safeguard against users executing plugins themselves ... main_loop now first checks that it has been started with the `serve` argument. If not, it prints an info message and quits.	2026-02-11 12:00:00 +01:00
ppom	a37a5e5752	release v2.3.0 ... - cross-rs project doesn't compile anymore: switching to debian12-amd64 only binary release - package virtual plugin in reaction .deb - package ipset plugin in separate .deb with its required libipset-dev dependency	2026-02-11 12:00:00 +01:00
ppom	a8651bf2e0	Removal of nft46 and ip46tables	2026-02-11 12:00:00 +01:00
ppom	b07b5064e9	Improve reaction-plugin developer documentation	2026-02-11 12:00:00 +01:00
ppom	b7d997ca5e	Slight change on the "no audit" sentence	2026-02-09 12:00:00 +01:00
ppom	cce850fc71	Add recommandation on ipset or nftables rather than plain iptables	2026-02-09 12:00:00 +01:00
ppom	109fb6d869	Adapt reaction core to plugin interface change	2026-02-09 12:00:00 +01:00
ppom	ae28cfbb31	cluster: adapt to plugin interface change	2026-02-09 12:00:00 +01:00
ppom	b0dc3c56ad	ipset: adapt to plugin interface change	2026-02-09 12:00:00 +01:00
ppom	57d6da5377	virtual: adapt to plugin interface change	2026-02-09 12:00:00 +01:00
ppom	12fc90535a	Change plugin interface: oneshot load_config and start ... Instead of multiple stream_impl / action_impl and one finish_setup. This made plugin implementations awkward: they often got some conf and couldn't determine if it was valid or not. Now they get all the conf in one function and don't have to keep partial state from one call to another. This has the other important benefit that configuration loading is separated from startup. This will make plugin lifecycle management less clunky.	2026-02-09 12:00:00 +01:00
ppom	62933b55e4	Start plugins after start commands ... Because stop commands run after plugins' shutdown, so it seems better that commands embrace ({ plugins }). Fix outdated comment about aborting on startup.	2026-02-09 12:00:00 +01:00
ppom	34e2a8f294	plugin: simpler crate version retrieval	2026-02-09 12:00:00 +01:00
ppom	41bc3525f8	Fix time-based test sometimes failing by increasing sleep	2026-02-09 12:00:00 +01:00
ppom	5ce773c8e5	cluster: ignore integration tests for now	2026-02-09 12:00:00 +01:00
ppom	6914f19fb8	fix assert_cmd::cargo_bin deprecation warning	2026-02-09 12:00:00 +01:00
ppom	7cd4a4305d	fix: merge plugins in configuration	2026-02-09 12:00:00 +01:00
ppom	c39fdecef3	ipset: add tests for configuration	2026-02-09 12:00:00 +01:00
ppom	885e6b7ef7	ipset: re-arrange spacing in logs	2026-02-09 12:00:00 +01:00
ppom	516e6956ab	fix double-printing of square brackets in plugin logs	2026-02-09 12:00:00 +01:00
ppom	79ec6d279f	ipset: Manuel e2e test does pass	2026-02-09 12:00:00 +01:00
ppom	a83c93ac9d	ipset: do not shutdown plugin when one action errors	2026-02-09 12:00:00 +01:00
ppom	47947d18db	ipset: Fix dumb bug due to future not awaited ... The edge case is so dumb, cargo is supposed to tell me about this >< Just learnt that Python never warns about this btw: https://trio.readthedocs.io/en/v0.9.0/tutorial.html#warning-don-t-forget-that-await	2026-02-09 12:00:00 +01:00
ppom	915e308015	Better plugin process management ... following stderr: task doesn't use shutdown anymore. It will simply follow stderr until the end of reaction, which at worst is a negligible memory leak if reaction continues running. I tried closing stderr on the plugin side with a raw syscall of the file descriptor, but reaction side doesn't see that stderr is closed. So I can't rely on that. Quitting when shutdown.wait() returns is too early, because that's also what makes reaction asking for the plugin to close(), and it can print important logs during its shutdown. The task ignoring all the shutdown part is dead simple and is most likely correct everytime. updated the wording of plugin-related errors. also replaced futures::select! { future, sleep() } with more concise and macro-less tokio::timeout.	2026-02-09 12:00:00 +01:00
ppom	41b8a661d2	Print on stderr instead of stdout ... ...stdout is already taken by remoc ;)	2026-02-09 12:00:00 +01:00
ppom	87a25cf04c	Extract ipset options from action options so that it's globally merged ... Actions don't manage sets anymore. Set options are merged at each new action, then Sets are managed by themselves.	2026-02-09 12:00:00 +01:00
ppom	d6b6e9096b	ipset: Add the add/del option, journal orders & deduplicate them	2026-02-09 12:00:00 +01:00
ppom	3ccd471b45	ipset: so much ~~waow~~ code	2026-02-09 12:00:00 +01:00
ppom	3a6260fa26	reaction-plugin-ipset: first work session	2026-02-09 12:00:00 +01:00
kol3rby	959c32c01e	Fix project not compiling on BSD & Solaris systems	2026-02-09 11:03:00 +01:00
ppom	05c6c1fbce	Fix tests ... I initially wrote those tests with a test secret key file in the same directory. Better having them write their own secret key file in their own dir than a dangling test file in source code and be sensitive to the directory tests are run in.	2026-01-19 12:00:00 +01:00
ppom	615d721c9a	cluster: Upgrade iroh to 0.95.1	2026-01-19 12:00:00 +01:00
ppom	19ee5688a7	Testing with clusters of up to 15 nodes. Fails at ~6 to 9 nodes. ... Still a "connection lost" issue. Happens irregularly. Nodes tend to ignore incoming connections because their id is too small. I should debug why it is the case. Nodes may succeed to recreate connections, but they should not lose connections on localhost like that...	2026-01-19 12:00:00 +01:00
ppom	fb6f54d84f	Disable test where one plugin is in multiple nodes of one cluster. Test pass!	2026-01-19 12:00:00 +01:00