[somethingwithproof]

Description

Stacked on #422 (base fix/review-hardening). Merge #422 first; this PR must merge after it and should be retargeted to develop once #422 lands.

Hostname resolution now handles IPv6, and the CI test data now contains an alert rule that fires.

• syslog_resolve_incoming_hosts() used IPv4-only gethostbyname(), so a sender known only by an AAAA record, or by an IPv6 address, was marked unresolved-. Names are now checked with dns_get_record() for A and AAAA, falling back to gethostbyname(). IPv4 and IPv6 literals need no lookup and do not use the lookup budget from Harden alert commands, upgrades and dashboards, and fix review findings #422. IPv6 text is normalised with inet_pton()/inet_ntop() (compressed, lower case; brackets removed, zone suffix kept) before it is compared with Cacti devices and stored. Harden alert commands, upgrades and dashboards, and fix review findings #422's lookup budget, its "left as received" handling when the budget is spent, and syslog_no_dns are unchanged. The function gains an optional last argument, $dns_records, so tests can supply DNS answers.
• tests/bin/populate_syslog_incoming.sh inserted its alert rule without suppression_schedule (text NOT NULL, no default), and with mismatched column and value counts, so the insert failed under strict mode and CI never had a rule to evaluate. The insert is fixed, the script uses set -euo pipefail, and the removal rule uses REPLACE INTO.
• The "force Syslog Plugin Poller" step now asserts that the alert fired (log line and a syslog_logs row for host bob). "Check Cacti Logs for Errors" still fails on any FATAL:, ERROR: or WARNING: line except that one alert line, matched literally.
• HostnameLookupBudgetTest used the literal 192.0.2.1 as a name that fails to resolve. Literals now count as resolved, so it uses no-such-sender.invalid.

Related Issue

None.

Motivation and Context

IPv6-only senders were reported as unresolved. The CI gap meant alert evaluation was never exercised end to end.

How Has This Been Tested?

Executed, on Cacti 1.2.x and on develop, on top of #422 head f93418a:

• Pest: 589 passed on PHP 8.4 on both cores. New tests in tests/Unit/HostAddressResolutionTest.php cover IPv4 literal; IPv6 compressed, expanded, mixed case, bracketed, mapped and zoned; A only, AAAA only, both, none and a failed query; hosts-file fallback; syslog_no_dns on; and the budget: AAAA-only names resolve within the budget, are left as received with one budget log line when it is exhausted, literals skip the budget, and no lookup is made when syslog_no_dns is on, with either budget.
• php tests/bin/patch-coverage.php coverage/clover.xml f93418a 100: 100.00% (31/31 lines).
• php tests/bin/check-i18n-pot.php f93418a: OK.
• CI-order replay against MariaDB on both cores: the five database regression scripts and cli_schema_upgrade_test.php pass (including review_fixes_paths_database_test.php, 124 checks), populate and processor give one alert-fired line and one syslog_logs row for bob, live_pipeline_test.php passes (39 checks) as www-data, and nothing remains after the log-check filter.

Not run: a real AAAA lookup against a live resolver (stubbed in tests); the GitHub workflow itself; the optional Xdebug dump and coverage merge from #422 (no Xdebug in the replay containers).

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)

Checklist:

• My code follows the code style of this project.
• My change requires a change to the documentation.
• I have updated the documentation repository accordingly.

README gains a short note on IPv6 senders.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud