#archive-library-discussion — Page 1

tough geode Mar 13, 2020, 2:22 PM

#

oh, right

snow crypt Mar 13, 2020, 2:23 PM

#

https://github.com/discordjs/discord.js/blob/master/src/client/Client.js#L134

slate nacelle Mar 13, 2020, 2:57 PM

#

@tough geode another reason would be that the id of the client user is not necessarily the id of the application.
This is only the case for very old applications however.

ornate topaz Mar 13, 2020, 2:58 PM

#

(Nadeko bot would be one example)

vernal atlas Mar 13, 2020, 5:41 PM

#

i can't seem to get TS to recognise that messages is meant to be an array?

#

using the normal send overloads, not from my PR

spice lava Mar 13, 2020, 5:46 PM

#

you can cast it <Channel>.send().then(msg => (msg as Message[]).length ...)

vernal atlas Mar 13, 2020, 5:50 PM

#

should have explained but, i made a PR to remove some duplicate types as parameters for the send overloads, but then crawl commented saying that would break it, but i can see the overloads arent even working as they should in the first place @spice lava

#

so, not trying to fix an issue on my end, but with the send overloads

#

i think i found the problem, the overloads with only options need to be defined first, i think its something to do with the fact StringResolvable is any

snow crypt Mar 13, 2020, 6:33 PM

#

i pointed that out earlier

#

because send takes StringResolvable it would not autocomplete or type check

vernal atlas Mar 14, 2020, 2:57 AM

#

ah

solemn kite Mar 14, 2020, 6:37 AM

#

are intents fully implemented yet?

#

I imagine the new cache way is to make way for the intents system instead because cache is gonna be redundant at some time

#

... right? cause I assume its gonna be because of the intents you're unable to really make sure that the dats you have in the cache is gonna be there so its gonna be deprecated soon?

copper laurel Mar 14, 2020, 7:35 AM

#

Uhh, no?

#

Just specify all the intents so that we can cache stuff

#

Intents are available but not yet required

#

But there are some issues where if intents arent specified cache will be missing, yes

ornate topaz Mar 14, 2020, 1:24 PM

#

Even if missing, it will definitely not be redundant though. That would mean that bot doesn't cache literally anything at all, while you can just use the intent to get data for cache back again (assuming your bot will be allowed to use all intents)

analog oyster Mar 14, 2020, 4:35 PM

#

https://i.imgur.com/7cVhGLj.png

#

'none' | 'all' | 'everyone' aren't types PES_Hands

real jetty Mar 14, 2020, 4:38 PM

#

No the typings are correct, they are valid types in TS

#

You just didn't use it correctly

#

What is the code you tried?

analog oyster Mar 14, 2020, 4:39 PM

#

https://i.imgur.com/FzLzffh.png

#

what can I even do about this

real jetty Mar 14, 2020, 4:40 PM

#

Yeah like that, now you're encountering "type widening"

#

You can do disableMentions: 'everyone' as 'everyone'

analog oyster Mar 14, 2020, 4:40 PM

#

oh, thank you

#

sorry for coming to the wrong channel ig

copper laurel Mar 14, 2020, 9:23 PM

#

Or define CLIENT_OPTIONS as ClientOptions

#

Might also work /superlate

thorny cosmos Mar 15, 2020, 11:03 AM

#

Ok, I just came across something, <Guild>.verificationLevel returns a VerificationLevel typedef, which has a type of String

#

However, when I see the typeof this thing, it says it returns a number actually

#

same goes for the ExplicitContentFilterLevel and so... I don't know if this was done with intention for showing the string type in the respective methods which takes a parameter as a string | number, I think it would be better if the numbers are shown along with the Level alongside in the TypeDef and make the Types as string | number

real jetty Mar 15, 2020, 2:53 PM

#

version 12 of discord.js is a poop

opal mauve Mar 15, 2020, 2:53 PM

#

please don't randomly complain here. this channel is for serious discussion about the library.

vernal atlas Mar 15, 2020, 3:10 PM

#

@thorny cosmos oh, i think that came from a PR i made, maybe a merge conflict or smth changed it without realising

#

https://github.com/discordjs/discord.js/pull/3742/files#diff-e5c54069adfa0d32480eb3cc9faeee20R256

#

interesting

#

hm

#

@thorny cosmos i assume you are talking about

#

https://github.com/discordjs/discord.js/blob/69d69f25b964b6a2bb13bd086696abec17782ca3/src/structures/Guild.js#L254 this line?

#

because it looks as it should to me

#

perhaps your version is out of date

thorny cosmos Mar 15, 2020, 3:18 PM

#

Oh, you are right@vernal atlas , I realised i am using an older commit of v12, well, will update then

vernal atlas Mar 15, 2020, 3:19 PM

#

👍

green sentinel Mar 15, 2020, 5:59 PM

#

I think Channel should have a partial field, since when you use client.on("channelUpdate",...) you get a Channel | PartialChannel.

#

I use typescript so to check if the channel is partial I either have to just channel.fetch() or do something like

(channel as PartialChannel).partial ? channel.fetch() : channel as Channel;

vivid field Mar 15, 2020, 7:33 PM

#

DMChannels have a property showing if it's a partial or not, other channels are never partials as guild channels are always cached

raven juniper Mar 15, 2020, 7:34 PM

#

you can remove guild channels from cache

green sentinel Mar 15, 2020, 7:41 PM

#

I think it just makes sense to put a partial property on a value that can possibly be partial. Call me crazy.. lolol

copper laurel Mar 16, 2020, 12:20 AM

#

Theres already a PR in place for this

#

GuildChannel#partial will always return false
DMChannel#partial will do the appropriate check
Typings will be improved so directly handling the Channel class shouldnt happen, since its supposed to be abstract

tough geode Mar 16, 2020, 11:46 AM

#

https://github.com/discordjs/discord.js/blob/9c8aaf1bbcc220299e65cadb6ac54e1a7f208576/src/client/Client.js#L389 shouldn't this check for only 'all' | 'everyone' | 'none' instead of the type string?

opaque vessel Mar 16, 2020, 11:53 AM

#

Probably

#

It's specified in the typings as such

unique axle Mar 16, 2020, 5:53 PM

#

this channel is for library discussion not support

split steeple Mar 16, 2020, 5:54 PM

#

sry

finite rover Mar 16, 2020, 10:51 PM

#

is there any reason djs doesn't use ms precision ratelimit?

tired kindle Mar 17, 2020, 12:11 AM

#

~~you can't~~ you don't receive milliseconds precision from the API

#

https://discordapp.com/developers/docs/topics/rate-limits#header-format

X-RateLimit-Reset:
Epoch time (seconds since 00:00:00 UTC on January 1, 1970) at which the rate limit resets

#

it's in seconds, from the Discord API

opal mauve Mar 17, 2020, 12:19 AM

#

@finite rover discord.js does use ms precision

finite rover Mar 17, 2020, 1:14 AM

#

it's in s unless you tell it to give you ms

#

@opal mauve would you mind showing me the relevant line?

#

i ctrl-f ed the repo to find where the x-ratelimit-precision header was and I couldn't find it but maybe i missed smthing

opal mauve Mar 17, 2020, 1:16 AM

#

@finite rover oh, my bad. it's included in @tacit crypt's REST PR which is still in draft form.

#

https://github.com/discordjs/discord.js/pull/3443

finite rover Mar 17, 2020, 1:17 AM

#

yeah exactly what i found as well

#

so any reason this hasn't been introduced sooner?

#

it shouldn't imply much change apart from sending the additional header right?

tacit crypt Mar 17, 2020, 1:18 AM

#

It wasn't really needed?

opal mauve Mar 17, 2020, 1:19 AM

#

the PR that's adding it isn't finished yet

tacit crypt Mar 17, 2020, 1:19 AM

#

I mean you're not gaining SO much speed from it

finite rover Mar 17, 2020, 1:19 AM

#

that's fair, honestly I'm purely asking for information here

tacit crypt Mar 17, 2020, 1:19 AM

#

Maybe I'll actually finish that PR one day

#

After rewriting the entire bucket system...again

finite rover Mar 17, 2020, 1:20 AM

#

I was not even talking about the PR that seems to introduce much more than just ms precision

#

I was mostly wondering if adding ms precision would needmore changes than it seems

#

and that maybe I don't know about

tacit crypt Mar 17, 2020, 1:21 AM

#

I mean, to keep it simple

Can we use it? Absolutely
Should we? Probably?
Is it that important we use it? Not really
Is it gonna need massive changes? Nop, one liner. Testing tho... Yike

finite rover Mar 17, 2020, 1:21 AM

#

Alright

#

thanks for the answer

#

That's what I thought, I wanted confirmation thank you

#

Also yes I guess it doesn't matter that much theshrug

real jetty Mar 17, 2020, 7:33 AM

#

I have been thinking. But in the message reaction remove event, when fetching cuz it's a partial. this.emoji could be a guild emoji what could trigger the error. But why does message reaction add event doesn't have that problem? think

real jetty Mar 17, 2020, 8:12 AM

#

@real jetty I'll be opening a pr today that fixes this issue with an explanation to y all this mess happens

#

Okie

real jetty Mar 17, 2020, 4:23 PM

#

There is a lot of text and stuff but basically you use only messageReaction.fetch() where it would also fetch the message automatically? (if the message was also partial) (In the new solution such to speak)

#

yh thats pretty much wat it was doing before, now it just uses Message#_patch() to end up with the same result

#

u would only use MessageReaction#fetch() if u care about the count prop otherwise u prolly just wanna fetch the message depending on wat ur doing

alpine pier Mar 18, 2020, 3:11 PM

#

I might be stupid but why is TextChannel.members not a GuildMemberManager like Guild.members but a normal Collection instead?

rigid trail Mar 18, 2020, 3:12 PM

#

because its a getter, which means it would have to be a pseudomanager, and it would also be redundant with the guild's member manager, overall its just not really worth it, at least in my opinion

alpine pier Mar 18, 2020, 3:13 PM

#

I see. Thank you meowmeltthumbsup

vernal atlas Mar 18, 2020, 4:05 PM

#

would it make sense that the 'newMessage' in messageUpdate technically couldn't be a partial?

#

listening to the 'raw' event, i see that the full object seems to be sent by the websocket, which is enough to construct a full message right

slate nacelle Mar 18, 2020, 4:07 PM

#

More of an edge edge case:
If you receive an embed update for a partial (or uncached) message, you would only have relevant ids and embeds.

vernal atlas Mar 18, 2020, 4:07 PM

#

ah

real jetty Mar 18, 2020, 7:40 PM

#

yo is there an API for the discordjs.org website that returns info of the docs in JSON, kind of like how NodeJS docs has a "View as JSON" format?

slate nacelle Mar 18, 2020, 7:48 PM

#

There is the docs branch with the relevant json files you could use: https://github.com/discordjs/discord.js/tree/docs

real jetty Mar 18, 2020, 7:57 PM

#

thanks

frozen bridge Mar 18, 2020, 9:55 PM

#

haiya! I was wondering how long it takes issues to be looked at and fixed.

rigid trail Mar 18, 2020, 9:59 PM

#

it really depends on how complex, how big of a change it is, and if there are people who bother to fix it ¯_(ツ)_/¯

copper laurel Mar 18, 2020, 10:11 PM

#

Discussing it here/general (if its a feature/change) or in support (if its not actually an issue with the lib) would speed things along

#

Since it gets more eyes looking at the problem

ornate topaz Mar 19, 2020, 9:05 AM

#

Actually, is adding information about needed permission to setChannel desription (https://github.com/discordjs/discord.js/pull/3959/files#diff-3598194ff2d32819aea8eaf51aeb1b09R157) necessary?
This would make it only method that has it mentioned.

unique axle Mar 19, 2020, 9:42 AM

#

My stand on that: either everywhere or nowhere
In the past we've always come to the conclusion that discord.js should not document upstream permissions, as they weren't even clearly documented to begin with. Discord has now fully documented permissions on the API documentation per endpoint.

However since we recently introduced the permission and hierarchy matters this might have changed.
If we really want to start documenting permissions I'd personally be in favour of adding a link to the used API endpoint rather than just the permissions as part of the method description. If that is even needed/wanted however is another point of discussion

frozen bridge Mar 19, 2020, 4:38 PM

#

@copper laurel It's a big issue thats causing bigger bots not to be able to start:
https://github.com/discordjs/discord.js/issues/3956

#

Staff have seen it

#

just wondering how long it will take to be fixed

#

as I can't keep a shard offline for too long

wild flax Mar 19, 2020, 4:41 PM

#

There is really not enough information to even figure out where its starting or coming from currently

#

Despite the high priority it has

frozen bridge Mar 19, 2020, 4:46 PM

#

Any way I can help ya get info or something

wild flax Mar 19, 2020, 4:49 PM

#

cc @slate nacelle maybe you have a better general idea where this could come from and know which information we'd need

#

cc @tacit crypt

tacit crypt Mar 19, 2020, 4:53 PM

#

I..

For one, can you get me debug logs from the shards, especially those that error?

manager.spawn(20); // Start 4 Shards

This spawns 20 shards, not 4, try changing that?

#

As for errors like Error: Something took too long to do I have NO idea where it's coming from. Check your code to ensure it doesn't have many hanging loops..

slate nacelle Mar 19, 2020, 5:00 PM

#

My thesis is that this is caused by a guild not being available (presence limit reached or discord outage, probably #1).

In detail:

Error: Something took too long to do.
This is a Client (shard) error.

Thrown on v11, once a timeout elapsed before all guilds were ready.
(Nothing you can do about this on v11 sadly.)

That was taken care of with v12 by readying the client after a (different) timeout instead of erroring.
(You'll have unavailable guild(s) then though, but that's probably not an issue, d.js-wise)

Error: Shard 1's Client took too long to become ready.
This is a ShardingManager error.

