Running db sqlite query within a go call

[ImmanuelSamuel]

The following code works without the 'go' based function call - what am I doing wrong here?

import db.sqlite

struct Dbs {
mut: 
db sqlite.DB
}

struct Login {
    id int
    name string
}

fn main() {
  db1 := sqlite.connect('test.db') or { panic(err) }
  shared dbs := &Dbs{db: &db1}

  go test(shared dbs)
}

fn test(shared dbs Dbs) ! {
  eprintln('test called')

  login := Login{name: 'Test'}

  sql dbs.db { insert login into Login } or { panic(err) }
  foo := sql dbs.db { select count from Login } or { panic(err) }
  eprintln('${foo}')

  eprintln('test finished')
}

[jorgeluismireles]

test could be used by several threads so is important to use lock:

fn test(shared dbs Dbs) ! {
	lock dbs {
		eprintln('test called')
		login := Login{name: 'Test'}
		sql dbs.db { insert login into Login } or { panic(err) }
		foo := sql dbs.db { select count from Login } or { panic(err) }
		eprintln('${foo}')
		eprintln('test finished')
	}
}

[JalonSolov]

Or, slightly nicer to the rest of the code:

fn test(shared dbs Dbs) ! {
	eprintln('test called')
	login := Login{name: 'Test'}
	lock dbs {
		sql dbs.db { insert login into Login } or { panic(err) }
		foo := sql dbs.db { select count from Login } or { panic(err) }
	}
	eprintln('${foo}')
	eprintln('test finished')
}

In other words, only lock when you need it. It's not needed around the eprintln() calls.

[ImmanuelSamuel]

Even with the lock the query is not executed. Did the above code work for you? (maybe I need to update V?! I am on v4.10 - fairly recent)

[jorgeluismireles]

You forgot to create the table, without shared things this works:

import db.sqlite

struct Dbs {
mut: 
db sqlite.DB
}

struct Login {
    id int
    name string
}

fn main() {
  db1 := sqlite.connect('test.db') or { panic(err) }
  dbs := &Dbs{db: &db1}
  test(dbs)!
}

fn test(dbs Dbs) ! {
  eprintln('test called')
  sql dbs.db { create table Login }!
  login := Login{name: 'Test'}
  sql dbs.db { insert login into Login }!
  foo := sql dbs.db { select count from Login }!
  eprintln('${foo}')
  eprintln('test finished')
}

C:\Users\jmireles\issues>v run db.v
test called
5
test finished

[ImmanuelSamuel]

Thanks for the response - Yes, I already have a table that I am creating things into and Yes I have already run the code without using 'go' and it works for me too - my main question as stated in the discussion title is how to do what you have shown above using 'go' and 'spawn' (async sql ops essentially) - Thank you for the self-contained code above - hopefully others can use this for testing.

[jorgeluismireles]

If your data must be accessed by more than one user you can use the veb module and share the database object in the endpoints functions and using lock and the context to manage timeouts and the like.

…

On Sun, Apr 13, 2025, 7:04 AM ImmanuelSamuel ***@***.***> wrote: Thanks for the response - Yes, I already have a table that I am creating things into and Yes I have already run the code without using 'go' and it works for me too - my main question as stated in the discussion title is how to do what you have shown above using 'go' and 'spawn' (async sql ops essentially) - Thank you for the self-contained code above - hopefully others can use this for testing. — Reply to this email directly, view it on GitHub <#24203 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA6A4WT7SB4TVCJTONWTUPD2ZJOGDAVCNFSM6AAAAAB3ANES62VHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTEOBRHE2DGMA> . You are receiving this because you commented.Message ID: ***@***.***>

[ImmanuelSamuel]

Yes, I have used veb before to do that - but that seems like a hacky way to do it. Don't want to import a whole web framework just to write to a database. I know that veb can access the same db connection across different coroutines - just want to know how to do the same thing without using the veb framework - there must be a way since veb does that. I tried reading through veb implementation but can't figure out how this is implemented in the code.

Also the above use case doesn't throw an error if its not able to access the connection - no debug messages either. Based on the responses here it seems like when people first read the code above they expect it to work based on the documentation - so afaik it seems like a bug to me. Didn't want to open an issue before discussing here to see if I missed something obvious.