Sequel ORM support #1277
Sequel ORM support #1277
Conversation
There was an attempt to write sequel adapter in cerebris#1030. I took commits from this PR. This commit fixes most of the failures that are result of applying these commits into much newer codebase. Most of the code is also copied from the original PR.
lukesarnacki
force-pushed
the
sequel-adapter
branch
from
3acdce5
to
368105a
Compare
|
I initially copied a lot of the code from the original PR but then I found a way to use original commits. There are still 4 failures on my local env and seems like CI has lots of them. So the first step would be to fix all of these failures for Forgot to write in the PR description: please let me know if you are still willing to accept this PR. And when it comes to current work, are you ok with presented fixtures setup (in short: active record schema is dumped to a file with a rake task; it is then loaded into the database when test are run)? |
|
@lukesarnacki I haven't taken a detailed look at this PR yet, but I am on board with adding the SequelRecordAccessor, though I wonder if it would be better as a separate gem. I'll try to dig into this soon and offer some guidance. The CI failure should now be fixed if you rebase against master. |
I want to revisit the subject of sequel ORM support that was started in this issue: #1006
There is a PR that was started quite some time ago with an attempt to do so.
I copied a lot of the code for fixtures setup and models definitionsI cherry picked as many commits as I could from the original PR so kudos for @Aryk for starting this. I didn't want to start from that PR as some of the code is outdated and I wanted to start from database setup only (no core changes yet).I want to start draft PR early for any suggestions that you may have and I want to discuss changes early to avoid pushing a lot of code. I am not sure what is the direction I want to go but I think I would like to change inheritance to composition and just delegate methods from
Resourceto object representing "active relation" or "sequel dataset" instead of inheriting from it. What do you think?