- CI moves to .gitea/workflows/ci.yml and calls the shared driver workflow
of NAPALM/driver-standard; .github/ is gone.
- license = "Apache-2.0" and a LICENSE file: the package declared no
license at all. README and CHANGELOG.md, py.typed, links to
git.netork.io, a dev extra with floors; .venv/ and build output ignored.
- Floors on every dependency: napalm-device-types>=4.21.1 (had none),
napalm>=4.0 (what napalm-device-types needs) and requests>=2.28.0.
Without them the lowest CI job resolves napalm 0.10, which does not
build; the tests pass at exactly these floors.
- mypy strict for the package; napalm_yealink.yealink is exempt for now
(20 errors), recorded in the quality baseline with get_facts'
complexity. Coverage floor 42 % (measured 42.61 %).
Closes#4
- The MD5 of the password is what the phone's login API expects, the wire
format and no protection of ours (the session runs over HTTPS):
hashlib.md5(..., usedforsecurity=False) (B324). The digest is the same.
- A failed logout in close() and an unreadable system info in get_facts()
now log at debug level (B110), so a swallowed failure leaves a trace; the
module gets its logger.
- The comment on the "yealink" fingerprint rule moves above it (ruff
format had split the line around it) and is in English.
Mechanical: ruff format, and ruff check --fix with the rule set every NAPALM
driver uses (E, F, I, UP, B at line length 100, Python 3.10). What ruff
cannot fix comes in the next commit.
get_health_metrics, get_ssids, get_radio_info and get_wireless_clients
only raised NotImplementedError. They are left over from when the
driver sat on AccessPointDriver, and a stub makes a capability check
(callable(getattr(driver, name))) answer yes. The application polls
SNMP itself (napalm-device-types docs/STRUCTURE.md); a phone has no
profile to declare.
Also stops tracking the .pyc and egg-info files .gitignore already
lists.
Closes#2
The same job as napalm-fritzbox and napalm-opnsense: Python 3.10, 3.11 and
3.12, pytest, then wheel and sdist. napalm-device-types comes from
git.netork.io first, because PyPI has an unrelated package of that name.
pyproject gains a dev extra with pytest, which the job installs.
Closes netork#117.
A Yealink is a desk phone. It inherited AccessPointDriver because nothing better
existed, and netOrk reads that role to populate its wireless page, its AP
profile pickers and its SSID drift view — so a phone appeared in all three.
napalm-device-types now has PhoneDriver. TYPE_LABEL comes from it, so the local
override is gone.
The two failing fingerprint tests were asserting the opposite of a deliberate
decision. HTTP_FINGERPRINT carries "yealink" as a heavy but *not* mandatory
rule, with the reason written next to it: some models answer / with a redirect
to their API page, whose body says "yealink" nowhere, and a mandatory rule
disqualifies the driver whenever its pattern is absent. Making it mandatory
would rule out a real phone.
The tests demanded it be mandatory, and the Cisco test then asserted `not
all(...)` over an empty list, which is False. Both now check the property that
actually protects against a false match: none of this driver's patterns appear
on a Cisco page, so it contributes no score and never wins.
OUI-Treffer (80:5E:0C, 80:5E:C0) sind der zuverlässigere Identifier.
mandatory=True disqualifizierte Geräte bei denen nur der JS-Redirect abgeholt
wurde statt der vollständigen Login-Seite.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Nur IEEE-verifizierte und an echten Geraeten bestaetigen OUIs behalten.
00:1A:8C gehoert AzureWave Technology (OpenWrt-WLAN-Chips), nicht Yealink.
24:9A:D8, 70:32:17, 34:6B:D3 entfernt bis Bestaetigung vorliegt.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Christian Manivong