danclay/eris-fleet

Cluster management for Discord bots using the Eris library.

eris-fleet's Issues

incorrectly typing id parameters as numbers

Describe the bug
context:

eris-fleet/src/util/IPC.ts

Line 49 in 4f4f119

public async fetchUser(id: number): Promise<any> {

eris-fleet/src/util/IPC.ts

Line 62 in 4f4f119

public async fetchGuild(id: number): Promise<any> {

eris-fleet/src/util/IPC.ts

Line 75 in 4f4f119

public async fetchChannel(id: number): Promise<any> {

eris-fleet/src/util/IPC.ts

Line 88 in 4f4f119

public async fetchMember(guildID: number, memberID: number): Promise<any> {

possibly more; unsure

JavaScript doesn't handle discord snowflake IDs properly as numbers; see

These types should be changed to strings (or bigints though thats just silly because they would need to be converted into a string instantly)

This library should also perhaps be tested? testing would've seen that this is an issue

Version N/A

To Reproduce N/A

workerID of undefined

It seems like letting the bot run I get the error of "TypeError: Cannot read property 'workerID' of undefined' occasionally appear. I don't have a consistent way of testing this other than letting the bot run.

Note: The bot was running on Ubuntu

Shutdown stops after failure inside shutdown function

Describe the bug
A clear and concise description of what the bug is.

Version
0.3.3

To Reproduce
Shutdown a worker with some non-functional code inside the the shutdown function

Expected behavior
The kill timeout to kick in and kill the worker forcefully

shard undefined

Describe the bug
when errors are logged, the log that is showen is 2|Counting | Cluster 6, Shard undefined | [DEPRECATED] Member#permission is deprecated. Use Member#permissions instead
Version
0.3.5

To Reproduce
log an eris-fleet error
Expected behavior
my expected behavior is that the shard that the error ocurrd on will show as in this log
2|Counting | Cluster 6, Shard 6 | [DEPRECATED] Member#permission is deprecated. Use Member#permissions instead

shutdowns giving "TypeError: Cannot redefine property: _token"

When using shutdownCluster or totalShutdown an error occurs:
TypeError: Cannot redefine property: _token

This error is caused by a bug in eris. The bug is fixed in the dev branch of eris. Use abalabahaha/eris#dev. Once eris's dev branch is merged into the master branch eris-fleet's eris dependency will be updated.

Prefix missing after restarting

It seems like the prefix of the cluster is missing after restarting the clusters using the ipc, is this intended?

Custom logging

Decent library so far but I'd like to be able to disable built-in logging so I can receive logging events on the main file (ie. index.js) and pipe them into my custom logging functionality.

So instead of something like this:

Or any sort of general logging like shard ready, it'd send a JSON object to an event (via IPC or just by using an event handler on the main Fleet) that can then be handled appropriately.

EDIT: Re clarifying that I'm suggesting a JSON object to be passed instead of just the errors that come right now.

Add more IPC shardstats properties

I would like to see users and guilds to be added to the shardstats function.

Edit: I mean the properties of the shards in the getStats function.

Incorrect amount of users on clusters

Describe the bug
The cluster is showing 1 when the shard itself is not in a guild

Version
v0.2.4

To Reproduce
Console log the clusters users amount

Expected behavior
Should display 0

How do I have the events emitted by Fleet in their own files?

Hi, got a question.

My event loader works for Discord events however I'm not sure how to access the events emitted by Fleet since const Admiral = new Fleet(...) is in index.js and cannot be accessed elsewhere.

I want them in separate files since I use classes which means I can access bot, ipc etc and can use webhooks for Discord logging. It also enables me to post stats received from the stats event to my own API.

Solved by danclay on Discord.

Shard disconnect error with undefined

Describe the bug
Shard disconnect with an error which is undefined, which later the shard resumes.

Version
v0.2.4

To Reproduce
I don't have a way to reproduce it as it seems it just happens while running.

Expected behavior
Show the actual error

Screenshots

Additional context
The issue occurred on a Ubuntu machine I have not tested if this occurs on my Windows machine.

Workers stuck and never connect

Workers sometimes launch after the connect message is sent.
Workers are launching after the connect message is sent from the queue being executed. This causes the log to get stuck at "shards spread."