This timeout can be disabled by setting the spawnTimeout to -1 or Infinity, may not be the fanciest solution, but should work.
https://discord.js.org/#/docs/main/stable/class/ShardingManager?scrollTo=spawn
(Nothing you can do about this on v11 sadly.)

frozen bridge Mar 19, 2020, 5:03 PM

#

@tacit crypt Same issue with 4, used 20 to make sure that only 1/20 of the servers had the bot down

tacit crypt Mar 19, 2020, 5:03 PM

#

Can you check the debug log for each shard?

frozen bridge Mar 19, 2020, 5:03 PM

#

I tried debug, nothing was wrong

#

I can start it with debug and send ya the logs 1 sec

tacit crypt Mar 19, 2020, 5:06 PM

#

Please do

frozen bridge Mar 19, 2020, 5:06 PM

#

The issue persisted in v12

#

Sharding Manager ➜ Launched shard 1
Shard 1 ➜ Using gateway wss://gateway.discord.gg/?v=6&encoding=json
Shard 1 ➜ [ws] [connection] Connecting to wss://gateway.discord.gg/?v=6&encoding=json
Shard 1 ➜ [ws] [connection] Identifying as a new session
Shard 1 ➜ [ws] [connection] Connected to gateway wss://gateway.discord.gg/?v=6&encoding=json
Shard 1 ➜ [ws] [connection] Setting a heartbeat interval for 41250ms

tacit crypt Mar 19, 2020, 5:07 PM

#

thats v11

#

i'm asking for v12

frozen bridge Mar 19, 2020, 5:07 PM

#

o

#

1 sec

#

Hmm it seems to be working now

#

If I get an offline guild ill try again

tacit crypt Mar 19, 2020, 5:18 PM

#

owo 👍

copper laurel Mar 19, 2020, 6:05 PM

#

Sorry, why was I pinged here? Is this relevant to me? I don't understand the sharding at all lol

frozen bridge Mar 19, 2020, 8:02 PM

#

Nope not anymore

#

@tacit crypt Yep it seems to work now, it just takes like a minute longer than the other shards. What could be causing this? (Nothing specifically singles this shard out in the code)

tacit crypt Mar 19, 2020, 8:05 PM

#

Because we wait for all guilds to come in. However, due to unavailable guilds, we wait 15 seconds. Aka

Say you have like 1000 guilds, and 5 are unavailable. We set a timer for 15s to mark the shard as ready regardless if all guilds came. We refresh that timer every time a guild comes. Once all but the last unavailable guilds came, the timeout takes over.

clever crypt Mar 19, 2020, 9:00 PM

#

I keep having problems with one unavailable guild - 7100 members.

Sometimes it doesn't come through until 60 seconds after coming ready, sometimes it's perfectly fine.

fallen arrow Mar 19, 2020, 9:02 PM

#

and this 15s timer is present in v12 but not v11, correct?

#

the solution i came up with is to manually run client.ws.connection.triggerReady(), not sure if its the best workaround

tacit crypt Mar 19, 2020, 9:11 PM

#

I keep having problems with one unavailable guild - 7100 members.

Sometimes it doesn't come through until 60 seconds after coming ready, sometimes it's perfectly fine.
@clever crypt blame discord

warped crater Mar 19, 2020, 10:03 PM

#

^ literally nothing you can do about it

#

after Discord's ready event they begin sending you guilds in the form of GUILD_CREATE packets

#

all d.js does is process them

glass hamlet Mar 20, 2020, 4:45 AM

#

Thanks @fallen arrow , that workaround is working partially for me. Gonna have to work with it for now

lofty birch Mar 20, 2020, 8:58 AM

#

So am I correct in thinking that the ready event is emitted when the bot connects, and the shard ready event is emitted when all of the guilds have been sent?

fallen arrow Mar 20, 2020, 9:18 AM

#

should be the opposite afaik

lofty birch Mar 20, 2020, 10:13 AM

#

Ah ok

real jetty Mar 20, 2020, 1:36 PM

#

https://discord.js.org/#/docs/main/stable/class/User?scrollTo=toString

When concatenated with a string, this automatically returns the user's mention instead of the User object.
That's not true, when doing someString + user, the valueOf method is used, resulting in the userID.
https://github.com/discordjs/discord.js/blob/stable/src/structures/Base.js#L37-L39
Doing somestring ${user} or user.toString() does work as expected, but still, I think the word "concatenated" isn't the right word for it.

old kelp Mar 20, 2020, 3:57 PM

#

I'm having a problem understanding the version 11.6.3 and 12.00

#

like what are they both

#

and as I know, master is the version 11.6.3, right?

frozen bridge Mar 20, 2020, 4:42 PM

#

@wild flax wyd ya reopen?

wild flax Mar 20, 2020, 4:42 PM

#

I commented.

frozen bridge Mar 20, 2020, 4:42 PM

#

o

#

yea just saw

#

One of the devs above thought that there was nothing we could do about it?

#

nope I misread

fallen arrow Mar 20, 2020, 5:22 PM

#

the only thing that could be done about it is to give it the same timer that v12 uses

real jetty Mar 20, 2020, 5:52 PM

#

return { guild };
What is the point in having the above over this:
return guild;

deft holly Mar 20, 2020, 5:54 PM

#

this is #archive-library-discussion 👀

real jetty Mar 20, 2020, 5:55 PM

#

Yea that is a line in the library

deft holly Mar 20, 2020, 5:56 PM

#

that's a question about the library code, but the answer would be it returns an object (shorthand for { guild: guild }), as to why, depends on context

real jetty Mar 20, 2020, 9:17 PM

#

message.channel.bulkDelete(2, true)```
works but return this error

(node:2068) UnhandledPromiseRejectionWarning: DiscordAPIError: 405: Method Not Allowed

#

 
  name: 'DiscordAPIError',
  message: '405: Method Not Allowed',
  path: '/api/v7/channels/%3C',
  code: 0,
  method: 'POST' }```

#

(Taken from #archive-djs-v12-deprecated )

#

Whats weird is from where %3C came from

raven juniper Mar 20, 2020, 9:39 PM

#

Without knowing the context, that sounds like it's in a DM

slate nacelle Mar 20, 2020, 9:44 PM

#

%3C is an encoded <
This method does not even hit that endpoint (POST /channels/:id)
(It impossibly could result in that error)
Bug reports should go to our issue tracker instead
(This does not include enough info though, see point 2)

clever crypt Mar 21, 2020, 2:17 AM

#

Is version 7 even available? https://discordapp.com/developers/docs/reference#api-versioning

vernal atlas Mar 21, 2020, 2:25 AM

#

doesnt seem to throw an error

#

but then again doing v9999 doesnt throw an error either

#

GWrjkKappaLUL

ornate topaz Mar 21, 2020, 2:53 AM

#

some people say v7 is "out" for multiple years now, but discord can't justify releasing it

tacit crypt Mar 21, 2020, 11:51 AM

#

v7 of rest is like v6 of rest but with better errors

vivid field Mar 21, 2020, 3:41 PM

#

Probably a rather nooby question, but why doesn't TextBasedChannel extend a regular Channel?

warped crater Mar 21, 2020, 3:42 PM

#

its an "interface"

#

and it's being implemented into whatever channels are "text based"

tough geode Mar 21, 2020, 3:45 PM

#

/\, it's never constructed, whatever it's implemented into in the lib extends Channel

warped crater Mar 21, 2020, 3:45 PM

#

in short, it's a class but its not used the "normal" way

real jetty Mar 21, 2020, 4:25 PM

#

guild.channels.create(name, options) and guild.roles.create({ data, reason }), I'd say it's better if GuildChannelManager#create would only accept one argument

opaque vessel Mar 21, 2020, 4:39 PM

#

Huh? It already does, the options are optional and the type defaults to a text channel

real jetty Mar 21, 2020, 4:44 PM

#

No I mean like guild.channels.create({ name: 'somevoice', type: 'voice', permissionOverwrites })

#

I'd say it's more consistent with other methods

raven juniper Mar 21, 2020, 4:49 PM

#

The object is for optional parameters, the name is not optional for channels, it is for roles (apparently)

#

You see the same format in emojis.create(), first parameters are required, and then an object for optional ones

real jetty Mar 21, 2020, 4:51 PM

#

Ahh so that's the logic behind it

ornate topaz Mar 21, 2020, 4:54 PM

#

regarding roles, discord defaults to new role name, additionally, it's name every role you create in app had at one point, since you can rename it after you create it

thorny copper Mar 21, 2020, 8:06 PM

#

I tried to update my bot to v12. It's horrible how Typescript's support has become worse than v11

#

Some examples:

#

https://cdn.discordapp.com/attachments/479323925726101514/690988941184073768/unknown.png

#

https://cdn.discordapp.com/attachments/664599917117046794/691011284262453278/unknown.png

clever crypt Mar 21, 2020, 8:20 PM

#

Manager#cache is a collection, which is an extended Map.
Typings for that would be for (const [id, channel] of server.channels.cache).
If you wish to loop over just the channels suffix the .cache access with .values().

thorny copper Mar 21, 2020, 8:20 PM

#

https://cdn.discordapp.com/attachments/664599917117046794/691013481318711296/unknown.png

clever crypt Mar 21, 2020, 8:20 PM

#

This is more of a support question that a complaint/suggestion/nit about the package.

thorny copper Mar 21, 2020, 8:20 PM

#

with .values()

#

I did find a solution tho

#

(server.channels.cache.array() as GuildChannel[])

clever crypt Mar 21, 2020, 8:22 PM

#

I have no problem with this code

for (const channel of <guild>.channels.cache.values()) {
    console.log(channel.permissionOverwrites);
}
``` *edited for your use case

thorny copper Mar 21, 2020, 8:23 PM

#

I don't think there is, but TS grabs it as a "any" type

clever crypt Mar 21, 2020, 8:23 PM

#

More or less a problem with your code editor

thorny copper Mar 21, 2020, 8:23 PM

#

Another fantasy

#

And this one is really from DJS

#

https://cdn.discordapp.com/attachments/484807050766057483/690985999282602034/unknown.png

#

fetch method doesn't take any arg, right?

clever crypt Mar 21, 2020, 8:24 PM

#

Correct

thorny copper Mar 21, 2020, 8:24 PM

#

https://cdn.discordapp.com/attachments/484807050766057483/690986061794639953/unknown.png

umbral aspen Mar 21, 2020, 8:24 PM

#

If you're calling fetch on the message, then yeah.

clever crypt Mar 21, 2020, 8:24 PM

#

What is message?

thorny copper Mar 21, 2020, 8:25 PM

#

client.on("message", (message) =>

umbral aspen Mar 21, 2020, 8:25 PM

#

Also wouldn't #archive-djs-v12-deprecated or #archive-updating-to-v13 be more suitable for this? Thonk

thorny copper Mar 21, 2020, 8:25 PM

#

@umbral aspen Actually I complaint more than I ask for help :x

real jetty Mar 21, 2020, 8:25 PM

#

For the collection loop, it should work fine on es6, how does your tsconfig look?

umbral aspen Mar 21, 2020, 8:26 PM

#

You seem to be asking questions that belong in the variety of subjects discussed / asked (about) in support center.

thorny copper Mar 21, 2020, 8:26 PM

#

Erf, I'm going to #archive-djs-v12-deprecated then

#

Sorry

spice lava Mar 21, 2020, 9:20 PM

#

We have GuildMember#bannable would there be a place for a general method like GuildMember.canban(GuildMember) for example. Or is this too high level same as addBlankField was removed

Edit: or would comparing roles positions by highest roles of each member be the preferred method without a custom implementation

rigid trail Mar 21, 2020, 9:29 PM

#

i really do not see a use for this

#

why dont you just check .bannable on the member Thonk

#

the last thing the lib should do is make wrappers for already kind of iffy permission getters blobsweats

spice lava Mar 21, 2020, 9:30 PM

#

As that would only count for the bot, as said this would be quite high level

rigid trail Mar 21, 2020, 9:30 PM

#

you want to see if a member can ban another? lol

#

again, dont see a use for it

#

its too specific of a method to be justified to be added, especially when it has no direct correlation with things the bot is doing

spice lava Mar 21, 2020, 9:31 PM

#

That's why I asked. Alright!

sly pollen Mar 22, 2020, 5:07 PM

#

Does djs use cluster module for sharding or no?

#

And does it scale to the number of cores available?

opal mauve Mar 22, 2020, 5:08 PM

#

the node cluster module isn't particularly useful for sharding

#

d.js uses normal child processes

sly pollen Mar 22, 2020, 5:15 PM

#

any reason why you guys wanted to go with normal child processes?

#

one thing i noticed about internal sharding is worker threads

#

would a build of several process that have several worker threads(shards) be a bad idea?

#

it's like a mix of internal sharding and process sharding except that you can get the most out of these processes?

opal mauve Mar 22, 2020, 5:17 PM

#

we went with child processes because we didn't need the functionality of the cluster module

#

worker thread based sharding was merged in https://github.com/discordjs/discord.js/pull/2908

sly pollen Mar 22, 2020, 5:20 PM

#

ah i see, but what about a mix between process sharding and threads sharding?

opal mauve Mar 22, 2020, 5:21 PM

#

unsure what advantage that would bring

sly pollen Mar 22, 2020, 5:28 PM

#

Instead of having 10 child processes with their own independent instances and their own memory being shards
We get to have 3 processes with 4 worker threads for each. These worker threads being shards (and every 4 worker threads do share the same memory)
instead of 10 child processes being shards, we get 12 workers being shards and only 3 processes

#

thats what i meant

undone ravine Mar 22, 2020, 6:11 PM

#

@sly pollen workers don't share memory with the main thread

copper laurel Mar 22, 2020, 10:55 PM

#

Does anyone know why the message event shares typings with messageDelete | messageReactionRemoveAll?

public on(
  event: 'message' | 'messageDelete' | 'messageReactionRemoveAll',
  listener: (message: Message | PartialMessage) => void,
): this;

When would client.on("message") emit a PartialMessage?

#

I can't think of a reason, so I feel like it should have its own typings

real jetty Mar 22, 2020, 11:23 PM

#

yh it should just be typed as Message. it never will be a partial within the message event, the author and content fields will always be present there so Message#partial will always be false

frozen bridge Mar 23, 2020, 11:10 PM

#

Haiya! I'm wondering if this is a bug or not... in djs 12.0.2, setting the position of a category to 0 puts it at the bottom of the channels, while setting it to 1 puts it under the channel in position 0

#

im using guild.channels.create

slate nacelle Mar 23, 2020, 11:13 PM

#

The position is passed to discord as-is: https://github.com/discordjs/discord.js/blob/stable/src/managers/GuildChannelManager.js#L121

frozen bridge Mar 23, 2020, 11:28 PM

#

Then why is setting the position to 0 having the category created at the bottom of the channel list?

#

guild.channels.create("Covid-19", {"type": "category", permissionOverwrites: [{id: guild.id, allow: "VIEW_CHANNEL", deny: "CONNECT"}], "position": 0})

rigid trail Mar 23, 2020, 11:40 PM

#

this is not at all the place for that question

copper laurel Mar 23, 2020, 11:40 PM

#

This is not a support channel

rigid trail Mar 25, 2020, 7:56 PM

#

you mean client#voiceStateUpdate?

#

it includes a voice state which has guild info..

#

not the right channel for this

#

please use a support channel

clever crypt Mar 25, 2020, 7:58 PM

#

?gotosupport

rich iglooBOT Mar 25, 2020, 7:58 PM

#

You are in #archive-library-discussion right now. Go to a support channel instead:

rigid trail Mar 25, 2020, 7:58 PM

#

on stable its a VoiceState for both params and on v11 its a guildmember so im not quite sure what you are suggesting

#

anyways its kind of offtopic since at this point its a code issue, not a library issue

copper laurel Mar 25, 2020, 10:44 PM

#

reaction.count isn't showing in docs on 12.0.2 for some reason

#

Possibly because // eslint-disable-next-line eqeqeq has been added between the JSDoc and the property?

slate nacelle Mar 25, 2020, 10:46 PM

#

~~I think there was a PR for that~~
This was fixed with https://github.com/discordjs/discord.js/pull/3980

uneven dawn Mar 26, 2020, 5:31 AM

#

I'm not sure the best practice on this, but instead of having a generic type like GuildChannel that's defined to extend Channel , shouldn't there be a GuildTextChannel that extends TextChannel etc.?

This way you don't have to cast everything that comes as GuildChannel to the text/voice/category/news/store equivalents

rigid trail Mar 26, 2020, 5:42 AM

#

so:

GuildTextChannel ---- TextChannel ---- GuildChannel ---- Channel```? what about DmChannels, those simply extend Channel and implement TextBasedChannel, similar to TextChannel which extends GuildChannel and implements TextBasedChannel

#

you will still have to cast, some methods can return any channel, say grabbing a channel from the GuildChannelManager, it can be text/voice/category, etc, so i fail to see how that solves anything, it seems to just add more clutter to inheritance

copper laurel Mar 26, 2020, 10:22 AM

#

Channel typings will definitely be improved in v13 with proper typescript

snow crypt Mar 26, 2020, 11:54 AM

#

I think messages could implement a guild typeguard

wild flax Mar 26, 2020, 11:54 AM

#

Like the non-working channel typeguard?

snow crypt Mar 26, 2020, 11:55 AM

#

I really don't know, but something to tell TS that if message.guild is present, message.channel is a TextChannel

plucky jay Mar 26, 2020, 11:55 AM

#

I don't think a typeguard like that is possible with TS

#

But maybe a GuildMessage or something?

snow crypt Mar 26, 2020, 11:56 AM

#

was about to say

wild flax Mar 26, 2020, 11:56 AM

#

That falls under what borger said right at the top there

snow crypt Mar 26, 2020, 11:56 AM

#

type GuildMessage = Message & {
  guild: Guild;
  channel: TextChannel;
  member: GuildMember;
};

type DMMessage = Message & {
  guild: null;
  channel: DMChannel; // can be partial(?)
  member: null;
};
``` and rest

