DISCONTINUED: Look at https://github.com/projektwahl/projektwahl-lit

sveltejs/svelte#2118

set default_transaction_read_only = false; CREATE DATABASE projektwahl;

INSERT INTO projects (title, info, place, costs, min_age, max_age, min_participants, max_participants, presentation_type, requirements, random_assignments) VALUES ('', '', '', 0, 0, 0, 0, 0, '', '', NULL);

SELECT column_default FROM information_schema.columns WHERE table_name = 'projects' AND column_name = 'random_assignments';

CREATE INDEX users_type_index ON users (type); CREATE INDEX users_project_leader_id_index ON users (project_leader_id);

nix develop
# TODO FIXME check that this is set - if we use it.
// TODO FIXME store password instead or use a longer living token or use oauth so we can refresh this
PROJEKTWAHL_ADMIN_ACCESS_TOKEN=`curl --data "grant_type=password&username=admin&password=admin&client_secret=secret&client_id=admin-cli" http://localhost:8888/auth/realms/master/protocol/openid-connect/token | jq -r .access_token`

# set in .env - warning: the .env file is stupid - don't quote anything and don't use comments
THE_BASE_URL=http://localhost:3000/
DATABASE_URL=postgres://projektwahl:changeme@localhost:54321/projektwahl
OPENID_URL=http://localhost:8888/auth/realms/projektwahl
CLIENT_ID=projektwahl
CLIENT_SECRET=
OPENID_ADMIN_URL=http://localhost:8888/auth/admin/realms/projektwahl/users

npm run dev
http://localhost:3000/setup

docker-compose stop
docker-compose rm db

psql -p 54321 -h localhost -U projektwahl
echo "EXPLAIN (ANALYZE, COSTS, VERBOSE, BUFFERS, FORMAT JSON) " | psql -p 54321 -h localhost -U projektwahl > analyze.json
# https://explain.dalibo.com/
EXPLAIN ANALYZE SELECT id,name,type FROM users ORDER BY id ASC,name ASC LIMIT (10 + 1); # why sorted after name
# https://www.postgresql.org/docs/current/runtime-config-query.html

-- TODO FIXME OPTIMIZE THIS KIND OF QUERY BY UNION-ing the parts together and also only ordering the id field
-- ALSO TEST IF THIS IS ACTUALLY FASTER
EXPLAIN ANALYZE SELECT id,name,type FROM users WHERE (('user0.94651659117602724' < name) OR ('user0.94651659117602724' = name AND 'voter' < type) OR ('user0.94651659117602724' = name AND 'voter' = type AND '0655c7e4-cc6a-4013-a0a5-d18b7ff48e44' < id) OR (NOT true AND NOT false)) AND name LIKE '%%' AND (true OR id = null) ORDER BY name ASC,type ASC,id ASC LIMIT (10 + 1);

EXPLAIN ANALYZE SELECT id,name,type FROM users WHERE ('user0.94651659117602724' < name) OR ('user0.94651659117602724' = name AND 'voter' < type) OR ('user0.94651659117602724' = name AND 'voter' = type AND '0655c7e4-cc6a-4013-a0a5-d18b7ff48e44' < id) ORDER BY name ASC,type ASC,id ASC LIMIT (10 + 1);

TODO

TODO FIXME THE ORDER OF THE UNIONS HERE IS WRONG AND MAY ALSO NEED TO BE ADJUSTED FOR BACKWARDS PAGINATION

EXPLAIN ANALYZE
(SELECT id,name,type FROM users WHERE ('user0.94651659117602724' < name) ORDER BY name ASC,type ASC,id ASC LIMIT (10 + 1))
UNION ALL
(SELECT id,name,type FROM users WHERE ('user0.94651659117602724' = name AND 'voter' < type) ORDER BY name ASC,type ASC,id ASC LIMIT (10 + 1))
UNION ALL
(SELECT id,name,type FROM users WHERE ('user0.94651659117602724' = name AND 'voter' = type AND '0655c7e4-cc6a-4013-a0a5-d18b7ff48e44' < id) ORDER BY name ASC,type ASC,id ASC LIMIT (10 + 1))
LIMIT 11;



EXPLAIN ANALYZE SELECT id,name,type FROM users WHERE ROW('user0.94651659117602724', 'voter', '0655c7e4-cc6a-4013-a0a5-d18b7ff48e44') < ROW(name, type, id) ORDER BY name ASC,type ASC,id ASC LIMIT (10 + 1);


OTHER ONE:

