* [feature] Read + Write tombstones for deleted Actors
* copyTombstone
* update to use resultcache instead of old ttl cache
Signed-off-by: kim <grufwub@gmail.com>
* update go-cache library to fix result cache capacity / ordering bugs
Signed-off-by: kim <grufwub@gmail.com>
* bump go-cache/v3 to v3.1.6 to fix bugs
Signed-off-by: kim <grufwub@gmail.com>
* switch on status code
* better explain ErrGone reasoning
Signed-off-by: kim <grufwub@gmail.com>
Co-authored-by: kim <grufwub@gmail.com>
* S3: add config value "proxy" for not redirecting
Signed-off-by: Mara Sophie Grosch <littlefox@lf-net.org>
* S3: document new config value "proxy"
* S3: add new config value "proxy" to test scripts
Signed-off-by: Mara Sophie Grosch <littlefox@lf-net.org>
* re-add eslint
* fix oauth url getting too long
* actually attach single emoji get and delete routes
* basic emoji details + deletion using rtk query
* refactor emoji upload to rtk query
* clean up old redux api+reducers for custom emoji
* fix validation order
* refactor custom emoji form fields
* remove unused requires
* cleanup, fix most eslint errors
* more small eslint fixes
* eslint & web bundle in CI
* fix directories, move lint+build to scripts
* working dirs
* test linter and build errors
* upgrade skulk for better error handling
* Revert "test linter and build errors"
This reverts commit dc047d5b444c41a90209d02bb4ed4893ccf0885f.
* fix max emoji size
* tiny bit of function documentation
* sign drone.yml
Co-authored-by: tsmethurst <tobi.smethurst@protonmail.com>
* [feature] Make instance thumbnail configurable via admin panel
* log db errors in InstanceToAPIInstance
* only update instance in db if necessary
* start adding tests
* finish test
* re-add eslint
* fix oauth url getting too long
* actually attach single emoji get and delete routes
* basic emoji details + deletion using rtk query
* refactor emoji upload to rtk query
* clean up old redux api+reducers for custom emoji
* fix validation order
* refactor custom emoji form fields
* remove unused requires
* cleanup, fix most eslint errors
* more small eslint fixes
* fix max emoji size
* tiny bit of function documentation
* greatly simplify httpclient request queuing
Signed-off-by: kim <grufwub@gmail.com>
* improved request queue mutex logic
Signed-off-by: kim <grufwub@gmail.com>
* use improved hashmap library
Signed-off-by: kim <grufwub@gmail.com>
* add warn logging when request queues are full
Signed-off-by: kim <grufwub@gmail.com>
* improve worker pool prefix var naming
Signed-off-by: kim <grufwub@gmail.com>
* improved worker pool error logging
Signed-off-by: kim <grufwub@gmail.com>
* move error message into separate field
Signed-off-by: kim <grufwub@gmail.com>
* remove old log statement
Signed-off-by: kim <grufwub@gmail.com>
* don't export worker message, it gets very spammy :')
Signed-off-by: kim <grufwub@gmail.com>
Signed-off-by: kim <grufwub@gmail.com>
* update go-store to v2.0.6: closer callbacks are now only ever called at most once
Signed-off-by: kim <grufwub@gmail.com>
* bump go-store => v2.0.7, go-mutexes => v1.1.4
Signed-off-by: kim <grufwub@gmail.com>
Signed-off-by: kim <grufwub@gmail.com>
* bump go-store version to v2.0.5, init kv.KVStore without initial clean (as we are using for storage, not as a key-value store)
Signed-off-by: kim <grufwub@gmail.com>
* remove newline
Signed-off-by: kim <grufwub@gmail.com>
Signed-off-by: kim <grufwub@gmail.com>
If set, the landing page user configuration value is used as a Gin
context parameter, which seems incorrect, since a normal request isn't
going to have a parameter named after an arbitrarily configured user.
Instead, the user name should be used directly when building the
redirect URL.
* use readcloser for content.Content
* call media postdata function no matter what
* return a readcloser from data func
* tidy of logic of readertostore
* fix whoopsie
The migration that adds the `admin_account_actions` table did so at the
same time as adding indexes onto the new table. This code was ran inside
a `RunInTx` function, but the table creation did not use the transaction
reference, while the creation of the indexes did. This could cause a
race between the table and index creations, depending on the scheduling
order. If the table creation did not win the race, then the migration
would fail.
This changeset corrects the table creation to also be done inside the
same transaction as the index creation.
Signed-off-by: Terin Stock <terinjokes@gmail.com>
Signed-off-by: Terin Stock <terinjokes@gmail.com>