#

some methods can return any channel

#

@wild flax why would that fail for message?

wild flax Mar 26, 2020, 11:59 AM

#

oh

snow crypt Mar 26, 2020, 12:01 PM

#

so would that work?

#

also, the channel typeguard works?

#

slate nacelle Mar 26, 2020, 12:04 PM

#

That part works, yes.
But breaks when attempting to use Promise#then on an union of Promises.
https://github.com/discordjs/discord.js/pull/3978

plucky jay Mar 26, 2020, 12:12 PM

#

@snow crypt But if you use message.guild instead, of message.channel.type === 'text', it won't work

snow crypt Mar 26, 2020, 12:12 PM

#

I know, that is why I say something like this could be used https://discordapp.com/channels/222078108977594368/682166281826598932/692703619455254579

#

type GuildMessage = Discord.Message & {
  guild: Discord.Guild;
  channel: Discord.TextChannel;
  member: Discord.GuildMember;
};

type DMMessage = Discord.Message & {
  guild: null;
  channel: Discord.DMChannel; // can be partial(?)
  member: null;
};

type AnyMessage = DMMessage | GuildMessage;

// @ts-ignore
const m : AnyMessage = get();
if (m.guild) {
  m.channel.permissionsFor
}

#

this works

tough geode Mar 29, 2020, 8:48 AM

#

why is the return type of CategoryChannel#children marked as nullable in docs (?Collection<Snowflake, GuildChannel>)? it's not nullable in typings, and it should not be, because it just returns Collection#filter(...), which returns Collection<K, V>, not nullable.
Docs link: https://discord.js.org/#/docs/main/stable/class/CategoryChannel?scrollTo=children

umbral aspen Mar 29, 2020, 8:59 AM

#

I think it should be nullable since a category doesn't necessarily need to have any children channels. In other words, a category can be empty.
This perhaps should be updated to it not being marked as nullable, or updating it so it is nullable and that it returns a collection with entries present in it or null.

tough geode Mar 29, 2020, 9:03 AM

#

categories can be empty, but .children is an empty collection for empty categories, not null, therefore it's misleading in my opinion.

thorny cosmos Mar 29, 2020, 9:06 AM

#

Yes, 1s3k3b is right, the jsdocs for .children is kind of wrong, since collections extend maps, they can give an empty collection object if no key value pair exists, here is what I just evaled on an empty category
Also the typings on Collection states that Collection#filter() returns a Collection, not ?Collection, so it makes sense for why should it not return ?Collection<Snowflake, GuildChannel>

snow crypt Mar 29, 2020, 9:18 AM

#

speaking of CategoryChannel, the setParent method is documented (because of GuildChannel), but it shouldn't be there at all

#

it looks like they tried to document it with a warning, but it doesn't show up https://github.com/discordjs/discord.js/blob/stable/src/structures/CategoryChannel.js#L19

tough geode Mar 29, 2020, 9:19 AM

#

well, it's inherited, but it has a warning

snow crypt Mar 29, 2020, 9:19 AM

#

it doesn't have

#

https://discord.js.org/#/docs/main/master/class/CategoryChannel?scrollTo=setParent

#

same thing for lockPermission

#

It can be removed by just deleting from the child's prototype (https://discordapp.com/channels/222078108977594368/222086648706498562/693752580664459295)

unique axle Mar 29, 2020, 9:32 AM

#

CategoryChannel#setParent should be there because the API technically allows it, despite throwing an error at this moment in time
Hence why instead the warning is introduced

snow crypt Mar 29, 2020, 9:32 AM

#

and what would it do?

unique axle Mar 29, 2020, 9:33 AM

#

as i just said, at this moment in time you can not tier categories like a file system, however the endpoint is there, so we decided to support it

snow crypt Mar 29, 2020, 9:46 AM

#

but it does practically nothing?

unique axle Mar 29, 2020, 9:55 AM

#

throwing an error at this moment in time
would not describe that as "nothing" per se

snow crypt Mar 29, 2020, 10:09 AM

#

i'm talking discord

#

like, you can set a topic to a voice channel with djs, but it really does nothing if you look at it practically

unique axle Mar 29, 2020, 10:11 AM

#

i fail to understand what you mean, the API throws an error, the request is rejected

snow crypt Mar 29, 2020, 10:11 AM

#

the api throws an error when trying to set a parent?

#

does the api throw or is it just the lib?

slate nacelle Mar 29, 2020, 10:15 AM

#

Depends on the error class you catch:
DiscordAPIError -> Discord threw it
DiscordjsError -> discord.js threw it

unique axle Mar 29, 2020, 10:16 AM

#

as said, API error, to be more precise:

DiscordAPIError: Invalid Form Body
parent_id: Categories cannot have subcategories

snow crypt Mar 29, 2020, 10:36 AM

#

well isn't this going with the head in the wall?

#

making a request that is going to fail is very strange to me

wild flax Mar 29, 2020, 10:39 AM

#

its not worth it to specifically handle that use case or somehow remove it from the documentation despite its inheritance

snow crypt Mar 29, 2020, 10:41 AM

#

but the library is there to interact with discord

#

could also split it to have a base GuildChannel, and a ParentedGuildChannel to have setParent on it, then have voice and text inherit from it, and categories would inherit from the base GuildChannel with no setParent

ornate topaz Mar 29, 2020, 2:29 PM

#

but the library is there to interact with discord
Well, it's a wrapper for Discord's API, so theoretically, if CategoryChannel#setParent is documented in API, there is no reason to not have it. It might reject today, but what if tomorrow discord will release subcategories?
It's a wrapper for whole API, not wrapper for some of its functions

copper laurel Mar 29, 2020, 10:53 PM

#

It exists because the API doesn't actually have different endpoints for different channel types the way Discord.js does. There isnt even a "set parent" API endpoint. Its literally just Modify Channel, specifying a parent_id
https://discordapp.com/developers/docs/resources/channel#modify-channel

setParent() is already just a helper function around an edit call really. The most logical place to share this inheritence is GuildChannel otherwise we'd be duplicating code across Text/Voice/News/Store channels

digital hawk Mar 30, 2020, 7:26 AM

#

can someone explain this change? https://github.com/discordjs/discord.js/commit/2be9ebaad2edca24f87f527397d4f2330f107a89#diff-4419bad7f25fd13106e6fe7516df7617

#

it doesnt play well with custom events and my event loader

#

i think having a [event: string]: any[] at the end of ClientEvents could potentially fix it

copper laurel Mar 30, 2020, 8:54 AM

#

Makes sense to me - I think we've effectively locked it down too far

slate nacelle Mar 30, 2020, 8:59 AM

#

You can augment the interface adding your own events there:

declare module 'discord.js' {
    interface ClientEvents {
        foo: [Map<string, any[]>];
    }
}

(Or that catch-all clause, too.)

snow crypt Mar 30, 2020, 9:00 AM

#

I think it should be event: string | symbol in the all-catch

#

now how do i make a pr again

slate nacelle Mar 30, 2020, 9:02 AM

#

This is intentional: https://github.com/discordjs/discord.js/pull/3944#discussion_r393221258

snow crypt Mar 30, 2020, 9:04 AM

#

so I guess we're not adding it back?

slate nacelle Mar 30, 2020, 9:04 AM

#

I'd be in favor to not do that.

snow crypt Mar 30, 2020, 9:05 AM

#

I'll do a pr with that note

#

@slate nacelle https://github.com/discordjs/discord.js/pull/4008

wild flax Mar 30, 2020, 9:32 AM

#

This doesn't make a lot of sense

#

why would you not extend the interface?

#

Why would you want to just allow any symbol or string?

snow crypt Mar 30, 2020, 9:33 AM

#

as written in the pull, i'm just waiting for feedback

wild flax Mar 30, 2020, 9:34 AM

#

You got feedback from 2 devs

snow crypt Mar 30, 2020, 9:34 AM

#

why not more ¯_(ツ)_/¯

vivid field Mar 30, 2020, 12:49 PM

#

Probably more of an api question, but is there a reason why guildMemberRemove gives you a member object whereas guildBanAdd gives you a user object?

rigid trail Mar 30, 2020, 12:49 PM

#

hackbans

#

you can ban a user by id which is not in the guild (see: Guild#members#ban())

snow crypt Mar 30, 2020, 1:17 PM

#

I would not be surprised if member.ban calls member.guild.members.ban

#

aaand... it does

rigid trail Mar 30, 2020, 1:18 PM

#

why would it not

digital hawk Mar 30, 2020, 4:03 PM

#

@slate nacelle I went ahead and just made my own interface in my bot which extends ClientEvents but I guess adding it in a d.ts works too

snow crypt Mar 30, 2020, 6:57 PM

#

so do i close the pr?

#

yep

tender surge Apr 2, 2020, 6:27 AM

#

Wish intellisense would still work if you allowed anything but yea if you allow anything now you either don't get autocompletion or you can't get the typings that are returned from each event

#

Also if you extend now you can easily type what's returned from your custom event instead of asserting

snow crypt Apr 5, 2020, 5:11 PM

#

Would v11 break if moved to node-fetch?

warped crater Apr 5, 2020, 5:27 PM

#

otherwise it would have been done already

tacit crypt Apr 5, 2020, 5:39 PM

#

Would it break? No.
Would it be breaking? Yes

undone ravine Apr 5, 2020, 5:42 PM

#

It wouldn't be breaking

tacit crypt Apr 5, 2020, 5:47 PM

#

breaking from a semver point, no?

rigid trail Apr 5, 2020, 5:49 PM

#

if nothing changes on the public API then i dont see how it would be semver major

undone ravine Apr 5, 2020, 5:49 PM

#

No, unless there are points where raw return values are given to a user from snekfetch/node-fetch

#

Changing dependencies is not a breaking change

slate nacelle Apr 5, 2020, 5:51 PM

#

We expose the raw errors from snekfetch when encountering 5xxs and the retry limit was reached:
https://github.com/discordjs/discord.js/blob/v11/src/client/rest/RequestHandlers/Sequential.js#L76
https://github.com/discordjs/discord.js/blob/v11/src/client/rest/RequestHandlers/Burst.js#L46

snow crypt Apr 5, 2020, 5:53 PM

#

if nothing changes on the public API then i dont see how it would be semver major
that's why i'm asking

rigid trail Apr 5, 2020, 5:56 PM

#

i mean, i believe there is little to be gained from switching to node fetch but a significant amount to lose if done improperly, also, v11 isnt really in development as much as v12, as long as snekfetch works i dont see a need

real jetty Apr 6, 2020, 7:51 PM

#

Would there be any way of adding in where an API call failed in the code? As in, more detailed stack traces

#

So instead of it coming from the library it would come from a method called from code the user had written

rigid trail Apr 6, 2020, 7:54 PM

#

you mean unhandled promise rejections? not really no, as async code is run in parallel to the stack elsewhere, therefore stack traces are hard to make and would be huge

exotic flicker Apr 6, 2020, 7:54 PM

#

v11-dev had a way to show accurate stack traces, but afaik node always had this missery with async stack traces

real jetty Apr 6, 2020, 7:57 PM

#

then what changed between v11-dev and now that required that code to be changed?

