# `Mailglass.Identifier`
[🔗](https://github.com/szTheory/mailglass/blob/v2.6.0/lib/mailglass/identifier.ex#L1)

Validates Postgres unquoted identifiers — the single source of truth for
schema, prefix, and table-name checks across mailglass.

A valid identifier is a letter or underscore followed by any combination of
letters, digits, and underscores, and is at most 63 bytes long (Postgres'
`NAMEDATALEN` limit — longer names are silently truncated by Postgres, which
can alias two distinct schemas onto one, so we reject them outright).

This grammar rejects anything that could be an injection vector — quotes,
semicolons, whitespace, dashes. Callers with a legitimate need for a quoted
identifier (mixed case, dashes) should pre-quote and adjust; mailglass does
not surface such options.

Validation failures raise `Mailglass.ConfigError` with `type: :invalid`. The
specific cause lives in `context.key` and `context.reason` — match on the
struct, never on the message string.

# `validate!`
*since 2.0.0* 

```elixir
@spec validate!(String.t(), atom()) :: String.t()
```

Validates `value` as a Postgres unquoted identifier, returning it unchanged.

`key` is the configuration key name surfaced in the raised error's context
(e.g. `:schema`, `:prefix`).

Returns the validated string so it can be used pipe-first. Raises
`Mailglass.ConfigError` with `type: :invalid` when `value` is not a binary,
exceeds 63 bytes, or does not match the unquoted-identifier grammar.

---

*Consult [api-reference.md](api-reference.md) for complete listing*