* replicate issue
* update go-structr to v0.8.10 with internal linked-list fix, small tweaks to caching of interaction requests
* remove debug function
---------
Co-authored-by: tobi <tobi.smethurst@protonmail.com>
* [docs] Don't run SQLite on networked storage
This updates our documentation to include warnings against running
SQLite on networked storage.
It also slightly reorganises and consolidates some admonitions on the
Getting Started page because it was getting a bit too busy with repeated
information.
* [docs] Reword a few things
* bump go-byteutil v1.2.0 -> v1.3.0 which has safer (as in long-term API consistency) byte <-> string conversions
* fix test relying on byteutil exported type no longer existing
* update go-storage dependency, for S3Storage manually call PutObject() so we can set content-type
* update calls to PutFile() to include the contentType
* ensure that errors checking status visibility / converting aren't dropped
* add some more context to error messages
* include calling function name in log entries
* don't error on timelining hidden status
* further code to ignore statusfilter.ErrHideStatus type errors
* remove unused error type
* add local instance status / domain / user counts
* add checks for localhost
* rename from InstanceCounts to LocalInstance
* improved code comment
* add error message to gin context on header blocked or not allowed
* remove the unused header filter tracking code (leaving OTEL TODOs in place)
* appease the linter
* improved federatingdb logging in cases of unknown iri / types, add new log methods
* whoops; forgot to wrap log argument in serialize{} !
* use debug instead of warn level
* switch last entry to Debug
* improvements to caching for lists and relationship to accounts / follows
* fix nil panic in AddToList()
* ensure list related caches are correctly invalidated
* ensure returned ID lists are ordered correctly
* bump go-structr to v0.8.9 (returns early if zero uncached keys to be loaded)
* remove zero checks in uncached key load functions (go-structr now handles this)
* fix issues after rebase on upstream/main
* update the expected return order of CSV exports (since list entries are now down by entry creation date)
* rename some funcs, allow deleting list entries for multiple follow IDs at a time, fix up more tests
* use returning statements on delete to get cache invalidation info
* fixes to recent database delete changes
* fix broken list entries delete sql
* remove unused db function
* update remainder of delete functions to behave in similar way, some other small tweaks
* fix delete user sql, allow returning on err no entries
* uncomment + fix list database tests
* update remaining list tests
* update envparsing test
* add comments to each specific key being invalidated
* add more cache invalidation explanatory comments
* whoops; actually delete poll votes from database in the DeletePollByID() func
* remove added but-commented-out field
* improved comment regarding paging being disabled
* make cache invalidation comments match what's actually happening
* fix up delete query comments to match what is happening
* rename function to read a bit better
* don't use ErrNoEntries on delete when not needed (it's only needed for a RETURNING call)
* update function name in test
* move list exclusivity check to AFTER eligibility check. use log.Panic() instead of panic()
* use the poll_id column in poll_votes for selecting votes in poll ID
* fix function name
tobi