Refuses an email that already exists in any letter case (EmailTakenError): the @unique index on PostgreSQL is case-sensitive, so a legacy Bob@x.test would not stop a new bob@x.test. A concurrent insert of the same canonical email loses on the index (P2002) and gets the same error.

Parameters

data

email

string

passwordHash

string

Returns

Promise<AuthUser>

Implementation of

UserSource.create

findByEmail()

> findByEmail(email): Promise<AuthUser | null>

Defined in: auth-prisma/src/index.ts:324

Emails are case-insensitive identities: new rows are stored canonical (trimmed, lowercased). Rows written before that may be mixed-case, so on PostgreSQL the lookup is case-insensitive and refuses ambiguity: when two rows differ only in letter case it throws AccountEmailAmbiguousError instead of picking one (run normalizeAuthUserEmails to find and fix them). On MySQL the column collation is already case-insensitive and the unique index rules out case-variant duplicates, so the exact lookup is authoritative.

Parameters

email

string

Returns

Promise<AuthUser | null>

Implementation of

UserSource.findByEmail

findById()

> findById(id): Promise<AuthUser | null>

Defined in: auth-prisma/src/index.ts:352

Parameters

id

string

Returns

Promise<AuthUser | null>

Implementation of

UserSource.findById

findByIds()

> findByIds(ids): Promise<PublicUser[]>

Defined in: auth-prisma/src/index.ts:363

One WHERE id IN (…) per chunk instead of one query per id. select lists only the non-credential columns, so the password hash is never even read; ids with no row are simply absent from the result, which stays in the order the caller asked for.