Tags

1.27.12rc1

Release 1.27.12rc1

1.27.11

Release 1.27.11

1.27.11rc5

Release 1.27.11rc5

1.27.11rc4

Release 1.27.11rc4

1.27.11rc3

Release 1.27.11rc3

1.27.11rc2

1.27.11rc1

Release 1.27.11rc1

1.27.10

## Improved performance
* Replays to 92M blocks in 47 hours on an AMD9900 with 64GB RAM and 2x4TB T700 nvme drives (full details here: https://gitlab.syncad.com/hive/hive/-/issues/556#note_200975).
* Most frequently used API calls are 3-4x faster due to switch to pure SQL queries on server side coupled with switch to postgresT as the web server.
* Reduced database storage and speeded up performance by eliminating unnecessary columns like ([preview field](https://gitlab.syncad.com/hive/hivemind/-/merge_requests/836), [img_url field](https://gitlab.syncad.com/hive/hivemind/-/merge_requests/842) and [indexes](https://gitlab.syncad.com/hive/hivemind/-/merge_requests/830).
* Most API calls that previously could take several seconds in problematic cases were optimized to eliminate such problems.

## New features
* Configurable timeout time for API calls to prevent server slowdown on problematic API requests
* [get_profiles API call](https://gitlab.syncad.com/hive/hivemind/-/merge_requests/820)

## Compatibility with older versions

### All the calls below are now limited to a max of 20 posts (parameter limit <= 20)

Note: there is an option to run a "special" API node that returns more than 20 posts, in case there is an app that can't live without the ability to fetch more than 20 full posts at a time.

```
condenser_api.get_discussions_by_blog
condenser_api.get_discussions_by_feed
condenser_api.get_discussions_by_author_before_date (Note: default amount fetched also increased from 10 to 20 to match other calls)
condenser_api.get_discussions_by_comments
condenser_api.get_replies_by_last_update
condenser_api.get_discussions_by_created
condenser_api.get_discussions_by_hot
condenser_api.get_discussions_by_promoted
condenser_api.get_discussions_by_trending
condenser_api.get_post_discussions_by_payout
condenser_api.get_comment_discussions_by_payout
condenser_api.get_blog

bridge.get_account_posts
bridge.get_ranked_posts

Same applies to these tag calls, but they were deprecated a long time ago and we don't think anyone is using them:
tags_api.get_discussions_by_blog
tags_api.get_discussions_by_comments
tags_api.get_discussions_by_author_before_date (max limit decreased to 20, also default increased from 10 to 20)
tags_api.get_discussions_by_created
tags_api.get_discussions_by_hot
tags_api.get_discussions_by_promoted
tags_api.get_discussions_by_trending
tags_api.get_post_discussions_by_payout
tags_api.get_comment_discussions_by_payout
```

Note that `bridge_api.get_discussion` (hardcoded limit of 2000 replies), `tags_api.get_discussion`, and `condenser_api.get_content` (hardcoded limit of 5000 replies) are still supported and haven't changed behavior since they only fetch one post. 

However: only the bridge_api version properly honors the observer parameter. Tags doesn't have the option and the option has never worked properly for condenser_api version (it ignores it). 

Therefore bridge_api.get_discussion version is HIGHLY recommended versus the other 2 versions. The other versions are deprecated and scheduled for likely removal in the future. The bridge_api is also by far the most popular version and will cache much better.

### `condenser_api/follow_api.get_following` calls will not work properly for users who follow more than 10K accounts

For users who follow large numbers of other accounts, this functionality was already basically broken because the query to fetch the accounts that were followed took too long to calculate. 

To speed this query up, get_following will now return random results for accounts that follow more than 10K accounts and paging will not work (so it is kind of useless to call this for accounts that follow more than 10K accounts, but now it won't put excessive loads on the API node itself).

Frontends may want to consider warning users if they attempt to follow more than 10K accounts and block such transactions. But I suspect that most such large follow lists are created by bots, not by using normal frontends.

Note that `get_followers` and `get_follow_list` calls have not changed.

### `list_subscribers orders subscribers by name instead of by id

### `list_comments` call has being removed from API
I polled devs a while back and no one reported using this call, nor do we see calls to it on our production site (and paging is sort of broken for it too). Eliminating this call will save substantial disk storage on API nodes.

### Removed hive_payments table and "promoted" feature

1.27.9

Release 1.27.9

1.27.8

Release 1.27.8

1.27.7rc16

Release 1.27.7rc16

1.27.7rc15

Release 1.27.7rc15

1.27.7rc14

Release 1.27.7rc14

1.27.7rc13

Release 1.27.7rc13

1.27.7rc12

Release 1.27.7rc12

1.27.7rc11

Release 1.27.7rc11

1.27.7rc10

1.27.6rc9

1.27.6rc8

1.27.6rc7