SELECT id,name,type FROM users WHERE ('user0.0033005226067721605' < name) OR ('user0.0033005226067721605' = name AND 'c7821916-0e40-4459-8746-2ad6bde37700' > id) ORDER BY name ASC,id DESC LIMIT (10 + 1);



SELECT id,name,type FROM users WHERE ('user0.94651659117602724' < name) OR ('user0.94651659117602724' = name AND 'c7821916-0e40-4459-8746-2ad6bde37700' > id) ORDER BY name ASC,id DESC LIMIT (10 + 1);

in AND umschreiben?


mit tricks in ROW umschreiben:

sure, das wird jetzt natürlich nen seq scan, weil type...
EXPLAIN ANALYZE SELECT id,name,type FROM users WHERE ROW(type, 'user0.04651659117602724', id) < ROW('wter', name, '0655c7e4-cc6a-4013-a0a5-d18b7ff48e44') ORDER BY type DESC, name ASC,id DESC LIMIT (10 + 1);


slow:
EXPLAIN ANALYZE SELECT id,name,type FROM users WHERE ROW(type, 'user0.04651659117602724', id) < ROW('voter', name, '0655c7e4-cc6a-4013-a0a5-d18b7ff48e44') ORDER BY type DESC, name ASC,id DESC LIMIT (10 + 1);
30ms

can this be improved? maybe with an and query or union?


this works AND IS THE FASTEST:

EXPLAIN ANALYZE
(SELECT name,type FROM users WHERE ('voter' = type AND 'user0.04651659117602724' = name AND '0655c7e4-cc6a-4013-a0a5-d18b7ff48e44' > id) ORDER BY type DESC, name ASC,id DESC LIMIT (10 + 1))
UNION ALL
(SELECT name,type FROM users WHERE ('voter' = type AND 'user0.04651659117602724' < name) ORDER BY type DESC, name ASC,id DESC LIMIT (10 + 1))
UNION ALL
(SELECT name,type FROM users WHERE ('voter' > type) ORDER BY type DESC, name ASC,id DESC LIMIT (10 + 1))
LIMIT 11;



maybe with AND:


EXPLAIN ANALYZE SELECT type,name FROM users WHERE ROW(type, 'user0.04651659117602724') < ROW('voter', name) ORDER BY type DESC, name ASC LIMIT (10 + 1);


EXPLAIN ANALYZE SELECT type,name FROM users WHERE type <= 'voter' AND (name >= 'user0.04651659117602724' OR type < 'voter') ORDER BY type DESC, name ASC LIMIT (10 + 1);


just for the trolls using INTERSECT

EXPLAIN ANALYZE
(SELECT type,name FROM users WHERE type <= 'voter')
INTERSECT ALL
(SELECT type,name FROM users WHERE (name >= 'user0.04651659117602724' OR type < 'voter'))
ORDER BY type DESC, name ASC LIMIT (10 + 1);



RESET ALL;
# docker run -it -e MARIADB_ALLOW_EMPTY_ROOT_PASSWORD=true -p 3306:3306 mariadb
# mysql -h 127.0.0.1 -u root

https://github.com/sveltejs/realworld

TODO FIXME https://datatracker.ietf.org/doc/html/draft-ietf-httpbis-rfc6265bis-02#section-5.3.7.1

Reset database (looses all data)

psql -p 54321 -h localhost -U projektwahl --db postgres

set default_transaction_read_only = false;
DROP DATABASE projektwahl;


BEGIN READ WRITE;
DROP TABLE IF EXISTS settings;
DROP TABLE IF EXISTS sessions;
DROP TABLE IF EXISTS choices;
DROP VIEW IF EXISTS present_voters;
DROP TABLE IF EXISTS users;
DROP TABLE IF EXISTS projects;
DROP TYPE IF EXISTS user_type;
COMMIT;

create-svelte

Everything you need to build a Svelte project, powered by create-svelte;

Creating a project

If you're seeing this, you've probably already done this step. Congrats!

# create a new project in the current directory
npm init svelte@next

# create a new project in my-app
npm init svelte@next my-app

Note: the @next is temporary

Developing

Once you've created a project and installed dependencies with npm install (or pnpm install or yarn), start a development server:

npm run dev

# or start the server and open the app in a new browser tab
npm run dev -- --open

Building

Before creating a production version of your app, install an adapter for your target environment. Then:

npm run build

You can preview the built app with npm run preview, regardless of whether you installed an adapter. This should not be used to serve your app in production.