1. Update URI
2. Drop the cursor from your database (because cursors are unique for every URI)
3. Apply the fix to the parser of commits

I'll make a new URI as default soon.

By default, the database indexes posts based on the local machine's time with datetime.now in server/database.py. However, this can cause timezone issues in a feed's database - e.g. when running from a server in a different country (happened to me today) or due to summertime timezone changes. This means posts in the feed get listed out of order.

I'd suggest changing the default to datetime.utcnow to avoid issues. (Will make a PR in a sec)

I have a repo with a screenshot of the results of the script.

I was trying to create a script to permanently delete a custom feed that was created using SkyFeed. Other feeds I had created with SkyFeed were successfully A) unpublished and then B) deleted without any lingering artifacts. But this particular "zombie" feed references in my repo won't go away, it keeps lingering to just be re-published and un-published, over and over.

In my script, I'm not sure if:

• I have entered the correct value for RECORD_NAME
• where I might obtain the correct value
• whether I need additional code
• or if what I want to do is actually achievable given that the feed was created on SkyFeed; I don't know which server holds the zombie "record" of my unpublished feed for instance

Thank you for any insight.

I recently switched my feed from the starter kit to this implementation but trying to look at the feed just gives the error "could not resolve identity did:web:[etc]"

any ideas what may be the issue?

Everything is running on https and I can reach all the endpoints through the domain.

Hi. I'm new to bsky. Came across your code when looking at AT Protocol

Your Feed example: https://bsky.app/profile/did:plc:kvwvcn5iqfooopmyzvb4qzba/feed/whats-alf
gives error:
could not resolve did document: did:web:feed.atproto.blue

Is this a bksy issue?

Hi, thanks for the python implementation. I have a feed running and would like to generate and serve more feeds from the same firehose. I feel like I’m missing something obvious here, but is it trivial to implement multiple feeds with the existing code? I assume I can create as many algo’s as I need, but I’m having trouble getting my head round how the filter should be set up for this. Does the existing logic easily accomodate multiple feeds?