#

well, ok probably a lot in all fairness but

slate nacelle Apr 6, 2020, 7:59 PM

#

There is a "round 2" for this in #3443.
A bit of the relevant code, but it's all over the place, really:
https://github.com/discordjs/discord.js/pull/3443/files#diff-b06340a028069dac07f7b735c4e096a5R12
https://github.com/discordjs/discord.js/pull/3443/files#diff-162f9af6d66ef5705d4310712d982f4bR85
Comment linking "round 1":
https://github.com/discordjs/discord.js/pull/3443#issuecomment-523510644

exotic flicker Apr 6, 2020, 8:02 PM

#

discords information on missing permissions / missing access is also lacking (like what permission is missing)
tho this is a feature request
https://github.com/discord/discord-api-docs/issues/1467

tacit crypt Apr 6, 2020, 8:34 PM

#

My implementation in that PR you linked @slate nacelle would only work on v8 sadly (which would be fine if we didn't have browser builds); it's to be changed

slate nacelle Apr 6, 2020, 8:34 PM

#

RIP

#

So basically like appells comment said.

tacit crypt Apr 7, 2020, 12:49 PM

#

Well, yes but actually no

tacit crypt Apr 8, 2020, 3:54 PM

#

TypeError: timeout.refresh is not a function
    at GABClient.handler (C:\Users\GAwesomeBot\Development\node_modules\discord.js\src\stores\GuildMemberStore.js:210:17)
    at GABClient.emit (events.js:323:22)
    at GABClient.EventEmitter.emit (domain.js:482:12)
    at Object.module.exports [as GUILD_MEMBERS_CHUNK] (C:\Users\GAwesomeBot\Development\node_modules\discord.js\src\client\websocket\handlers\GUILD_MEMBERS_CHUNK.js:18:10)
    at WebSocketManager.handlePacket (C:\Users\GAwesomeBot\Development\node_modules\discord.js\src\client\websocket\WebSocketManager.js:391:31)
    at WebSocketShard.onPacket (C:\Users\GAwesomeBot\Development\node_modules\discord.js\src\client\websocket\WebSocketShard.js:437:22)
    at WebSocketShard.onMessage (C:\Users\GAwesomeBot\Development\node_modules\discord.js\src\client\websocket\WebSocketShard.js:295:10)
    at WebSocket.onMessage (C:\Users\GAwesomeBot\Development\node_modules\ws\lib\event-target.js:120:16)
    at WebSocket.emit (events.js:311:20)
    at WebSocket.EventEmitter.emit (domain.js:482:12)``` If anyone wants to have a look, I'm a bit busy ![sad](https://cdn.discordapp.com/emojis/585885410257928194.webp?size=128 "sad")

rigid trail Apr 8, 2020, 3:56 PM

#

GuildMemberStore 🤔

tacit crypt Apr 8, 2020, 3:57 PM

#

yes, i'm not on latest cause we're rewriting soon

rigid trail Apr 8, 2020, 3:58 PM

#

what node version are u running? cause timeout.refresh was added in v10

tacit crypt Apr 8, 2020, 4:49 PM

#

nodejs 12.something

#

OH

#

ik the issue nvm

frozen bridge Apr 8, 2020, 4:57 PM

#

Does djs use the Server Members Intent when using guild.members.cache.get() etc ?

slate nacelle Apr 8, 2020, 4:58 PM

#

No, that's just cache, it does not interact with anything outside of your ram.

frozen bridge Apr 8, 2020, 4:59 PM

#

ah ok thx

lofty birch Apr 8, 2020, 6:26 PM

#

Just want to point out that the api seems to now show public_flags on a user object, you may want to look into that as this seems new

slate nacelle Apr 8, 2020, 6:27 PM

#

Once this gets documented on Discord's end (https://github.com/discord/discord-api-docs), this should be a simple thing to add.

Here we go, a PR: https://github.com/discord/discord-api-docs/pull/1490

tacit crypt Apr 8, 2020, 10:31 PM

#

OWO

#

fucking

#

FINALLY

clever crypt Apr 9, 2020, 2:43 AM

#

Consider yourself PR'd kappa
https://github.com/discordjs/discord.js/pull/4060

keen cosmos Apr 9, 2020, 2:56 AM

#

Doesn't the data come in with a regular fetch? instead of needing a separate utility function couldn't users just use User#fetch

thorny cosmos Apr 9, 2020, 5:01 AM

#

For User#fetch(), you would need the user object to call fetch() on them, and assuming you don't have any user object, in that case what will you do?

rigid trail Apr 9, 2020, 5:01 AM

#

fetch from the users manager

thorny cosmos Apr 9, 2020, 5:02 AM

#

Of course there is a method User#fetch, but it is used on Partial User Objects. Assuming you don't have that also, then how would you use? Hence you need UserManager#fetch(), i.e., a separate function exists for it @keen cosmos

copper laurel Apr 9, 2020, 5:47 AM

#

I dont think that was the point being made

#

Are flags not included in a User payload - do they have a separate endpoint that requires User#fetchFlags() instead of just User#fetch()?

copper laurel Apr 9, 2020, 6:09 AM

#

Tested - not part of the data you get about users when you identify but it is the same fetch endpoint

keen cosmos Apr 9, 2020, 7:03 AM

#

yeah, I meant that the data comes in from the same fetch endpoint

#

@thorny cosmos also you're not restricted to using fetch() on partial options

thorny cosmos Apr 9, 2020, 7:04 AM

#

yes I meant that it can be used on any valid User object, it is mostly used for partial user objects but Ok

ornate topaz Apr 9, 2020, 7:19 AM

#

mostly used does not in any way limit it to that functionality or use case

#

guild.fetch() also exists, but did you ever had a need to fetch a guild? tmk you would use it only if you would at any point need maximumPresences or maximumMembers

real jetty Apr 9, 2020, 9:54 AM

#

@keen cosmos if the User isn't partial then User#fetch() will just return the cached user, so it would never end up patching the user and updating the flags

#

and the reason y Guild#fetch() will always patch the props mentioned above is cus it will always make the api call and call _patch() (https://github.com/discordjs/discord.js/blob/stable/src/structures/Guild.js#L588-L596)

vivid field Apr 9, 2020, 10:04 AM

#

is there a reason why bulkDelete didn't move to the MessageManager?

ornate topaz Apr 9, 2020, 11:25 AM

#

it was never under MessageStore

snow crypt Apr 9, 2020, 2:26 PM

#

i think the question is more like why isn't it on messagemanager?

vivid field Apr 9, 2020, 6:19 PM

#

exactly, with "moving" i meant v11 -> v12 and not old v12 -> new v12

rigid trail Apr 9, 2020, 9:05 PM

#

i just didnt think about moving it when implementing managers, i think its fine and its not a big deal but i do see why you would expect it to be on the manager, unfortunately the switch cannot be made as it is semver major

#

i kept all methods the same from dataStores to managers with the exception of renaming MessageStore#remove to MessageManager#delete for clarity

#

@vivid field

loud moat Apr 9, 2020, 9:15 PM

#

So v13 soon

rigid trail Apr 9, 2020, 9:16 PM

#

i wouldnt exactly say soon lol, its a full rewrite and we dont even have a gateway yet, id expect the basic things like rest/gateway/ some of the core to be done by the end of 2020

undone ravine Apr 9, 2020, 10:19 PM

#

the method could still be moved and leave the old one in place as a shortcut to it and deprecated

rigid trail Apr 9, 2020, 10:19 PM

#

i find this useless personally

undone ravine Apr 9, 2020, 10:20 PM

#

if the method would be more appropriate in the new location, it is absolutely not useless

ornate topaz Apr 9, 2020, 10:23 PM

#

at that point it would be more of a change just for the sake of a change

undone ravine Apr 9, 2020, 10:24 PM

#

no, it'd be a change for the sake of logical consistency

#

is there any functional difference? no, of course not

#

but if that is where that method should be, then that is where it should be

#

no reason to hold it back in an improper place because "meh, it doesn't change anything"

rigid trail Apr 9, 2020, 10:25 PM

#

meh i just think there is no need

undone ravine Apr 9, 2020, 10:25 PM

#

you yourself have already essentially admitted that it makes sense to be on the manager

#

therefore there is a need

rigid trail Apr 9, 2020, 10:25 PM

#

it does, i still dont think the change should be made

undone ravine Apr 9, 2020, 10:26 PM

#

you're making that decision from a purely emotional standpoint with no logic behind it

rigid trail Apr 9, 2020, 10:27 PM

#

for all we know bulkDelete could be renamed/moved somewhere else in djs-next, and managers could not exist either, i think deprecating it is not really the right move

undone ravine Apr 9, 2020, 10:27 PM

#

deprecation doesn't guarantee removal

#

it is not a commitment

ornate topaz Apr 9, 2020, 10:28 PM

#

but if it's deprecated, devs should start to look into new way, which literally does not exist now

undone ravine Apr 9, 2020, 10:29 PM

#

the new way is the correct location for the method

#

which would exist at the time of deprecation

raven juniper Apr 11, 2020, 7:44 PM

#

@real jetty that sort of question belongs in #archive-commando-deprecated not in this channel

#

commando is a framework for d.js it's not part of the library

rigid trail Apr 11, 2020, 9:00 PM

#

you might want to read #3485 to get an idea of why things changed, its mostly to separate the API and cache methods. im not quite sure what you mean added to the client class, would you mind elaborating on that?

#

there are no new classes by the way, they only replaced the data store classes

#

while yes, they do break code significantly and many users may not see the need, which is unfortunate, the changes greatly benefitted the maintainability of the library for the future

#

and where would the api methods go? that would be the same as datastores but with a different name

#

the point of managers was to separate the cache and api, you cant quite do that without breaking users code

#

but what you are describing are data stores

wild flax Apr 11, 2020, 9:04 PM

#

though maybe actually extending the class into type-specific subclasses (guildcollection, channelcollection, etc) would have been a better move to avoid breaking existing code
this is essentially what was happening with stores and why we moved away from it

rigid trail Apr 11, 2020, 9:05 PM

#

this would still have all of the downsides of datastores

rigid trail Apr 11, 2020, 9:24 PM

#

add is an internal method to add to the cache and remove is an artifact with a pr to remove it from typings

#

yes

#

the GuildMemberRoleManager is a pseudomanager, its add/set methods call the api

#

no problem

slate nacelle Apr 12, 2020, 4:23 PM

#

Would this work?

#

Looks good, on a first glance.

#

sure

topaz mango Apr 12, 2020, 4:35 PM

#

Don't think it's a breaking change.

rigid trail Apr 12, 2020, 4:36 PM

#

This is a breaking change tho real_eyes

#

at least thats what it looks like to me, im not sure if private things like the data resolver count

snow crypt Apr 12, 2020, 4:40 PM

#

it is like changing the rest manager

#

you're not supposed to touch it

#

hold up, wouldn't it break things that call if it they are not also changed?

#

because in resolveBase64 it still checks for buffer

rigid trail Apr 12, 2020, 4:52 PM

#

it seems to me that this isnt a clean change and that it could silently affect other things that call it, and that more tests need to be written

rigid trail Apr 12, 2020, 5:19 PM

#

Wouldnt that just nullify the point of the pr monkaHMM

opal mauve Apr 12, 2020, 5:20 PM

#

technically, APIMessage doesn't document files as anything other than "Object", so I don't think this would qualify as breaking.

#

honestly a bit of an oversight in documentation, but i think it would serve our purposes here if we wanted to make this change

lofty birch Apr 13, 2020, 11:23 AM

#

I do think a separate method for uploading local files as streams would be helpful though, I do quite a bit of sending images and my ram usage jumps up when doing so

warped crater Apr 13, 2020, 11:23 AM

#

Doesn't this entirely contradict past changes, however

#

with time all of the send* (e.g. sendEmbed, sendFile) methods were removed in favor of the universal send

snow crypt Apr 13, 2020, 12:02 PM

#

@drowsy niche you can still make separate DataResolver methods

#

like, a different resolving for local / remote that will then be called accordingly

snow crypt Apr 13, 2020, 3:47 PM

#

Regarding patterns in MessageMentions, can't IDs be 16-character long?

ornate topaz Apr 13, 2020, 4:30 PM

#

considering they are being somewhat increased, and that Discord was released in may, no snowflake from after may can be 16 characters long

snow crypt Apr 13, 2020, 4:31 PM

#

so it will be 17,19 for now, and 17,20 after there can't be no more 19digit ids?

ornate topaz Apr 13, 2020, 4:34 PM

#

yes

#

we're already at 699296380694429736

#

so theoretically

wild flax Apr 14, 2020, 12:25 PM

#

What of all of this do you actually send to discord?

keen cosmos Apr 14, 2020, 7:47 PM

#

@drowsy niche I messed around with your test code and it seems like the files property of the json response from httpbin is being reset when sending a stream, maybe that's the issue?

tribal slate Apr 15, 2020, 2:58 AM

#

setSelfDeaf literally does not work

#

Not sure if I'm doing something wrong or if its because of the library itself

#

It shouldn't return anything

wispy harbor Apr 15, 2020, 2:59 AM

#

(actually it should)

tribal slate Apr 15, 2020, 2:59 AM

#

Yeah but like

#

yikers

rigid trail Apr 15, 2020, 3:00 AM

#

the return type is documented so im not sure what the problem is

tribal slate Apr 15, 2020, 3:00 AM

#

Its supposed to self deafen the bot

#

But its not doing that, and just returning false

rigid trail Apr 15, 2020, 3:01 AM

#

well, thats not something for this channel

tribal slate Apr 15, 2020, 3:01 AM

#

I thought this was a problem with the library

#

Which is why I posted here

#

Is it a problem with the docs? I'm really not sure

snow crypt Apr 15, 2020, 8:16 AM

#

not sure what should I do now tho, the form-data repo is pretty dead
@drowsy niche create @discordjs/form-data?

tacit crypt Apr 15, 2020, 2:06 PM

#

@drowsy niche ```ts
declare const body: import('node-fetch').Body;

body.body```

#

Per their docs, this is totally valid

#

Since res.body is a Body class, it has a body property (stupid, I know)

#

Do you think that might fix it? (so we don't have the issue where form-data is..borked)

#

In this case, body.body is a NodeJS Readable Stream, instead of a WHATWG ReadableStream

#

Wait the hell?

#

Am I seeing this wrong

#

Oh I totally did see this wrong, fricking docs

#

Well this might explain why it doesn't work without that form-data PR

tacit crypt Apr 15, 2020, 2:30 PM

#

oh, gotcha

keen cosmos Apr 16, 2020, 6:57 AM

#

@drowsy niche so the issue is that form-data doesn't like anything that's not a regular Readable stream, right?

wild flax Apr 16, 2020, 10:13 AM

#

@drowsy niche if you want we can probably keep a fork and keep it up to date with upstream

wild flax Apr 16, 2020, 10:47 AM

#

well, I'd rather have a fork with the PR merged (in our fork) than rely on someone else fork, if you know what I mean

#

yes, but we have to fork form-data first is what im saying

#

have a fork that is under our control, in case the original PR author deletes his or whatever

#

probably

#

anything is better than relying on some 3rd party fork here.

