Allow diacritics/accents characters for User schema in SQL Server DB - TypeORM

[jiriludvik]

Summary of proposed feature
Hello, recently I discovered a problem when working with database sessions stored in SQL Server (TypeOrm adapter). My setup uses Azure Active Directory (oauth2.0) to sign users in. The thing is that as I have there user with name "Jiří" - this creates new record in DB table "users" with name = "JiYí".

Having checked the docs for MS SQL schema seems the problem is that column "name" is of type varchar. Another problem could be in schema definition which states the same.

Purpose of proposed feature
Would it be possible to update required definitions to use nvarchar instead of varchar to handle accented characters?

Potential problems
Not sure about potential problems as I'm quite new in Javascript and Next.JS field ;)

[iaincollins]

Oh no! Sorry you ran into this issue and thank you for raising it.

Yes that change seems sensible, if it's mangling names that's a bug.

[ndom91]

Re: diacritics and other special latin characters, usually you want to set the db charset to utf-8. I did a bit of digging on what to do there with typeorm and came across this: typeorm/typeorm#390.

You basically just want to add "charset": "utf8mb4_unicode_ci" to the connection options. Unfortunately this won't fix existing databases / apps, but new installations will profit from this change. For existing databases / apps, you can try converting a table / db like so:

ALTER DATABASE dbname CHARACTER SET utf8 COLLATE utf8_general_ci;
ALTER TABLE tbl_name CONVERT TO CHARACTER SET utf8 COLLATE utf8_general_ci;

[balazsorban44]

@ndom91 is there anything to do here in the adapter code?

[ndom91]

I believe you can define the utf8 character set when defining the db.

Might make sense to append to our seed file.

EDIT: after some research, it seems this is the default character set with mysql8 👍

So nothing to do as far as I see, should be supported out of the box