fix: diff to show the default rather than "unknown"

This (including the rest of the commit message) was AI-generated.

In `_compute_diff`, the expected support was derived via a raw dict lookup
on `expected_all` with "unknown" as the fallback.  Features not explicitly
listed in the server profile (e.g. most features absent from the xandikos
compatibility hints) were therefore shown as expected="unknown" in the diff
output, instead of their actual default ("full" for standard server-features).

Fix by replacing the raw lookup with `self.expected_features.is_supported()`
which traverses the feature hierarchy and applies FeatureSet defaults.

prompt: why does `caldav-server-tester --name xnadikos --diff` report the expectations as "unknown" rather than default "full"?  search.recurrences.expanded.exception is now reported to be supported by xandikos, while it's configured in ~/caldav/caldav/compatibility_hints.py to be unsupported.  Please check if this is due to code changes on our side (either here or in the caldav library) or xandikos side. Also, why is this visible when doing `cd ~/caldav ; pytest -k 'xandikos and compat' --pdb`, but `pytest -k 'xandikos` passes without problems?

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: tobixen

CHANGELOG.md

@@ -6,6 +6,11 @@ This file should adhere to [Keep a Changelog](https://keepachangelog.com/en/1.1.
 
 This project should adhere to [Semantic Versioning](https://semver.org/spec/v2.0.0.html), though some earlier releases may be incompatible with the SemVer standard.
 
+## [Unreleased]
+
+### Fixed
+- `--diff` reported expected support as `unknown` for features not explicitly listed in the server profile. `_compute_diff` now uses `is_supported()` to resolve defaults (typically `full`) instead of falling back to a hard-coded `"unknown"`.
+
 ## [1.1] - 2026-04-24
 
 This release works with caldav 3.2 and higher.

src/caldav_server_tester/checker.py

@@ -172,7 +172,7 @@ def _compute_diff(self) -> dict:
         all_keys = set(observed) | set(expected_all)
         for key in all_keys:
             obs_support = observed.get(key, {}).get("support", "unknown")
-            exp_support = expected_all.get(key, {}).get("support", "unknown")
+            exp_support = self.expected_features.is_supported(key, str)
             if obs_support != exp_support:
                 diff[key] = {"expected": exp_support, "observed": obs_support}
         return diff