Files
mtusi2026lab2/db/migrations/002_utf8_binary_strings.sql
T

18 lines
1.2 KiB
SQL

-- PostgreSQL text/varchar reject the NUL character, but the public API accepts
-- arbitrary valid UTF-8 code points in display_name, title and message text.
-- Store exact UTF-8 bytes in bytea while enforcing code-point limits at API.
-- Existing rows are converted losslessly; safe with persistent lab2 volumes.
ALTER TABLE users DROP CONSTRAINT users_display_name_check;
ALTER TABLE chats DROP CONSTRAINT chats_title_check;
ALTER TABLE messages DROP CONSTRAINT messages_text_check;
ALTER TABLE users ALTER COLUMN display_name TYPE bytea USING convert_to(display_name, 'UTF8');
ALTER TABLE chats ALTER COLUMN title TYPE bytea USING convert_to(title, 'UTF8');
ALTER TABLE messages ALTER COLUMN text TYPE bytea USING convert_to(text, 'UTF8');
-- A valid UTF-8 code point occupies at most four bytes.
-- Pydantic enforces 1..100/4000 code points and no whitespace-only strings.
ALTER TABLE users ADD CONSTRAINT users_display_name_octets CHECK (octet_length(display_name) BETWEEN 1 AND 400);
ALTER TABLE chats ADD CONSTRAINT chats_title_octets CHECK (octet_length(title) BETWEEN 1 AND 400);
ALTER TABLE messages ADD CONSTRAINT messages_text_octets CHECK (octet_length(text) BETWEEN 1 AND 16000);