The KQL parser produced its own validation errors but imported them from the
search service's query package. Move them into pkg/kql and let the search
backend consume kql.IsValidationError, so the parser stops depending on a
service package.
both engines now agree on names, titles, tags, paths, types, sizes, dates,
hidden flags, facet values and wildcards. quotes only delimit phrases and
the equals operator matches the whole field value, following the kql spec.
the index name carries a generation so a changed mapping starts on a fresh
index, MIGRATION.md says how to fill it.
Four cases covering the distinct grammar paths touched by the
new Key rule: dotted key in a TextPropertyRestrictionNode,
multi-level key (stressing the Kleene part), dotted key in a
GroupNode (the k:Key? branch), and a dot on the value side
that must stay a literal character.
Introduce a dedicated Key rule (Char+ ("." Char+)*) so property
restriction keys can contain dots. Queries like
audio.artist:Motörhead or photo.cameraMake:Apple now parse.
Goal: keep the data structure aligned with the query namespace.
A driveItem exposes nested fields as dotted paths in its API
response (photo.cameraMake, location.latitude); users should be
able to query with the same path, without admin-configured slot
mappings or alias tables like Microsoft requires.
Both bleve and OpenSearch already treat dots as the nesting
separator for field paths and pass unknown keys through
unchanged — only the KQL grammar blocked dotted keys via
Char <- [A-Za-z]. The new Key rule lifts that restriction; no
compiler changes or field-alias maintenance is needed.
Values are unaffected — dots in values still parse as literal
characters.
Dominik Schmidt