mirror of
https://github.com/discourse/discourse.git
synced 2026-08-07 13:19:19 +08:00
0a35ae8367
This is the first of a series of changes that rework how the converter and importer handle uploads. Right now too many things are called "uploads": the table in the IntermediateDB, the uploads.db working DB and its `uploads` table, and of course Discourse's own `uploads` table. That makes the code and SQL around uploads harder to follow than it should be. The IntermediateDB table stores where to find each file (path, URL or blob data) - it's not an upload yet - so it's now called `upload_sources`. Purely mechanical rename, no behavior change: the table, its `Upload` model (now `UploadSource`), the generated schema SQL, and every query that touched the IntermediateDB table. The shard-merge consolidator resolves the model by singularizing the table name, so `upload_sources` -> `UploadSource` keeps working with no change there. Untouched on purpose: `*upload*_id` reference columns, `mapped.ids`, the source Postgres `uploads` table the converter reads from, the output uploads.db `uploads` table (reworked in a later change), and the importer step named `uploads`. |
||
|---|---|---|
| .. | ||
| lib | ||
| spec | ||
| .rspec | ||
| Gemfile | ||
| migrations-converters.gemspec | ||
