Conversation
|
closes #178 |
|
An unrelated TODO: missing test for column-level CHECK constraints. |
|
The reason I squash merge is that an overwhelming majority of people who contribute to libraries are not very commit hygienic. They don't usually rebase or clean up their work for whatever reason. So squashing is easier for me to manage, and revert whole sale if something was horridly wrong. I can rebase merge this, so it's not a big deal. |
There was a problem hiding this comment.
I intentionally kept the create and create_if_not_exists separate for the executeddl stuff. Ya it's repeated, but it's clear what the function is for.
There was a problem hiding this comment.
This was copied from ecto_sql. I could revert the changes if you'd like to.
There was a problem hiding this comment.
🤔 it probably is better to stay as close to ecto_sql as possible. I do remember there being a reason why I split them other than "just feels" but I can't recall.
|
@mochaaP let me know when you are at a good point to stop for these set of changes. |
|
I'm good for now, try https://reviewable.io/reviews/elixir-sqlite/ecto_sqlite3/184 if force-pushing is noisy for reviews |
2 participants
Note: do not squash merge this PR. Commits in this PR are self-contained and atomic, so squash merging breaks the semantics.
Namely, this branch implements check constraints via
Ecto.Migration.constraint/3.