Adding CI

[hjmtql]

It will be easier to maintain and contribute to this awesome library.

[hjmtql]

separate tests that use pg backend because they may cause conflicts related to transaction.

[hjmtql]

I often got error when running test before

| Error: (IntegrityError { code: "23505", column: "", constraint: "pg_type_typname_nsp_index", dataType: "", detail: "Key (typname, typnamespace)=(people, 2200) already exists.", file: "nbtinsert.c", hint: "", internalPosition: "", internalQuery: "", line: "570", message: "duplicate key value violates unique constraint \"pg_type_typname_nsp_index\"", position: "", routine: "_bt_check_unique", schema: "pg_catalog", severity: "ERROR", table: "pg_type", where_: "" })

[hjmtql]

There was a transaction overlapping when the error occured.

↓ postgres logs

db_1  | 2021-02-05 05:30:01.581 UTC [81] LOG:  connection received: host=192.168.176.1 port=44946
db_1  | 2021-02-05 05:30:01.581 UTC [82] LOG:  connection received: host=192.168.176.1 port=44948
db_1  | 2021-02-05 05:30:01.581 UTC [83] LOG:  connection received: host=192.168.176.1 port=44950
db_1  | 2021-02-05 05:30:01.585 UTC [82] LOG:  connection authorized: user=init database=purspg
db_1  | 2021-02-05 05:30:01.585 UTC [81] LOG:  connection authorized: user=init database=purspg
db_1  | 2021-02-05 05:30:01.586 UTC [83] LOG:  connection authorized: user=init database=purspg
db_1  | 2021-02-05 05:30:01.589 UTC [81] LOG:  statement: BEGIN TRANSACTION
db_1  | 2021-02-05 05:30:01.590 UTC [81] LOG:  statement: 
db_1  | 	  DROP TABLE IF EXISTS people;
db_1  | 	  CREATE TABLE people (
db_1  | 	    id INTEGER PRIMARY KEY,
db_1  | 	    name TEXT NOT NULL,
db_1  | 	    age INTEGER
db_1  | 	  );
db_1  | 2021-02-05 05:30:01.592 UTC [82] LOG:  statement: BEGIN TRANSACTION
db_1  | 2021-02-05 05:30:01.592 UTC [83] LOG:  statement: 
db_1  | 	        DROP TABLE IF EXISTS people;
db_1  | 	        CREATE TABLE people (
db_1  | 	          id INTEGER PRIMARY KEY,
db_1  | 	          name TEXT NOT NULL,
db_1  | 	          age INTEGER
db_1  | 	        );
db_1  | 	
db_1  | 	        DO $$
db_1  | 	        BEGIN
db_1  | 	          IF NOT EXISTS (SELECT 1 FROM pg_type WHERE typname = 'account_type') THEN
db_1  | 	            CREATE TYPE ACCOUNT_TYPE as ENUM (
db_1  | 	              'business',
db_1  | 	              'personal'
db_1  | 	            );
db_1  | 	          END IF;
db_1  | 	        END$$;

[hjmtql]

It might have been best if it could be split in the code, but it was difficult so I split them using npm-run-all.
@Kamirus How do you think the solution is good?

[hjmtql]

fix #39

[Kamirus]

Thank you for your effort! CI is very much welcome indeed.
Errors you've encountered are probably caused by some parallel execution - do they still occur? Or separating guides from other tests helped?

Also maybe @paluh might know more.

[hjmtql]

@Kamirus

Thank you for your quick reply!

Or separating guides from other tests helped?

-> Yes it does.
npm-run-all --sequential runs in sequence, so it doesn't seem to cause parallel execution problems.

I have never faced any error at the moment (on my local and github actions) since I separated PG tests.
test-branch: https://github.com/hjmtql/purescript-selda/runs/1839387218

[Kamirus]

Cool thank you very much!
@paluh do you think we should add anything or this is ready to merge?

[paluh]

Thanks a lot @hjmtql for this contribution! This is really cool. For a long time we weren't even sure whether it is possible to setup such a CI :-)
When purescript-cookbook finishes migration to github actions we are probably nearly ready to go!

@paluh do you think we should add anything or this is ready to merge?

I think we can merge this.
I'm not sure if we want to rebase it on the devel branch (with force push) and fix possible problems there before the final merge to master or we are going to merge devel later and fix all problems on the master. Either way should work for our small team :-P
But speaking a bit more seriously we should probably think about reliable and official branching flow for the project ;-)

[hjmtql]

I used docker-compose newly, so I added short notes to README.md

[Kamirus]

Great! @hjmtql would you like to add anything else or do you think it's ready to merge?

[hjmtql]

nothing. yeah it's ready to merge! 2021年2月7日(日) 2:37 Kamil Listopad <notifications@github.com>:

…

Great! @hjmtql <https://github.com/hjmtql> would you like to add anything else or do you think it's ready to merge? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#41 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABUBNTKE4YGERUJR4ZC32K3S5V455ANCNFSM4XEERUEQ> .