Fix uid: search to match by prefix instead of substring

[harriiinnii]

Summary

khard list displays a short unique UID prefix for each contact (e.g. a for a contact whose UID starts with a and is the only one with that prefix). However, uid:a matched by substring, so it would match both aaabbbb and bbbaaaa instead of only aaabbbb. This made the displayed prefix unusable as a selector.

Problem

parse("uid:xxx") returned a plain FieldQuery("uid", "xxx"), which inherits TermQuery.match and does term in value.lower() — a substring check. The displayed prefix is a prefix, not a substring, so the two behaviours are inconsistent.

Solution

Add a UidQuery subclass of FieldQuery that overrides _match_union to use str.startswith for string values. Update parse() to return a UidQuery when the field is "uid".

The change is intentionally minimal: only the UID field is affected, and the rest of the query machinery (AndQuery, OrQuery, equality, hashing, __str__) works as before.

Testing

• Added TestUidQuery in test/test_query.py covering: prefix match, substring-only non-match, exact match, case insensitivity, empty term with and without uid set, and disambiguation of two contacts sharing a common prefix.
• Added test_uid_field_creates_uid_query_instance to TestParser verifying that parse("uid:xxx") returns a UidQuery.
• All existing tests are preserved unchanged.

Fixes #327

[lucc]

Looks good, I just have some nitpicks.

[lucc]

I activated the ci pipeline so you can see that your changes break some other tests. Please run all test locally to speed up the feedback cycle.

The breakage is not severe, its just that some uid queries are being compared and their type now obviously changed. Please update the breaking tests as well.