#

Or installing a commit from a PR

wild flax Apr 16, 2020, 11:07 AM

#

published under @discordjs/form-data

#

3.0.1

#

youll have to make it a non-draft

#

👍

fallen arrow Apr 16, 2020, 11:45 AM

#

Are partial guildChannels supported? This throws errors in the ChannelManager if the channel is not cached but the guild is, i was wondering if it would make sense to provide a partial channel or to simply return if either of them are not available instead: ```js
class InviteCreateAction extends Action {
handle(data) {
const client = this.client;
const channel = client.channels.cache.get(data.channel_id);
const guild = client.guilds.cache.get(data.guild_id);
if (!channel && !guild) return false;

const inviteData = Object.assign(data, { channel, guild });
const invite = new Invite(client, inviteData);

warped crater Apr 16, 2020, 12:44 PM

#

this doesn't seem quite right

#

normally all guilds should be cached when they are received via the GUILD_CREATE packet, which provides all channels

#

there should never be a case where a channel is not cached

#

same with roles & guilds

fallen arrow Apr 16, 2020, 12:53 PM

#

yes i am aware of that, the only use case for this is when the developer uncaches channels to save memory, which is more common than one would think

#

hence why the question if it would make sense to support such use cases since we already have partials in many other scenarios, or if this is out of the scope of the library

warped crater Apr 16, 2020, 1:00 PM

#

as stated in other places & conversations related to cache, going by

uncaches channels to save memory, which is more common than one would think
would cause a lot of changes and checks all over the place for no particular reason

The client options provide various ways of keeping cache smaller for things that are not needed by the core library.

A pretty good summary of it can be seen here: https://github.com/discordjs/discord.js/pull/3995#issuecomment-604441660

#

Channels are data discord.js absolutely depends on internally at the moment, not just in the situation that you pointed out just now

On top of that, I honestly doubt this is "more common than one would think", given that channels are the least of your problems when thinking about cache.

The only real place you can save actual valuable memory in is the messages cache & potentially the members.

wild flax Apr 16, 2020, 1:05 PM

#

We currently only really use or support partials whenever discord gives us partials. And channels dont fall under that category

fallen arrow Apr 16, 2020, 1:08 PM

#

alright thanks

frozen bridge Apr 18, 2020, 2:58 PM

#

@wild flax not sure if this is ping important but people have been having the issue on v12. Is there a fix on that version as well? https://github.com/discordjs/discord.js/issues/3956

wild flax Apr 18, 2020, 2:59 PM

#

@tacit crypt

#

do we not trigger ready even if not all guilds are ready?

#

Im pretty sure we do, or do we only do that on internal sharding?

tacit crypt Apr 18, 2020, 3:00 PM

#

we do, but the timeout waits 15 seconds before doing that

wild flax Apr 18, 2020, 3:00 PM

#

ok but what if you use the sharding manager

tacit crypt Apr 18, 2020, 3:00 PM

#

..same?

wild flax Apr 18, 2020, 3:00 PM

#

doesnt look like it

#

lol

tacit crypt Apr 18, 2020, 3:00 PM

#

The default sharding manager timeout is probably too small by default

#

I mean the cycle is:

Bot gets Discord READY, we set a 15s timeout
For each GUILD_CREATE, we refresh the timeout
If there's still unavailable guilds, the 15s timeout will trigger, marking the client as READY

#

Sooo

#

TL;DR: increase your sharding manager timeout and it should be fine

lofty birch Apr 18, 2020, 3:04 PM

#

Using a sharding manager I get the error it took to long to ready sometimes

tacit crypt Apr 18, 2020, 3:04 PM

#

..As I said

#

Increase the timeout

lofty birch Apr 18, 2020, 3:05 PM

#

According to what you said it should just mark as ready after the timeout with the remaining guilds being unavailable

tacit crypt Apr 18, 2020, 3:05 PM

#

Yes, but the default timeout is too short in the manager

#

Speaking of, @lofty birch do you have any of the optional deps installed? (zlib-sync, bufferutil & utf-8-validate)?

lofty birch Apr 18, 2020, 3:07 PM

#

I don't believe I do

tacit crypt Apr 18, 2020, 3:08 PM

#

Can you try installing those and seeing if that helps?

#

Otherwise, try increasing the timeout on the shardingmanager to say, a minute

#

(manager.spawn(undefined, undefined, 60000)

lofty birch Apr 18, 2020, 3:12 PM

#

Will do in a bit

frozen bridge Apr 19, 2020, 12:18 AM

#

Will User#Flags ever be supported in v11?

copper laurel Apr 19, 2020, 1:08 AM

#

No, v11 will not receive new features

snow crypt Apr 19, 2020, 4:41 PM

#

https://github.com/discordjs/discord.js/blob/master/src/structures/Guild.js#L468
get nameAcronym() {
  return this.name.replace(/\w+/g, name => name[0]).replace(/\s/g, '');
}

Where does the regex come from?

#

I am asking as it gives incorrect results, and I can't find the source for it.

ornate topaz Apr 19, 2020, 5:23 PM

#

probably from nowhere, as acronym is generated by client from usually first letters

snow crypt Apr 19, 2020, 5:29 PM

#

so the acronym-ing is done somewhere in the client js?

ornate topaz Apr 19, 2020, 5:30 PM

#

by client i meant the discord app

wild flax Apr 19, 2020, 5:30 PM

#

Yeah its not an API thing

snow crypt Apr 19, 2020, 5:32 PM

#

by client i meant the discord app
that's what i meant

#

like, in the js of the app

#

there https://raw.githubusercontent.com/DJScias/Discord-Datamining/master/2020/2020-04-18/40abc31adc473eefed44.js

#

@opaque robin that's not how you deprecate

#

And you don't resolve the promise

opaque robin Apr 19, 2020, 5:34 PM

#

@snow crypt tbh I searched the server for util.deprecate and afaik from what I've seen, it returns the function with the deprecation warning.
But if you say I'm incorrect, I'll take that

snow crypt Apr 19, 2020, 5:35 PM

#

Currently the method returns a function

opaque robin Apr 19, 2020, 5:36 PM

#

Oh, so I should be calling the function it returns

slate nacelle Apr 19, 2020, 5:36 PM

#

This is what we did in v11: https://github.com/discordjs/discord.js/blob/v11/src/structures/Guild.js#L1693

snow crypt Apr 19, 2020, 5:36 PM

#

was about to say, do it at the bottom of the file with the prototype

opaque robin Apr 19, 2020, 5:36 PM

#

Ah

#

Thanks space,

#

that makes more sense

snow crypt Apr 19, 2020, 5:37 PM

#

yeah, thanks, space 😞

opaque robin Apr 19, 2020, 5:38 PM

#

I should still be using const util = require('util') instead of directly requiring in the prototype I presume.

snow crypt Apr 19, 2020, 5:38 PM

#

or destructure deprecate from it

opaque robin Apr 19, 2020, 5:44 PM

#

Alright 👍

#

@snow crypt Wouldn't the async await mean it wouldn't be returning a promise anymore

snow crypt Apr 19, 2020, 5:49 PM

#

nope, just the opposite

opaque robin Apr 19, 2020, 5:52 PM

#

ah right, mb

#

@vernal atlas What would be the more appropriate tag to use for that?

I think <info> would be appropriate, not too sure.

snow crypt Apr 19, 2020, 5:55 PM

#

No, just an actual description of the typedef

opaque robin Apr 19, 2020, 5:56 PM

#

  /**
   * An object containing information about a guild's vanity url.
   * @typedef {Object} VanityData
   * @property {?string} code Vanity URL invite code, not the full url
   * @property {?number} uses How many times this invite has been used, must be fetched at least once
   */

snow crypt Apr 19, 2020, 5:56 PM

#

or just vanity, remove url

#

or data

vernal atlas Apr 19, 2020, 5:59 PM

#

not sure what would be better (as to info or warn tag)

snow crypt Apr 19, 2020, 6:00 PM

#

i think just vanity is enough

#

if someone would want to correct it later it's an ez pr

opaque robin Apr 19, 2020, 6:01 PM

#

Might as well get everything done in the one PR tbh

#

so @typedef {Object} Vanity

#

remove URL in Vanity URL invite code

#

and that should be good yeah?

snow crypt Apr 19, 2020, 6:27 PM

#

by client i meant the discord app
do you think ```js
t.getAcronym = function(e) {
return null != e ? e.replace(/'s /g, " ").replace(/\w+/g, (function(e) {
return e[0]
})).replace(/\s/g, "") : ""
}

#

I think it is

#

so yea, it should be name.replace(/'s /g, " ").replace(/\w+/g, e => e[0]).replace(/\s/g, "")

snow crypt Apr 19, 2020, 6:50 PM

#

pr'd it

snow crypt Apr 19, 2020, 7:51 PM

#

What prevents #2910 from being merged?

deft holly Apr 19, 2020, 7:55 PM

#

I'd like to know about ^ as well, left a comment two weeks ago and haven't gotten a response :P

snow crypt Apr 20, 2020, 6:37 AM

#

@clever crypt I found that piece of code with the raw website, didn't need to actually download it

#

looked for "acronym", found uses of "getAcronym", searched its definition and got the actual one

copper laurel Apr 20, 2020, 7:40 AM

#

User#flags doesn't seem to be documented, not sure if it's missing or the docgen for 12.2.0 didn't work properly or something

#

I'm mobile so not checking or PRing myself

snow crypt Apr 20, 2020, 7:40 AM

#

on master it's there

#

https://discord.js.org/#/docs/main/master/class/User?scrollTo=flags

#

oh, you're correct

#

a PR to document it was merged after 12.2.0

#

copper laurel Apr 20, 2020, 7:47 AM

#

Ahh it's after, all good

#

In mobile and didn't check

vapid echo Apr 20, 2020, 12:07 PM

#

hey, any idea how long it takes for v12 bug fixes to get ported to v11? please let me know if this is wrong channel for this

#

i'm specifically looking for this bugfix: fix up WebSocketShard errors. it was released with v12.2.0 a few days back

lofty birch Apr 20, 2020, 5:57 PM

#

new api feature: https://github.com/discord/discord-api-docs/pull/1528

#

was announced in discord dev server also

ornate root Apr 20, 2020, 7:18 PM

#

@vivid field but client.channels gives you a ChannelManager

#

and you can use .fetch on it

vivid field Apr 20, 2020, 7:19 PM

#

oh wait, i thought they used Guild#channels

#

also unrelated to this channel

ornate root Apr 20, 2020, 7:21 PM

#

it was about djs

copper laurel Apr 20, 2020, 10:21 PM

#

@vapid echo v11 is no longer maintained and will not receive further non-critical fixes. v12 is the stable branch which will receive maintenance patches and releases

vapid echo Apr 21, 2020, 4:10 AM

#

Oh. I still see v11 releases though. Are you saying v11.6.4 (released 15 days ago) was the last one? @copper laurel

rigid trail Apr 21, 2020, 4:11 AM

#

if there are significant bugs which can be fixed pretty easily then a quick patch to v11 may be done, but feature updates are out the window

vapid echo Apr 21, 2020, 4:12 AM

#

Right, I am looking for a bugfix. Like I mentioned above, specifically this one: fix up WebSocketShard errors. it was released with v12.2.0 a few days back

#

Just wondering if it will make it to v11

copper laurel Apr 21, 2020, 4:12 AM

#

Can you please reference the PR you're referring to

vapid echo Apr 21, 2020, 4:12 AM

#

Sure

#

https://github.com/discordjs/discord.js/pull/3722

rigid trail Apr 21, 2020, 4:13 AM

#

iirc that was a pretty big change, and the websocket from v11 to v12 has changed a lot, so fixes on v12 can completely break v11

copper laurel Apr 21, 2020, 4:13 AM

#

Its highly unlikely because the websocket was signficantly changed in the v12 release

#

So these changes wouldnt even apply to v11

rigid trail Apr 21, 2020, 4:13 AM

#

yeah its pretty significant

vapid echo Apr 21, 2020, 4:13 AM

#

Oh

#

Okay, we can explore the possibility of moving to v12 then

proven heron Apr 21, 2020, 7:04 AM

#

I feel like displayAvatarURL’s dynamic key should be true by default. It seems like it would be more beneficial as more people will probably be after the gif instead of the webp

wild flax Apr 21, 2020, 7:06 AM

#

Sadly thats a breaking change and wont happen

proven heron Apr 21, 2020, 7:08 AM

#

Explain breaking change. Almost everything you change will break pre-existing code?

wild flax Apr 21, 2020, 7:16 AM

#

https://semver.org/

#

Changing a default to be something else falls under MAJOR

fickle flare Apr 22, 2020, 2:36 AM

#

Has anybody reported any issues with this promise hanging?

#

https://github.com/discordjs/discord.js/blob/master/src/client/websocket/WebSocketShard.js#L176

#

It doesn't reject, it doesn't resolve... just dies 😫

tough geode Apr 22, 2020, 6:11 AM

#

it adds listeners which either resolve or reject the Promise

real jetty Apr 22, 2020, 9:40 AM

#

I was looking for a library to write a Bot in and when looking at the discord.js Github, i've found this: https://github.com/discordjs/discord.js-next, what should anyone expect from it? Will any bot written in the current Discord.js work in next or would i have to write the my Bot again?

wild flax Apr 22, 2020, 9:41 AM

#

The next version, no, yes

real jetty Apr 22, 2020, 9:43 AM

#

I know it seems to be in a pretty early stage to say anything but you have any ETA planned? I wonder how long i could expect my bot to work using the current version.

#

Or you plan to support the current version in some sort of LTS?

spice lava Apr 22, 2020, 1:05 PM

#

So I found a typings issues. With the following code: ts msg.channel.delete().catch() where channel is TextChannel | DMChannel | NewsChannel fails with ```
TS2349: This expression is not callable.
Each member of the union type '(<TResult = never>(onrejected?: (reason: any) => TResult | PromiseLike<TResult>) => Promise<DMChannel | TResult>) | (<TResult = never>(onrejected?: (reason: any) => TResult | PromiseLike<...>) => Promise<...>) | (<TResult = never>(onrejected?: (reason: any) => TResult | PromiseLike<...>) => Promise<...>)' has signatures, but none of those signatures are compatible with each other.


