Module route_naming 

Canonical REST route-segment derivation for a model name.

cratestack#345: three independently-implemented algorithms (the server’s real Axum route registration in cratestack-macros::axum::model::routes, and near-identical split_words-based reimplementations in cratestack-client-typescript::naming and cratestack-client-dart::idents) all computed what is supposed to be the same wire contract — a model’s REST route path — and drifted apart for model names containing a literal _ (legal today; no parser grammar restriction). to_snake_case/pluralize here are exactly the server’s original algorithm, byte-for-byte, moved into this crate (already a shared dependency of cratestack-macros, cratestack-client-typescript, and cratestack-client-dart) so every consumer imports the same implementation instead of reimplementing it.

Do not reimplement these. Any code deriving a model’s REST route must call model_route_segment (or to_snake_case + pluralize directly, in that order) from here.

Functions

model_route_segment

The canonical REST route segment (no leading /) for a model name — pluralize(&to_snake_case(model_name)). The server’s Axum route registration and every client generator’s route derivation must produce this exact string for a given model name; call this instead of composing to_snake_case/pluralize locally so the composition order can’t drift either.

pluralize

English pluralization used for both REST route segments and generated table/column names — the single implementation cratestack-core and cratestack-migrate both call (cratestack#504; previously two hand-synced copies, cratestack-migrate::naming::pluralize mirroring this one, that had already drifted apart on this exact rule).

to_snake_case

Converts value to snake_case by inserting a single _ before every non-initial uppercase character. Every other character — including any pre-existing _ — passes through unchanged.