feat: add API to batch insert identities

[hperl]

This PR does:

• add a batch API for creating multiple identities in one request
• add a generic batch.Create function to create multiple records in one INSERT statement
• uses batch.Create in creating multiple identities (with up to four individual queries depending on the structure of the identitites)

All new code is tested extensively, because the single-identity create is now a special-case of the multiple-identities create.

Related issue(s)

Fixes ory/network#266

Checklist

• Add new API endpoint and definitions
• Implement a generic batch handler
  (Optional:) add delete action and dry_run paramenter

[codecov]

Codecov Report

Merging #3157 (1cac084) into master (08a3447) will increase coverage by 0.24%.
The diff coverage is 90.96%.

❗ Current head 1cac084 differs from pull request most recent head 21605bf. Consider uploading reports for the commit 21605bf to get more accurate results

@@            Coverage Diff             @@
##           master    #3157      +/-   ##
==========================================
+ Coverage   77.68%   77.92%   +0.24%     
==========================================
  Files         319      320       +1     
  Lines       20162    20401     +239     
==========================================
+ Hits        15663    15898     +235     
- Misses       3301     3303       +2     
- Partials     1198     1200       +2     

Impacted Files	Coverage Δ
persistence/sql/identity/persister_identity.go	80.03% <83.69%> (+1.71%)	⬆️
identity/handler.go	86.06% <87.03%> (+0.29%)	⬆️
persistence/sql/batch/create.go	91.04% <91.04%> (ø)
driver/config/config.go	82.87% <100.00%> (ø)
identity/extension_verify.go	97.56% <100.00%> (+0.33%)	⬆️
identity/identity.go	91.75% <100.00%> (+0.44%)	⬆️
identity/manager.go	80.17% <100.00%> (+2.07%)	⬆️
identity/test/pool.go	100.00% <100.00%> (ø)
x/xsql/sql.go	93.33% <100.00%> (ø)

... and 2 files with indirect coverage changes

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[aeneasr]

Thank you! This looks pretty good already, but I had a hard time understanding how the persister creates the batches. I will need to do another pass.

All new code is tested extensively, because the single-identity create is now a special-case of the multiple-identities create.

Not on my watch ;) This code is essential and has security implications. There can not be confusion of credentials or addresses when creating things in batch. Yet, there is not a single test assuring that we are actually creating things correctly when creating multiple things at once. Please add appropriate tests to prove that your changes work as
expected.

Please add dedicated persister tests and manager tests as well. We will need to use them when adjusting the persister for our enterprise use cases in the Ory Network.

Thanks!

[aeneasr]

Ok, took another look at the persister itself. I think the code falls into the classic Go "for range over a slice of pointers" trap. I think this could have serious implications when processing 2...n elements

[aeneasr]

Very nice :) One potential test could be added to ensure the merging of verification addresses works correctly :)