Seems to be an issue introduced with v12 as the Channel typings changed from `public delete(): Promise<Channel>;` to `public delete(reason?: string): Promise<this>;`

Granted this would be an easy fix if you cast a type like `(msg.channel as TextChannel).delete().catch()` but I thought that this issue shouldn't appear in the first place. On the other hand I am not sure on if this should even be fixed as typings in general would be fucked if you just return `Promise<Channel>` there as this wouldnt be very detailed. 

Introducing types on the top level classes like `TextChannel` would introduce the same issue. So I would like to know some comments on this one as I would say all approaches are not perfect because there are actually cases where `channel` could be both, `DMChannel` and `TextChannel`

vernal atlas Apr 22, 2020, 1:26 PM

#

interesting

#

there was something similar to this with the channel union types right?

spice lava Apr 22, 2020, 1:48 PM

#

The issue is the same with channel.fetch().catch() btw, exact same reason as above

slate nacelle Apr 22, 2020, 2:32 PM

#

This is the same issue as the union types, I mean it is an union type after all.

spice lava Apr 22, 2020, 2:33 PM

#

So this is a known issue/problem?

slate nacelle Apr 22, 2020, 2:35 PM

#

Not in this context, but I guess?
I don't think we can do much about this except using Channel instead of the union and let you assert / instanceof concrete types.

spice lava Apr 22, 2020, 2:35 PM

#

That's what I thought. Is there already a PR for it? (will look myself but maybe you know more)

slate nacelle Apr 22, 2020, 2:35 PM

#

I don't think so.

spice lava Apr 22, 2020, 2:36 PM

#

time to get into 2+ 2- PRs again

tacit crypt Apr 22, 2020, 4:23 PM

#

It doesn't reject, it doesn't resolve... just dies 😫
@fickle flare thats..impossible, theres always at least ONE event that runs

fickle flare Apr 22, 2020, 4:25 PM

#

What can I do to help diagnose the problem?

#

or track the source

#

I'm not using standard Node, but until v11 it was working flawlessly

tacit crypt Apr 22, 2020, 4:26 PM

#

Wait, what

fickle flare Apr 22, 2020, 4:44 PM

#

Yes, Its running from inside FiveM's compilation of Node

#

I know there was some history and people here didn't like it, but they actually solved the "window" global parameter pollution and it was working as it should

#

.... until I tried to update to v12 yesterday 😢

vernal atlas Apr 23, 2020, 6:12 PM

#

the data.embed_enabled here is deprecated, in favour of data.widget_enabled, im going to do a PR, assuming it would it be better to deprecate Guild#embedEnabled, in favour of Guild#widgetEnabled, is deprecating a property as simple as just adding the @deprecated tag?

#

https://github.com/discordjs/discord.js/blob/67a74c33e1b124d609bee65f82656e7bf5b09764/src/structures/Guild.js#L208

#

(relevent API docs PR to show its deprecated, https://github.com/discord/discord-api-docs/pull/1536/files)

copper laurel Apr 23, 2020, 8:34 PM

#

You use util.deprecate iirc, @deprecate only JSDocs it

tacit crypt Apr 23, 2020, 11:55 PM

#

I'm pretty sure that's only for functions

snow crypt Apr 24, 2020, 5:48 PM

#

no

#

also props

#

Permissions#_member on v11

opaque vessel Apr 24, 2020, 5:49 PM

#

Is this a bug in the client? Whenever I deleted an uncached message but listening to it as partial, the .system property is always true, but that's not correct?

real jetty Apr 24, 2020, 7:33 PM

#

Suggestion to add field inline limiter, so you can use two values each line if you want to

#

That's quite easy to implement, but really not a feature for D.JS imo

#

Especially since .addBlankField has also been removed

real jetty Apr 24, 2020, 8:14 PM

#

smugshh

ornate topaz Apr 24, 2020, 8:15 PM

#

addBlankField was just a shortcut for addField('\u200b', '\u200b'), that's why it was removed. You can still do it on your own, adding that as 3rd inline field will "fill" the 3rd inline space

However, that will render badly on devices that do not have ability to render any form of inline fields inline. Additionally, embed rendering is purely up to discord - as example can be not-that-recent embed update that allowed to have 3 inline fields and a thumbnail.

vernal atlas Apr 25, 2020, 3:02 AM

#

so i was actually looking around trying to figure out why user#flags can be undefined, found this, kinda odd? how the API sends it but its not picked up?

fiery geyser Apr 25, 2020, 8:05 AM

#

Seems that the API only gives information about public_flags for some users on startup for some reason, manually fetching the user will return the property
I don't think the library can do anything about this
https://github.com/discord/discord-api-docs/pull/1490#issuecomment-613546566

snow crypt Apr 25, 2020, 8:23 AM

#

There is still User#fetchFlags

vernal atlas Apr 25, 2020, 1:24 PM

#

ah, assumed that it was just being lost somewhere

vernal atlas Apr 26, 2020, 1:28 AM

#

Message#suppressEmbeds seems to be broken when using the allowedMentions client option
(suppressEmbeds calls Message#edit, which uses APIMessage.resolveData and adds in the allowedMentions default option)

would the best way be to just call the endpoint directly in suppressEmbeds rather than calling this.edit?

i feel like theres a better way but maybe thats just because its 2:30am and i haven't slept GWrjkKappaLUL

copper laurel Apr 26, 2020, 1:39 AM

#

broken in that it will unintentionally also edit the allowedMentions?

tacit crypt Apr 26, 2020, 2:07 AM

#

Reminder: message edits do not ping the user/role(s)

snow crypt Apr 26, 2020, 12:39 PM

#

would ids be bigints in djs@next?

wild flax Apr 26, 2020, 12:45 PM

#

possibly, not much is set in stone yet for djs 13

vernal atlas Apr 26, 2020, 12:55 PM

#

broken as in the API returns an error (if said message is authored by another user)

(i thought i mentioned that but must've slipped my mind lul) @copper laurel

copper laurel Apr 26, 2020, 1:07 PM

#

Ahh right, makes sense

#

As for BigInt - why? I don't see any value in them being handled as integers, strings works perfectly well.

wild flax Apr 26, 2020, 1:10 PM

#

tbf most sensible languages handle them as a numbers

#

there isnt any particular reason why we couldnt

tacit crypt Apr 26, 2020, 3:46 PM

#

Except we have the constant conversion between BigInt<->String for REST

timid pewter Apr 26, 2020, 4:58 PM

#

Not sure if I should make a new issue for this so I am just gonna try asking here first.

https://github.com/discordjs/discord.js/pull/4087#issuecomment-619242608

I didn't get a response from any of the lib devs on this besides a reaction which doesn't help clarify to me what you guys plan on doing for this? Should TS developers expect breaking changes without warning in the future on what should be a stable branch?

surreal geode Apr 26, 2020, 5:37 PM

#

No, updating typings is not a breaking change

timid pewter Apr 26, 2020, 5:40 PM

#

if it breaks projects it is a breaking change, and this broke projects aka breaking change

#

If DJS doesn't want to properly support TS devs then thats different, but any TS dev using TS properly the way it was intended broke from this PR

#

All im asking is to just give a proper bump to the semver, im fine with the change ill make the changes to my code but dont implement a change that can break projects in a minor bump plz

#

npm i on the vps and projects breaks RIP!

surreal geode Apr 26, 2020, 5:42 PM

#

You don't bump major because you want to, you bump it because it should be bumped. Typings are non-code changes

timid pewter Apr 26, 2020, 5:43 PM

#

When a PR breaks a project it should be bumped

#

the question is whether or not DJS is wanting to properly support TS developers

#

If DJS wants to properly support TS, then it should enforce proper warnings of breaking changes that DO break TS projects. If it doesnt then thats the devs choice and right to choose and myself to live with

#

But come one folks, this is a pretty simple ask to not introducing breaking changes for TS devs in a non-major bump 🙏

warped crater Apr 26, 2020, 5:48 PM

#

This is not how things work

surreal geode Apr 26, 2020, 5:48 PM

#

This is not a breaking change. If you're extending the functionality of D.js, that's on you

warped crater Apr 26, 2020, 5:48 PM

#

You don't bump the major version because X gets messed up by it

#

I highly advise you read this webpage: https://semver.org/

timid pewter Apr 26, 2020, 5:48 PM

#

a major version is bumped when an incompatible api change is made

#

🎉

warped crater Apr 26, 2020, 5:49 PM

#

correct, yet the API wasn't changed.

timid pewter Apr 26, 2020, 5:49 PM

#

this is incompatible with the previous commit aka breaking

warped crater Apr 26, 2020, 5:49 PM

#

just typings behaivour

timid pewter Apr 26, 2020, 5:49 PM

#

yes it was

warped crater Apr 26, 2020, 5:49 PM

#

this does not fall under semver major

surreal geode Apr 26, 2020, 5:49 PM

#

@timid pewter How was this a breaking change

warped crater Apr 26, 2020, 5:49 PM

#

no parameters were moved/removed, no methods were removed

timid pewter Apr 26, 2020, 5:49 PM

#

The API was clearly changed, it prevents a lot of behavior

#

it prevents emitting events that arent listed in those types

#

anyone that was using that behavior which was supported by DJS broke

ornate topaz Apr 26, 2020, 5:50 PM

#

Api didn't change. It was typings file that changed, it's not exactly part of library source code

warped crater Apr 26, 2020, 5:50 PM

#

this, typings changes alone will never fall under anything that isnt semver patch

timid pewter Apr 26, 2020, 5:50 PM

#

the typings are what provide the api to ts

surreal geode Apr 26, 2020, 5:50 PM

#

It doesn't prevent anything. tsc now tells you types don't match

timid pewter Apr 26, 2020, 5:50 PM

#

aka the api did change for ts devs

#

It doesn't prevent anything. tsc now tells you types don't match
No it prevents it. Bots wont start

warped crater Apr 26, 2020, 5:51 PM

#

once again, this is not how semver works, also see what Daro said

timid pewter Apr 26, 2020, 5:51 PM

#

TBH daro i was just hoping for a answer from a lib dev and then ill be on my way

ornate topaz Apr 26, 2020, 5:51 PM

#

You can launch compiled code anyway, so nothing broke

timid pewter Apr 26, 2020, 5:52 PM

#

I don't think you guys understand how working on a TS project works and how massively this PR effected end users

strange igloo Apr 26, 2020, 6:37 PM

#

Hey, I'm implementing support for chunk_count, chunk_index and nonce properties in GUILD_MEMBERS_CHUNK on discord.js. In the event emit, should I better put count and index into an object or leave them as two separate parameters?
Like... (members, guild, { count, index }, nonce) or (members, guild, count, index, nonce) ?

slate nacelle Apr 26, 2020, 6:40 PM

#

I think one single object would be good, in case they add more props.

wild flax Apr 26, 2020, 6:40 PM

#

^

#

id also be on board with that

#

generally more in support of objects in all cases almost

strange igloo Apr 26, 2020, 6:41 PM

#

One single object containing the three properties so?

wild flax Apr 26, 2020, 6:41 PM

#

yup

strange igloo Apr 26, 2020, 6:41 PM

#

Fine 👌

snow crypt Apr 26, 2020, 7:45 PM

#

as for the bigints, i just wanted to know if they'll be used as they exist now, not insisting they should / must be used, but just showing the possibility

rigid trail Apr 26, 2020, 7:45 PM

#

for snowflakes? there is no real reason to

#

the library has functions for extracting data out of snowflakes without using a bigint, and dapi gives snowflakes as strings

snow crypt Apr 26, 2020, 7:52 PM

#

@rigid trail read above (like, really)

#

https://discordapp.com/channels/222078108977594368/682166281826598932/703948421614272523

#

that conversion everytime seems unnecessary when thinking about it, but wouldn't it actually fit it better?

rigid trail Apr 26, 2020, 7:54 PM

#

i agree with monbrey in saying its a useless conversion, most users dont have to do any integer things on ids, and if they do its an easy construction of a bigint

snow crypt Apr 26, 2020, 7:54 PM

#

this is true

#

i also see no need to do math on ids

#

but the data type exists now, why not use it?

rigid trail Apr 26, 2020, 7:55 PM

#

just because something exists doesnt mean its a good idea to use it

strange igloo Apr 26, 2020, 7:55 PM

#

BigInt may also confuse begineers (it's already complicated sometimes as strings so I don't imagine...)

rigid trail Apr 26, 2020, 7:55 PM

#

all the math you have to do on snowflakes theres a deconstruct method for it

#

there really is not much math u can do on snowflakes

snow crypt Apr 26, 2020, 7:56 PM

#

(should there even be in the first place?)

rigid trail Apr 26, 2020, 7:56 PM

#

what

snow crypt Apr 26, 2020, 7:56 PM

#

forget it

rigid trail Apr 26, 2020, 7:56 PM

#

the deconstruct is used internally too

snow crypt Apr 26, 2020, 7:56 PM

#

timestamps, right?

rigid trail Apr 26, 2020, 7:57 PM

#

making it a bigint is quite frankly, pretty useless for most if not all users, and detrimental, as it will be a pretty gigantic change

#

also inconsistent with dapi as snowflakes are strings when sent, not integers

snow crypt Apr 26, 2020, 7:59 PM

#

aren't they strings to support languages that dont have support for 64b ints?

