pharo-rdbms / Pharo-SQLite3

Community-owned official SQLite3 binding for Pharo
MIT License
22 stars 20 forks source link

GLORP integration #23

Closed gcotelli closed 4 years ago

gcotelli commented 4 years ago

Depends on pharo-rdbms/glorp#24 . Some tests are failing without the changes in the PR.

astares commented 4 years ago

@gcotelli Please do not adopt the UDBCSQlite3Driver - this one is deprecated and discontinued in my local git account.

I made the SQLite3 driver is now officially available on https://github.com/pharo-rdbms/Pharo-SQLite3 (so it is totally independent from UDBC and more cleaned up)

(so we should change UDBCSQLite3Platform into SQLite3Platform also)

gcotelli commented 4 years ago

@astares sorry if I wasn't clear I've adapted the GlorpDatabaseDriver subclass but to use the code in the repo you mentioned (that it's this one by the way), the connectivity code is the one you want.

gcotelli commented 4 years ago

And yes we need to merge UDBCSQLite3Platform and SQLite3Platform, having two classes for the same platform makes no sense to me. But first I really want to get the CI running the Glorp tests against all the supported database backends and release a version before doing more disruptive changes.

gcotelli commented 4 years ago

This PR run the GLORP tests so it tackles part of #17

gcotelli commented 4 years ago

PR is ready for review, you can check the build run status on https://github.com/ba-st-dependencies/Pharo-SQLite3/runs/903054742?check_suite_focus=true

PierceNg commented 4 years ago

And yes we need to merge UDBCSQLite3Platform and SQLite3Platform, having two classes for the same platform makes no sense to me. But first I really want to get the CI running the Glorp tests against all the supported database backends and release a version before doing more disruptive changes.

Apols I didn't see this earlier. Background: I created UDBCSQLite3Platform as subclass of SQLite3Platform because SQLite3Platform is from Glorp upstream. With the subclass, I made the necessary adaptations to get Glorp+SQLite+Pharo to work without touching SQLite3Platform. And the rationale was that it could be 'cleaner' to sync changes such as these with upstream Glorp, should the need arise.