Now the open function will follow through a bit more and insure that the
proper sequence of events happens during the Open call. This required
some refactoring and a few mild changes in the onEvent func.
* Added ability to change the online status
* Add structs for rick presence implementation
* Refractor and publicise UpdateStatusData
* Add UpdateStatusComplex for raw status data sending
* Case gameType to int, stopped it compiling
* Might want to gofmt.
Doesn't do it on save because Gogland removed it and their new thing I
can't make sense of.
* Revert "Added ability to change the online status"
This reverts commit 235cd15a8eebbec070cb95a5853295387bceae1c.
* Change gametypeto match
* Move RLock to UpdateStatusComplex
* Fix#406: reconnect() can be called while still connected
* Add memberMap to speed up member queries
* Fix error return value and remove deletion
* Fix GuildAdd member map initialization edge case
* Add rate limit buffering support
- Break request into requestInner
- Break LockBucket into LockBucketObject
- Change getBucket to GetBucket so it can be externally accessed
- Add RequestBuffer interface
- Add RequestBuffer pointer to Bucket struct
- Add RequestPostWithBuffer() function to Session
* Remove internal implementation, export ratelimiter instead
Reference Issue: https://github.com/bwmarrin/discordgo/issues/440
This endpoint was declard twice in endpoints.go, and used in the
GuildInvites() method on Session values. I replaced the misspelled
invocation.
All tests pass.
* Fix#406: reconnect() can be called while still connected
* Add memberMap to speed up member queries
* Fix error return value and remove deletion
* Fix GuildAdd member map initialization edge case
* add custom ratelimits
* check for nil ratelimiter
* Don't expose custom ratelimits to Session
* attempt to fix race conditions
* use defer instead
* Slightly improved ratelimiter
You shouldn't need to change the ratelimiters ratelimits while its
running so I removed the functions SetCustomRatelimit and
RemoveCustomRatelimit.
runtime.LockOSThread does not do anything for performance and is instead
intended for interfacing with libraries that use thread local data structures.
Closes#402
* Add heartbeat ACK response and error handling
- Error when sending a heartbeat now triggers a reconnection
- Op7 now triggers a reconnection
- Session now tracks the last heartbeat ACK that was recieved. If the
last ACK is more than FailedHeartbeatAcks*heartbeatinterval in the past,
this is treated as a dead connection and a reconnection is forced.
* Address @iopred comments
* ContentWithMentionReplaced on roles and nicks (Fixes#208)
* Compatibility
* Like this? 🤔
* More efficient regexp
* Tweaked a little
* Tweaked a little more
* Tweaked even more
* Removed unessecary crap condition that is useless
* Disallow voice channel
* Moved regexp declaration
