Signed-off-by: DL6ER <dl6er@dl6er.de>

…y (e.g., broken pipe). Also use TCP protocol for upstream server name resolution

Signed-off-by: DL6ER <dl6er@dl6er.de>

Signed-off-by: DL6ER <dl6er@dl6er.de>

…ugh the environment

Signed-off-by: DL6ER <dl6er@dl6er.de>

Signed-off-by: DL6ER <dl6er@dl6er.de>

Signed-off-by: DL6ER <dl6er@dl6er.de>

Signed-off-by: DL6ER <dl6er@dl6er.de>

Signed-off-by: DL6ER <dl6er@dl6er.de>

Signed-off-by: DL6ER <dl6er@dl6er.de>

Add Clang compiler support and update project to C17 standard

Signed-off-by: DL6ER <dl6er@dl6er.de>

Reset config values after env forcing

Signed-off-by: DL6ER <dl6er@dl6er.de>

Signed-off-by: DL6ER <dl6er@dl6er.de>

Improve internal DNS resolver

Read config files from new location

Signed-off-by: DL6ER <dl6er@dl6er.de>

…ave been made in between and add a test ensuring they remain in sync in the future

Signed-off-by: DL6ER <dl6er@dl6er.de>

Signed-off-by: DL6ER <dl6er@dl6er.de>

… this in the log files to not needlessly change for format. Open for discussion

Signed-off-by: DL6ER <dl6er@dl6er.de>

Signed-off-by: DL6ER <dl6er@dl6er.de>

Update embedded CivetWeb server

Improve incomplete installations support

… 0.0.0.0/::, and known Umbrella IP blocking pages) accompanied with suitable CI testing. This change improves handling of externally blocked queries in two ways:

1. Once we know the domain is externally blocked, we don't forward it again for the same client (it is currently forwarded each time)
2. We can recognize cache content with the specified addresses/flags as being upstream blocked and don't return them as "OK (cached)" through the API/in the database

Signed-off-by: DL6ER <dl6er@dl6er.de>

…ress, the entire query should be blocked. This is already the case but this commit adds an explicit CI test for this desired behavior.

Signed-off-by: DL6ER <dl6er@dl6er.de>

…or configuring how long relpies from upstream are cached

Signed-off-by: DL6ER <dl6er@dl6er.de>

Do not refuse to (re)start if there is another pihole-FTL process already running *inside* a docker container on the same host. Currently, (re)starting on the host is prevented when FTL detects during startup that another process is running on the machine (the one inside the container) even when they can obviously run alongside perfectly fine. This is caused by the host being able to see *all* processes, not only the one in its own cgroup.
This actually hinders development/testing natively on the host. This commit fixes this by not considering pihole-FTL processes running in downstream (!) docker containers as duplicate instances.

Signed-off-by: DL6ER <dl6er@dl6er.de>

…g = false). This will happen only once and may greatly ease debugging in case unintended side-effects (most often it will by typos in the old setupVars.conf)

Signed-off-by: DL6ER <dl6er@dl6er.de>

Allow other FTL process inside docker

…ig type to a non-allocated type after free'ing.

Signed-off-by: DL6ER <dl6er@dl6er.de>

Signed-off-by: DL6ER <dl6er@dl6er.de>

Signed-off-by: DL6ER <dl6er@dl6er.de>

Signed-off-by: DL6ER <dl6er@dl6er.de>

Co-authored-by: Adam Warner <me@adamwarner.co.uk>
Signed-off-by: Dominik <DL6ER@users.noreply.github.com>

Fix possible double-free and avoid memory leaking

Implement DNS caching for queries blocked upstream

Signed-off-by: DL6ER <dl6er@dl6er.de>

Signed-off-by: DL6ER <dl6er@dl6er.de>

Use ftl-build:v2.7 builder for Actions

Restart FTL on change of webserver.api.cli_pw

… the last gravity update via the API

Signed-off-by: DL6ER <dl6er@dl6er.de>

…umber of available (= online and assigned) cores when computing the CPU utilization of the system

Signed-off-by: DL6ER <dl6er@dl6er.de>

Add GET /api/stats/summary -> .gravity.last_update

Fix CPU % calculation

…PR #3124 removed the redirection feature from the login page

Signed-off-by: DL6ER <dl6er@dl6er.de>

Remove appending ?target=... when redirecting to login page

Make v5 -> v6 migration verbose