Add metadata-2 spec #501
Add metadata-2 spec #501
Conversation
Like every other spec
|
Indeed you said that and you are right, doing SUB in pre-registration is useful too because of auto join, otherwise the client has to send SYNC again (even for a small channel). Not devastating but still. And, indeed, perhaps at some point people would want to include it in channel history or there is some other use case for it that we cannot think of right now. Better to get it right. |
| This subcommand returns which keys the client is currently subscribed to. | ||
|
|
||
| The server responds with zero or more `RPL_METADATASUBS` numerics. The server MAY return the keys in any order. The server MUST NOT list the same key multiple times in a response to this subcommand. | ||
|
|
There was a problem hiding this comment.
Without a batch or ending numeric, this is kinda tricky to deal with.
There was a problem hiding this comment.
Batches are a dependency of this spec
- Corrected some keys in example displays - Corrected a couple of missing ':' prefix in an example displays - Deduplicated `RPL_METADATASUBS` in the numerics table - Replaced RPL_ placeholders with their actual numeric value in examples Only thing I didn't touch is [this one](ircv3#501 (comment)) purely because the spec already mentions that `metadata-2` requires `batch` to work, so shouldn't be a problem =] Thank you Herringway for taking the time to point these out!
- Corrected some keys in example displays - