Add a username field to the users table (deploy 3?)

Right now, this will always have the same value as gh_login on the users
table and login on the linked accounts table. All of these values will
get updated when we get a new gh_login value.

Eventually, we're going to have to decouple the concept of crates.io
"username" from the logins of the various account(s), which you may or
may not want to update when you update your GitHub or other login, and
which may or may not conflict with other users' crates.io usernames.

But for now, set up for that future and defer the harder decisions until
later by making this field always get the same value as gh_login.

Author: carols10cents

crates/crates_io_database/src/models/user.rs

@@ -25,6 +25,7 @@ pub struct User {
     pub account_lock_until: Option<NaiveDateTime>,
     pub is_admin: bool,
     pub publish_notifications: bool,
+    pub username: Option<String>,
 }
 
 impl User {
@@ -110,6 +111,7 @@ pub struct NewUser<'a> {
     pub gh_id: i32,
     pub gh_login: &'a str,
     pub name: Option<&'a str>,
+    pub username: Option<&'a str>,
     pub gh_avatar: Option<&'a str>,
     pub gh_access_token: &'a str,
 }
@@ -139,6 +141,7 @@ impl NewUser<'_> {
             .do_update()
             .set((
                 users::gh_login.eq(excluded(users::gh_login)),
+                users::username.eq(excluded(users::username)),
                 users::name.eq(excluded(users::name)),
                 users::gh_avatar.eq(excluded(users::gh_avatar)),
                 users::gh_access_token.eq(excluded(users::gh_access_token)),

crates/crates_io_database/src/schema.rs

@@ -870,6 +870,12 @@ diesel::table! {
         is_admin -> Bool,
         /// Whether or not the user wants to receive notifications when a package they own is published
         publish_notifications -> Bool,
+        /// The `username` column of the `users` table.
+        ///
+        /// Its SQL type is `Nullable<Varchar>`.
+        ///
+        /// (Automatically generated by Diesel.)
+        username -> Nullable<Varchar>,
     }
 }
 

crates/crates_io_database_dump/src/dump-db.toml

@@ -216,6 +216,7 @@ account_lock_reason = "private"
 account_lock_until = "private"
 is_admin = "private"
 publish_notifications = "private"
+username = "public"
 [users.column_defaults]
 gh_access_token = "''"
 

crates/crates_io_database_dump/src/snapshots/[email protected]

@@ -13,7 +13,7 @@ BEGIN ISOLATION LEVEL REPEATABLE READ, READ ONLY;
     \copy "metadata" ("total_downloads") TO 'data/metadata.csv' WITH CSV HEADER
     \copy "reserved_crate_names" ("name") TO 'data/reserved_crate_names.csv' WITH CSV HEADER
     \copy "teams" ("avatar", "github_id", "id", "login", "name", "org_id") TO 'data/teams.csv' WITH CSV HEADER
-    \copy (SELECT "gh_avatar", "gh_id", "gh_login", "id", "name" FROM "users" WHERE id in (     SELECT owner_id AS user_id FROM crate_owners WHERE NOT deleted AND owner_kind = 0     UNION     SELECT published_by as user_id FROM versions )) TO 'data/users.csv' WITH CSV HEADER
+    \copy (SELECT "gh_avatar", "gh_id", "gh_login", "id", "name", "username" FROM "users" WHERE id in (     SELECT owner_id AS user_id FROM crate_owners WHERE NOT deleted AND owner_kind = 0     UNION     SELECT published_by as user_id FROM versions )) TO 'data/users.csv' WITH CSV HEADER
 
     \copy "crates_categories" ("category_id", "crate_id") TO 'data/crates_categories.csv' WITH CSV HEADER
     \copy "crates_keywords" ("crate_id", "keyword_id") TO 'data/crates_keywords.csv' WITH CSV HEADER

crates/crates_io_database_dump/src/snapshots/[email protected]

@@ -60,7 +60,7 @@ BEGIN;
     \copy "metadata" ("total_downloads") FROM 'data/metadata.csv' WITH CSV HEADER
     \copy "reserved_crate_names" ("name") FROM 'data/reserved_crate_names.csv' WITH CSV HEADER
     \copy "teams" ("avatar", "github_id", "id", "login", "name", "org_id") FROM 'data/teams.csv' WITH CSV HEADER
-    \copy "users" ("gh_avatar", "gh_id", "gh_login", "id", "name") FROM 'data/users.csv' WITH CSV HEADER
+    \copy "users" ("gh_avatar", "gh_id", "gh_login", "id", "name", "username") FROM 'data/users.csv' WITH CSV HEADER
     \copy "crates_categories" ("category_id", "crate_id") FROM 'data/crates_categories.csv' WITH CSV HEADER
     \copy "crates_keywords" ("crate_id", "keyword_id") FROM 'data/crates_keywords.csv' WITH CSV HEADER
     \copy "crate_owners" ("crate_id", "created_at", "created_by", "owner_id", "owner_kind") FROM 'data/crate_owners.csv' WITH CSV HEADER

migrations/2025-02-19-013433_add_username_to_user/down.sql

@@ -0,0 +1,4 @@
+DROP INDEX IF EXISTS lower_username;
+
+ALTER TABLE users
+DROP COLUMN username;

migrations/2025-02-19-013433_add_username_to_user/up.sql

@@ -0,0 +1,6 @@
+ALTER TABLE users
+-- The column needs to be nullable for this migration to be fast; can be changed to non-nullable
+-- after backfill of all records.
+ADD COLUMN username VARCHAR;
+
+CREATE INDEX lower_username ON users (lower(username));

src/controllers/crate_owner_invitation.rs

@@ -56,7 +56,7 @@ pub async fn list_crate_owner_invitations_for_user(
                     .iter()
                     .find(|u| u.id == private.inviter_id)
                     .ok_or_else(|| internal(format!("missing user {}", private.inviter_id)))?
-                    .login
+                    .username
                     .clone(),
                 invitee_id: private.invitee_id,
                 inviter_id: private.inviter_id,

src/controllers/session.rs

@@ -146,6 +146,7 @@ pub async fn save_user_to_database(
     let new_user = NewUser::builder()
         .gh_id(user.id)
         .gh_login(&user.login)
+        .username(&user.login)
         .maybe_name(user.name.as_deref())
         .maybe_gh_avatar(user.avatar_url.as_deref())
         .gh_access_token(access_token)

src/index.rs

@@ -153,6 +153,7 @@ mod tests {
         let user_id = diesel::insert_into(users::table)
             .values((
                 users::name.eq("user1"),
+                users::username.eq("user1"),
                 users::gh_login.eq("user1"),
                 users::gh_id.eq(42),
                 users::gh_access_token.eq("some random token"),