#

which now, js isn't

rigid trail Apr 26, 2020, 8:01 PM

#

yes, but the support isnt as seamless as other languages like python, it really serves no purpose yielding a bigint, instead of a string which can easily be converted to a bigint if somebody needs it, which again, 99% of users wont need bigint snowflakes

snow crypt Apr 26, 2020, 8:03 PM

#

actually, yeah probably

rigid trail Apr 26, 2020, 8:04 PM

#

so this change would be:

Gigantic (most bots rely on the id property)
Pretty useless
Mostly just an annoyance
Thats my take on it

snow crypt Apr 26, 2020, 8:06 PM

#

well yeah, useless - probably

#

anyway, v13 would be a gigantic change on its own, so bigint snowflakes shouldn't be the thing we're afraid is going to break alot

rigid trail Apr 26, 2020, 8:07 PM

#

just because other breaking changes will happen doesnt mean we should break things just because

#

realistically, what use will this bring, for most users, absolutely nothing lol, it will just be annoying

#

and unlike something like managers, it doesnt really have a purpose for the library itself either

snow crypt Apr 26, 2020, 8:10 PM

#

again, just a possibility ¯\_(ツ)_/¯

lofty birch Apr 26, 2020, 8:11 PM

#

There's absolutely no point to use bigints here, it's like making a change for the point of making a change which is really rather stupid

ornate topaz Apr 26, 2020, 8:12 PM

#

aren't they strings to support languages that dont have support for 64b ints?
which now, js isn't
Note that BigInt wasn't "merged" into native number

rigid trail Apr 26, 2020, 8:14 PM

#

number is still a primitive, and it still has the same size, bigint is a specialty object if you need to deal with large numbers

alpine pier Apr 26, 2020, 8:33 PM

#

I mean people accidentally use numbers for ids already and it's causing them errors everywhere... and while you could argue that that's a beginner level problem, you can't deny that it will just bring more errors when people are trying to handle them like number, forgetting that yes, bigints are sth different

copper laurel Apr 26, 2020, 10:07 PM

#

The fact that the language has BigInts does not make them a good or well supported option and it will do nothing to address people who use IDs as numbers. There is no use case I can think of for making the switch, and as Vlad pointed out, they need to be sent as strings to the API anyway

tacit crypt Apr 26, 2020, 10:17 PM

#

I mean can we use BigInts? Totally! Should we? nah

alpine pier Apr 26, 2020, 10:22 PM

#

Monbrey that's not what I meant. Messing up Bigints and numbers is even easier than Strings and numbers. And using bigints is also needlessly complicated judging by this mdn doc. So the few places where you'd want to have ids as numbers (which I can't think of rn) wouldn't benefit from it in the slightest anyways

copper laurel Apr 26, 2020, 10:34 PM

#

Yeah I know, I was agreeing with you, or I meant to be. BigInts are not a useful implementation for us to be using

real jetty Apr 27, 2020, 5:53 PM

#

This is a snippet from the discord.js sharding manager. Can someone tell me why the await statement is not outside the for loop? (github link: https://github.com/discordjs/discord.js/blob/026691702da7d1fcf9a959f00082f07f1e14b13d/src/sharding/ShardingManager.js#L209)

vernal atlas Apr 27, 2020, 5:54 PM

#

because the promises array is defined in the for loop?

exotic flicker Apr 27, 2020, 5:54 PM

#

if you look at it, there are delays getting pushed into the promises array to delay the next loop run

real jetty Apr 27, 2020, 5:58 PM

#

I get that. But this way the shards would spawn serially. I'm just wondering if it would be better if they spawn async?

#

Or am I missing something?

slate nacelle Apr 27, 2020, 6:00 PM

#

There is a 1/5 rate limit for identifying, this avoid failing attempts at such.

real jetty Apr 27, 2020, 6:06 PM

#

Got it. Thanks!

tawny warren Apr 27, 2020, 7:27 PM

#

i wanted to know if the owner of a guild is always cached or not, because intents are giving me an headache with this

raven juniper Apr 27, 2020, 8:05 PM

#

no, also a question for support channels not here @tawny warren

#

guild owners are guildmembers same as any other member and are subject to cache, use guild.ownerID instead

umbral aspen Apr 28, 2020, 8:06 PM

#

Wouldn't it be logical for the GuildDelete action to use a typing method in it's condition (for example, ClientUser#typingIn()) to return a value to determine whether or not the client user is typing in the provided channel to not call TextChannel#stopTyping redundantly? Or am I missing the point here?

real jetty Apr 28, 2020, 9:01 PM

#

TextChannel#stopTyping() already has a check for this https://github.com/discordjs/discord.js/blob/stable/src/structures/interfaces/TextBasedChannel.js#L234 (which is pretty much the same to how ClientUser#typingIn() does it)

dusk dagger Apr 29, 2020, 9:46 AM

#

Hey, hope it's the good channel to ask. I'm wondering if the fetch function is of any use in the RoleManager. As of my understanding, the cache should basically always be up-to-date, so is there a reason the method exist?

Source of what I'm talking about is there: https://github.com/discordjs/discord.js/blob/stable/src/managers/RoleManager.js#L48

tacit crypt Apr 29, 2020, 9:55 AM

#

Not everyone keeps everything cached and there's no issues having a fetch method just in case

dusk dagger Apr 29, 2020, 10:00 AM

#

What do you mean by "not everyone keeps everything cached"? Is the cache opt-out? And so, would you recommend using fetch() over cache.get()? That was my understanding that is was the preferred approach but wanted to be sure.

tacit crypt Apr 29, 2020, 10:42 AM

#

No, the cache isn't "opt-out" but some people might manually clear the cache

dusk dagger Apr 29, 2020, 10:45 AM

#

Alright. But either way, the best practice would be to use fetch, right?

tough geode Apr 29, 2020, 10:51 AM

#

not unless you modify cache. cache should always be up-to-date, as you said earlier, too.

dusk dagger Apr 29, 2020, 11:01 AM

#

I know I'm exaggerating, but is there any way that the cache would somehow miss roles? Packet loss, problem with the Discord API, or whatever? A very rare occurrence, but still possible?

copper laurel Apr 29, 2020, 11:04 AM

#

not really no

dusk dagger Apr 29, 2020, 11:11 AM

#

So fetch is only there as a short-hand and if we manually remove entries from the cache, then. Got it, thank y'all

rigid trail Apr 29, 2020, 12:09 PM

#

its also there for api coverage

#

just like guild manager and channel manager's fetch, its not much use to most people but some people may use it and the library should allow users to be able to use the whole api

umbral aspen Apr 29, 2020, 2:50 PM

#

Ah, I didn't notice that. Thanks for informing me, @real jetty. \👍

snow crypt Apr 29, 2020, 10:28 PM

#

what is the reason for dynamic in image options being false by default?

copper laurel Apr 29, 2020, 10:33 PM

#

breaking changes most likely

snow crypt Apr 29, 2020, 10:36 PM

#

thought of that, but why wouldn't it be true by default in the first place?

orchid stirrup Apr 29, 2020, 11:42 PM

#

When was discord.js made?

outer night Apr 29, 2020, 11:51 PM

#

This isn't a question exactly appropriate for this channel, but the first commit of discord.js was Aug 9, 2015

#

If you have more questions about the history of discord.js please ask in #archive-offtopic

ornate topaz Apr 30, 2020, 6:27 AM

#

@snow crypt well, because it would be breaking. before dynamic you would always get an image in format you wanted ({ format: 'png' }). Since png is not gif, you get png.
To get a gif when person has it, and a png when person doesn't have it, you would have to check if avatar hash starts with a_ and pass format appropiatelly. That's exactly what dynamic does.
if it would be true by default, you would have exactly 0 idea why the hell you pass { format: 'png' } and get a gif, but only for few people.

snow crypt Apr 30, 2020, 6:47 AM

#

didn't know dynamic didn't exist when the initial switch to functions was done, makes sense

drifting knot Apr 30, 2020, 8:13 PM

#

discord.js-next/pull/24 is just updating a lockfile and adding 2 empty typescript files

surreal geode Apr 30, 2020, 9:32 PM

#

It's a WIP PR. It's to track changes by that contributor

ornate topaz Apr 30, 2020, 10:34 PM

#

Draft PRs exist, and this one isn't marked as one. is your #25 also wip for tracking, or is it somewhat ready?

rigid trail Apr 30, 2020, 10:36 PM

#

there isnt even much you can do with the pr when there is no rest, no client, and no gateway lol

copper laurel Apr 30, 2020, 10:51 PM

#

I should also point out that any PRs which are nothing more than a copy-pasting or port of existing code with some type annotation slapped on are unlikely to be accepted

surreal geode May 1, 2020, 5:31 PM

#

@ornate topaz It's for tracking. @copper laurel I know it might not be accepted

ornate topaz May 1, 2020, 10:14 PM

#

@surreal geode If it's for tracking, why is it not marked as draft? Draft PRs are unmergable until marked as ready

surreal geode May 1, 2020, 11:55 PM

#

Oops sorry. It's marked as draft now

topaz mango May 2, 2020, 12:16 PM

#

Should docstrings for Collection mention that sort and sorted are stable starting with Node 12?

snow crypt May 4, 2020, 8:10 AM

#

Shouldn't there be a clone method for MessageEmbeds?

#

also freeze

#

more like permissions handle it

#

that would also prevent actual embeds from actual message be changed

topaz mango May 4, 2020, 8:14 AM

#

You can already do new MessageEmbed(embedToClone)

lofty birch May 4, 2020, 8:16 AM

#

I've seen this, and seen that the default method is process, just wondering what would be better about using process vs worker? https://discord.js.org/#/docs/main/stable/typedef/ShardingManagerMode

topaz mango May 4, 2020, 8:17 AM

#

Also, worker threads are not experimental anymore since Node 12.x

snow crypt May 4, 2020, 8:26 AM

#

@topaz mango i know, but you can also do new Permissions(permissionsToClone), but as they can be frozen, the methods check for it and operate on clone if frozen

#

shouldn't this be done with embeds too?

#

also, it's non-breaking to update

topaz mango May 4, 2020, 8:29 AM

#

What's the use case?

snow crypt May 4, 2020, 8:30 AM

#

same as for permissions

#

making Message#embeds array elements to be frozen

#

see as GuildChannel#permissionsFor returns a Readonly<Permissions>, because both memberPermissions and rolePermissions (internal) call .freeze() on the permission before returning

slate nacelle May 4, 2020, 8:32 AM

#

Making Message#embeds contain frozen elements is breaking.
As much as I'd like them to be frozen.

snow crypt May 4, 2020, 8:37 AM

#

@slate nacelle how is it?

#

the interface would be the same

#

but MessageEmbed methods would check if it is frozen, and if it is - return a modified clone

#

again, like bitfields do

slate nacelle May 4, 2020, 9:07 AM

#

They'd no longer return this but a new embed, changing it's behaviour in a non backwards compatible way.

topaz mango May 4, 2020, 11:26 AM

#

Does https://github.com/discordjs/collection have a minimum Node requirement?

#

Context for the question: I want to either clarify (if Node below 12 is supported) or remove the mention of sort being not necessarily stable.
Because in Node 12+ it should be stable.

snow crypt May 4, 2020, 12:09 PM

#

@slate nacelle oof

#

well

#

can we have a branch for breaking changes on v12?

raven juniper May 4, 2020, 12:10 PM

#

That's v13, aka next

snow crypt May 4, 2020, 12:33 PM

#

but next is gonna be ts

#

and unit tested

#

and more

#

i mean a branch parallel to master, but that accepts breaking changes

rigid trail May 4, 2020, 12:34 PM

#

but semantically that would be v13, which is djs next

topaz mango May 4, 2020, 12:35 PM

#

So no further breaking changes are expected to the current codebase? Even if Discord does something unexpected to the API that needs them?

rigid trail May 4, 2020, 12:35 PM

#

it is highly unlikely discord will make unexpected breaking changes to the api and require people to bump to the new api version

#

but no, v12 will not have breaking changes

topaz mango May 4, 2020, 12:38 PM

#

By definition, yes. My question is "can there be a v13 before d.js-next gets ready".

rigid trail May 4, 2020, 12:39 PM

#

im no dev but for that to happen djs-next would have to be made v14, @slate nacelle would that be acceptable in the unlikely event discord required breaking changes

slate nacelle May 4, 2020, 12:43 PM

#

No clue, we'll see it if it comes.

snow crypt May 4, 2020, 1:20 PM

#

@rigid trail can't it just be pre-v13-dev or something?

rigid trail May 4, 2020, 1:20 PM

#

wtf its still v13

#

semantically its v13 and djs next would be v14

snow crypt May 4, 2020, 1:24 PM

#

still feels like it should exist

copper laurel May 4, 2020, 1:47 PM

#

If there is a need for a breaking change, then PR it to master and a v13 will have to be released

#

The one being discussed here however is really not worth a semver major bump on its own

humble wyvern May 4, 2020, 2:07 PM

#

I saw, by doing a mistake in my code, that GuildMember.hasPermission(permission) doesn't throw any error when the permission isn't in the good format (for example, a typo in the permission flags).
The execution just stop silently.
I wonder if it's useful to open a PR to fix that or if it's just done on purpose? I hope it's the good channel for, if not, really sorry (don't hesitate to @ me)

lofty birch May 4, 2020, 2:16 PM

#

doesn't occur for me

#

(owner and admins will always return true unless you specify to discount that)

humble wyvern May 4, 2020, 2:18 PM

#

My precise mistake was :

message.guild.me.hasPermission("VIEW_AUDIT_LOG,MANAGE_WEBHOOKS", {checkAdmin : true})```
In guild where the bot didn't get admin permission, the execution just silently stopped without any message error (don't ask my why I did add it this way...)

lofty birch May 4, 2020, 2:20 PM

#

humble wyvern May 4, 2020, 2:20 PM

#

It's most probably an edge case

#

Strange

topaz mango May 4, 2020, 2:20 PM

#

Probably you're eating exceptions somehow without displaying

humble wyvern May 4, 2020, 2:22 PM

#

I log all unhandledRejection, but I will take a better look. Happy to see it's not due to d.js, and my bad k3llyOK

#

Thanks and sorry Sweat

surreal geode May 4, 2020, 2:55 PM

#

I don't think that's how bitfields work

#

Pass in an array ['VIEW_AUDIT_LOG', 'MANAGE_WEBHOOKS'] instead

frozen bridge May 4, 2020, 8:53 PM

#

Are Discord API endpoint changes going to be added to v11?

unique axle May 4, 2020, 8:56 PM

#

Oct 7th: intents are required - will not be happening for v11
Nov 7th: domain changes are required - doesn't really matter as v11 is already not usable anymore

drifting knot May 5, 2020, 12:18 AM

#

in djs 12 Manager#fetch is typed as Promise<Thing>, where Thing could be a user/channel/whatever. this is assuming that the fetch operation always succeeds and finds something. is this by design or a mistake?

rigid trail May 5, 2020, 12:22 AM

#

its not assuming that, the promise could reject

drifting knot May 5, 2020, 12:29 AM

#

does the promise reject when the snowflake isn't found

#

looking at source code, no

#

god i can't wait until djs next is released and i can bathe in first class ts support

opal mauve May 5, 2020, 12:32 AM

#

not sure what specific manager you're referring to, but at least the channel manager seems to have the appropriate logic

copper laurel May 5, 2020, 1:01 AM

#

Yeah not seeing the issue, most will include a call like this: const data = await this.client.api.users(id).get()
That handles and throws DiscordAPIErrors, so the fetch call, which is just an async function, would reject

snow crypt May 5, 2020, 8:59 AM

#

Shouldn't channel type group be removed?
bots can't be added to group dms, so kinda seems like a weird choice to keep it

lofty birch May 5, 2020, 9:02 AM

#

They could be added to group dms before iirc, so some still may be there

slate nacelle May 5, 2020, 9:03 AM

#

Bots can fetch invites to group dms, which yields you such a channel.

