[Nexthop][fboss2-dev] Add fboss2-dev config hostname CLI

[vybhav-nexthop]

Pre-submission checklist

• I've ran the linters locally and fixed lint errors related to the files I modified in this PR. You can install the linters by running pip install -r requirements-dev.txt && pre-commit install
• pre-commit run

Summary

Adds fboss2-dev config switch hostname <name>, which sets the optional
sw.hostname field in the agent switch config.

The command is nested under the config switch command group (introduced by
the admin-distance command) rather than at the top level, grouping
switch-level settings that do not belong to a specific network object.

• Action level: HITLESS — applied via
  ThriftConfigApplier::updateSwitchSettings (newSwitchSettings->setHostname);
  no SAI ChangeProhibited guard.
• Upsert: if hostname is already set to the requested value, the
  command is a no-op.
• Validation: the hostname argument is validated against RFC 952/1123
  using RE2 (alphanumeric plus hyphens, max 63 chars per label, total max
  253 chars).
• Field: switch_config.thrift:SwitchConfig.hostname (optional string) —
  "Overrides the system hostname, useful in ICMP responses." When unset, the
  agent falls back to the OS hostname at config-apply time
  (getLocalHostname()), so the value is always populated in switch state.

Test Plan

Unit tests (CmdConfigHostname{,Absent}TestFixture)

9 tests, all passing in CI: argument validation (valid, FQDN, empty args,
empty string, too-many args, invalid format), set-new hostname, already-set
no-op, and set-when-field-absent.

Integration test (ConfigHostnameTest) — on DUT gold407

The integration test reads the current hostname from switch state
(getSwitchSettingsField, always populated) inside a try/catch that returns
"" if absent, sets a test value, verifies it via the switch-state thrift
path, and restores the original hostname in TearDown. It does not read the
optional config field directly, so it does not need any getSwConfigField
default-value helper — Fboss2IntegrationTest.h is unchanged from upstream.

Results from the internal CLI Integration Test (DUT) run on 2026-07-03
(device gold407):

[----------] 2 tests from ConfigHostnameTest
[ RUN      ] ConfigHostnameTest.SetAndVerifyHostname
  Current hostname: 'gold407'
  [Step 2] Setting hostname to 'cli-e2e-test-hostname'...
  Running CLI command: config switch hostname cli-e2e-test-hostname
  Verified: hostname is 'cli-e2e-test-hostname'
  Running CLI command: config switch hostname gold407      (restore)
[       OK ] ConfigHostnameTest.SetAndVerifyHostname (173 ms)
[ RUN      ] ConfigHostnameTest.IdempotentSet
  Running CLI command: config switch hostname cli-e2e-idempotent-hostname
  Running CLI command: config switch hostname cli-e2e-idempotent-hostname
  Running CLI command: config switch hostname gold407      (restore)
[       OK ] ConfigHostnameTest.IdempotentSet (120 ms)
[----------] 2 tests from ConfigHostnameTest (294 ms total)

Ran as part of the 25-test fboss2 CLI integration suite; all 25 passed
(total 50.5 s).

Review Findings

Pre-publication review (single-pass automated reviewer). One item raised —
unchecked *config.sw() dereference in CmdConfigHostname.cpp — reviewed
and dismissed: it is the established pattern across all fboss2 config
commands (e.g. CmdConfigArp.cpp, CmdConfigCopp.cpp); sw is reliably
present for a loaded agent config. No issues above threshold.

Presubmission Checks

• clang-format (--style=file --dry-run --Werror): pass on all added/modified C++ sources.
• pre-commit: pass (no violations on changed files).

[joseph5wu]

Consider to use CmdConfigSwitch to manage individual Switch Settings

[meta-codesync]

@joseph5wu has imported this pull request. If you are a Meta employee, you can view this in D110790091.

[meta-codesync]

@joseph5wu merged this pull request in 3aed671.