punyaslokdutta tries to fold storage into one backend
A proposal to drop a second storage service and lean entirely on the tool the fork already runs, then quietly shelved.
This fork stores its files - uploads, downloads, signed links - on Cloudflare's object-storage service, which needs its own account, keys, and setup. punyaslokdutta pointed out that Supabase, the backend the project already uses for logins and its database, can do all the same file jobs. The pitch: cut the extra service entirely, and run one fewer vendor account and one fewer set of credentials.
The tradeoff was operational, not about features. Fewer moving parts to provision and manage, in exchange for putting all your eggs in the Supabase basket. The proposal was closed after a couple of days without being merged, so it never landed. Whatever the objection was - durability, cost of moving data out, or discomfort with leaning so hard on a single vendor - it wasn't written down.
Spotted something wrong? Or know the PR text has fresher detail than the writeup above?