snow crypt May 5, 2020, 9:04 AM

#

oh, ok

snow crypt May 5, 2020, 2:37 PM

#

@slate nacelle continuing the embed discussion, what if we added .clone() that just returns a new this.constructor(this)

#

or add freezing, but not freeze current non-frozen stuff

copper laurel May 5, 2020, 11:16 PM

#

.clone() would be redundant, just construct it normally. The whole thing is currently non-frozen so you can't add freezing to it

snow crypt May 6, 2020, 6:16 AM

#

Why is https://github.com/discordjs/discord.js/pull/2910 not being merged?

snow crypt May 6, 2020, 12:42 PM

#

@slate nacelle why do you check if the max listeners are not zero?

slate nacelle May 6, 2020, 12:43 PM

#

Because 0 is used as "no limit", incrementing would set the limit to 1, decrementing it would throw.

snow crypt May 6, 2020, 12:43 PM

#

and what about #2910?

real jetty May 6, 2020, 6:43 PM

#

Would a pr that set the default id of fetch to 0 be useful ? I say that because when you do not provide anything it returns a 404 instead of 'uknown user'

rigid trail May 6, 2020, 6:44 PM

#

it would be a breaking change

real jetty May 6, 2020, 6:45 PM

#

Why ? 🤔

rigid trail May 6, 2020, 6:45 PM

#

it changes the response

real jetty May 6, 2020, 6:46 PM

#

yes I see 🤔

snow crypt May 6, 2020, 9:04 PM

#

@opaque robin could you add what space requested in #4103? I really wanna see it merged

#

also, the d.js error is not imported

loud moat May 6, 2020, 10:02 PM

#

Why doesn't it throw an argument error if you call fetch without passing an user object instead of doing a bad request?

copper laurel May 6, 2020, 10:05 PM

#

This seems more like a support question, but you'd really need to be more specific about which fetch endpoint at least, there are a lot

rigid trail May 6, 2020, 10:06 PM

#

that would be breaking btw, so that could not be changed either ways

loud moat May 6, 2020, 10:08 PM

#

I was replying to the thing above

obsidian copper May 7, 2020, 1:27 AM

#

https://github.com/discordjs/discord.js/issues/4151 any ideas on the situation with this?

#

Causing major problems with one of my shards right now, sending 3 messages per command

copper laurel May 7, 2020, 1:29 AM

#

There's an open PR aiming to fix it

obsidian copper May 7, 2020, 1:31 AM

#

Alright ^ ty

opaque robin May 7, 2020, 5:42 AM

#

@snow crypt k I’ve just been busy with other stuff

tawny warren May 7, 2020, 7:57 AM

#

i have a suggestion for a future version: a user/guildmember hybrid object without an user property

#

my reasoning: intents can get messy and sometimes member.user can return undefined and everything goes downhill

copper laurel May 7, 2020, 8:52 AM

#

There's a few issues with that. First of all, the User and GuildMember objects/endpoints are completely different in the Discord API - they aren't a hybrid and as an API wrapper, it wouldn't be a correct implementation.
Second, while in theory a GuildMember object could extend User and include those properties, its not a good design pattern. It means we'd have to update every GuildMember hybrid object whenever a shared User property changes, instead of just the single User object which is referenced by member.user

#

In a broader sense about your concern with intents - I agree. The way discord.js is currently built does rely a lot on its internal caches for a lot of things, and those need specific intents in order to be filled and for that functionality to work. Making any changes to that behaviour would be huge, and almost definitely breaking. It'll be considered as part of the discord.js-next / v13 rewrite

snow crypt May 7, 2020, 1:59 PM

#

@slate nacelle how is it cached?

tawny warren May 7, 2020, 7:27 PM

#

isn't it there a way to iterate through getters/setters and peoperties and attach them to an object so the references are kept? or i just understand the core of references wrong?

snow crypt May 7, 2020, 7:37 PM

#

you could use what is called delegation

#

so when you access member.tag it will actually be member.user.tag

#

?npm delegates

rich iglooBOT May 7, 2020, 7:38 PM

#

snow crypt May 7, 2020, 7:38 PM

#

but there's really no reason for that

drifting knot May 8, 2020, 6:55 AM

#

i was going to write up some TS enums with pretty comments for gateway opcodes, where would those files belong in the ideal djs@next repo?

snow crypt May 8, 2020, 8:36 AM

#

@drifting knot i guess it will be the types package

drifting knot May 8, 2020, 8:43 AM

#

thought so

#

will open PR in a day or so

digital hawk May 8, 2020, 10:39 AM

#

would this affect d.js https://github.com/npm/rfcs/blob/latest/accepted/0025-install-peer-deps.md

#

npm v7 looks like it will install peerdeps automatically

#

and I know d.js has peerdeps which might conflict like sodium and libsodium

snow crypt May 8, 2020, 11:40 AM

#

sounds critical

#

maybe opening an issue would raise more attention

exotic flicker May 8, 2020, 11:41 AM

#

there are some solutions in the file, I guess we would need to do some of those

snow crypt May 8, 2020, 3:02 PM

#

are private props in WebSocketShard intentionally not shown in the docs?

exotic flicker May 8, 2020, 3:02 PM

#

what props are you referring to?

snow crypt May 8, 2020, 3:03 PM

#

anything defined with Object.defineProperty

#

https://github.com/discordjs/discord.js/blob/master/src/client/websocket/WebSocketShard.js#L118

slate nacelle May 8, 2020, 3:05 PM

#

Probably not

snow crypt May 8, 2020, 3:07 PM

#

pr incoming blobreach

tacit crypt May 8, 2020, 4:16 PM

#

All of those are private props you don't need access to

#

I guess they're missing an @name although then again; you won't need those variables ever

vernal atlas May 8, 2020, 5:07 PM

#

@snow crypt i think both would be the better option

#

the fix i suggested also fixes another issue i mentioned here a couple days ago

snow crypt May 8, 2020, 5:09 PM

#

which one?

vernal atlas May 8, 2020, 5:09 PM

#

to where calling suppressEmbeds throws an API Error if you have the allowedMentions client option enabled, and said message is authored by another user

tacit crypt May 8, 2020, 6:06 PM

#

If the content isn't present

#

👀

vernal atlas May 8, 2020, 6:06 PM

#

well, i tested passing undefined, but that removed the mentions

#

instead of keeping them

tacit crypt May 8, 2020, 6:07 PM

#

That sounds like an API bug

vernal atlas May 8, 2020, 6:07 PM

#

could be

tacit crypt May 8, 2020, 6:08 PM

#

~~aka test it and report tbh~~

vernal atlas May 8, 2020, 6:14 PM

#

ah, no i was mistaken before, both null and undefined do the same

thorny copper May 9, 2020, 8:57 AM

#

I don't know if it's already been asked, but it'd be awesome to have a property User#nitroType which returns either 0, 1 or 2, as API returns, or <null>, NITRO_CLASSIC or NITRO

unique axle May 9, 2020, 9:04 AM

#

nitroType requires an authenticated request (through oauth2, with the identify scope). As we do not provide any utility for oauth2 requests having such a field would not bring much if any benefit
All the information a API request from a bot for a user returns:

{
  id: '83886770768314368',
  username: 'Souji',
  avatar: 'c97d04911a8bc6f059900e913649fd1b',
  discriminator: '0001',
  public_flags: 640
}

ref(oauth2): https://discord.com/developers/docs/topics/oauth2

thorny copper May 9, 2020, 9:09 AM

#

GET /users/{user.id} doesn't require oath. Otherwise, how could you return User#locale?

copper laurel May 9, 2020, 9:24 AM

#

That endpoint doesn't, but when requested by a bot token, it does not return information such as locale and nitroType. Retrieving that data does require full OAuth, as Souji said

thorny copper May 9, 2020, 9:26 AM

#

Then how do you return User#locale?

unique axle May 9, 2020, 9:29 AM

#

the full return from that endpoint is in my initial response.
we also can not provide the locale, despite it being documented it will always be null, there used to be a dummy value for bots which was provided, however this is no longer true
the property still existing in the library is an oversight

thorny copper May 9, 2020, 9:34 AM

#

Sad :c

#

Thanks

#

Can we hope that, one day, we can have support for oauth2 requests?

lofty birch May 9, 2020, 9:44 AM

#

That requires a web server, which is not the purpose of djs.

unique axle May 9, 2020, 9:47 AM

#

unlikely, discord.js is a wrapper for the Discord bot API
the purpose of oauth2 is to grant a token to do requests as the user, which is no longer part of the bot API
the request also requires an endpoint for the redirect

thorny copper May 9, 2020, 9:48 AM

#

ok, thanks

digital hawk May 9, 2020, 1:57 PM

#

maybe opening an issue would raise more attention
@snow crypt i'll open an issue when npm v7 is close to release as rn the latest seems to be 6.14.5

snow crypt May 9, 2020, 1:58 PM

#

@digital hawk open it now so we can fix it ahead of time

digital hawk May 9, 2020, 2:01 PM

#

ok

strange igloo May 9, 2020, 5:25 PM

#

Hey, quick question regarding one of the library dependencies
The choice of node-fetch for > 12.x has any particular reason beside snekfetch being deprecated?

#

Okay, thanks 👌

lofty birch May 10, 2020, 1:09 PM

#

Either way one that works some of the time would be better than none at all

real jetty May 10, 2020, 1:09 PM

#

A fix really isn't that hard, it just seems like they made a mistake while implementing

#

I can PR one, I just wanted to know the reason why it got removed before I open one

tacit crypt May 10, 2020, 1:16 PM

#

There will be proper rest stack traces once I get to actually re-doing my REST PR

#

space pls

real jetty May 10, 2020, 1:17 PM

#

Ahh great! Then I won't open a PR

wild flax May 10, 2020, 6:03 PM

#

It's nice that you can, but please don't be overconfident and think we didn't have internal discussions and trials before reverting it.

#

It wasn't "just" a silly mistake. Also theres performance implications with async stacktraces on top.

lofty birch May 10, 2020, 6:05 PM

#

could be nice to have an option to use async stacktraces then, so it's off by default, but can be enabled

wild flax May 10, 2020, 6:06 PM

#

No such thing without a working implementation

ornate topaz May 10, 2020, 6:07 PM

#

why have a feature that doesn't work all the time? you would explicitly enable async traces, just to get that [object Object] when you least expect it

tacit crypt May 10, 2020, 6:31 PM

#

Altering the stack trace isn't the solution

#

And it only works in v8 and not in any browsers that don't run v8b(like Mozilla or Safari)

#

The solution is implementing something that doesn't mess with the stack

tender field May 11, 2020, 9:06 AM

#

In which case there can be multiple activities to a presence ? (=> Why is it an array (GuildMember.presence.activities) ?)

lofty birch May 11, 2020, 9:08 AM

#

Custom Status + Other presence

unique axle May 11, 2020, 9:09 AM

#

many more cases actually, discord just sends all activities, no matter which are displayed in the client GUI
spotify, games, code editors, custom status, etc.
this used to not be the case, but was added as feature (by discord) around the time when v12 neared end of development

tender field May 11, 2020, 9:11 AM

#

oh ok thanks... ^^

#

maybe that should be written in the docs, because it is not really clear when this can happen ^^

tacit crypt May 11, 2020, 9:25 AM

#

Simplest example would be rich presence and spotify

real jetty May 12, 2020, 1:28 AM

#

has something been done about this, bots can ping roles now

#

i'm in a bit of a panic since this is going to lead to some serious abuse

#

disabledmentions doesnt seem to have anything to stop role pings right now

#

other then not formatting them at all

#

or maybe i'm just stupid

ornate topaz May 12, 2020, 1:29 AM

#

https://canary.discordapp.com/channels/222078108977594368/222078108977594368/693190178965159956

real jetty May 12, 2020, 1:31 AM

#

Shouldn't that mean it's off by default?

#

I updated discord.js quite recently so something seems off

ornate topaz May 12, 2020, 1:33 AM

#

yes, by default it's not set

#

just like disableMentions is

real jetty May 12, 2020, 1:33 AM

#

So should I just set allowedMentions to 'none' in the client

#

Docs said it's the default value but maybe I'm just reading it wrong

#

I tested and other bots seem to be able to ping as well

#

Oh - I figured it out. Sorry to bother aquaDerp

real jetty May 12, 2020, 7:19 PM

#

Hi there, I was wondering about support for the new js runtime, deno. It hasn't hit its 1.0 release yet, so i understand if development hasn't started yet, but I think it would be interesting to see a d.js module for deno

#

If not, i'd be happy to help when deno gets released. good learning opportunity for me :P

vernal atlas May 13, 2020, 6:50 AM

#

regarding the issue #4284, this seems to be the problem, https://github.com/discordjs/discord.js/blob/2583ad5da7db5ab92dfa357909f9c713db0a2981/src/structures/Message.js#L479-L484, but why was that even done in the first place? seeing as it causes the event to be emitted twice

#

or at least, done in a way that emits the event

vernal atlas May 13, 2020, 9:13 AM

#

g#4121

rich iglooBOT May 13, 2020, 9:13 AM

#

real jetty May 13, 2020, 11:11 AM

#

@vernal atlas methods that cause an event to be emitted called the handler itself as a fallback (for cases when the ws wasn't connected or w/e), the handler then has some logic that checks if its alr been handled but that seems to be missing on MessageReactionAdd. there was a discussion a while ago about always assuming the ws event would come thru and not call the handler itself but instead handle the response within the method itself to avoid double emissions (which was done on Message#delete() / TextBasedChannel#bulkDelete() a couple of months ago iirc) so #4285's approach seems fine except it needs to call MessageReaction#_add

vernal atlas May 13, 2020, 11:12 AM

#

KidThumbsUp

opaque vessel May 13, 2020, 1:44 PM

#

Hi hi! Would I be right in stating this thrown TypeError is improper?
https://github.com/discordjs/discord.js/blob/master/src/managers/GuildMemberRoleManager.js#L93

It has a grammatical error: TypeError [INVALID_TYPE]: Supplied roles is not an Role, Snowflake or Array or Collection of Roles or Snowflakes..

The other TypeError just above on line 85 seems fine though! In comparison, the latter TypeError includes all possiblities of the accepted types, but the former TypeError doesn't include them all which I'm not sure is intentional 😅