Fix sequence deletion after nextval call

[agscpp]

CREATE SEQUENCE serial_01 CYCLE;
SELECT nextval('serial_01');
DROP SEQUENCE serial_01;

Tarantool error: DropSequence: Can't drop sequence 'serial_01': the sequence has data

[agscpp]

Hi, @gmoshkin .

Can you please take a look at my pull request?

[gmoshkin]

Hi, thanks for the patch!

Could you explain, how are the provided SQL commands related to the test?

CREATE SEQUENCE serial_01 CYCLE;
SELECT nextval('serial_01');
DROP SEQUENCE serial_01

The code in question will not be invoked when this SQL is executed.

Could you also add a test, so that we don't break the fix accidentally in the future.

[agscpp]

This is an abstract example that I used to compare the behavior with other DBMSs.

Okay, I'll write tests for this logic.

[agscpp]

@gmoshkin
I wrote the tests

[gmoshkin]

Also please update CHANGELOG.md, we try keeping it up to date with every merge we make into the master branch.

[agscpp]

@gmoshkin

[gmoshkin]

thanks for the test! There's a problem though, this hard-coded id is volatile as it may change if another test is added (if someone adds another sequence). We could make it much more robust if we got the id from the Sequence struct. Unfortunately there's currently no method for this, but it's very easy to add it.

Do you mind adding fn id(&self) -> u32 for the Sequence struct and using it in this test?

This will also be potentially useful for other users of tarantool-module as it will allow explicit deletion of sequences, by adding the ability to get the id of the instance.

[gmoshkin]

Apologies, I wrote this comment yesterday, but failed to submit the review

[agscpp]

I have corrected the code based on your comments. Please take a look again.

[agscpp]

@gmoshkin

[agscpp]

@gmoshkin

[agscpp]

@gmoshkin
ping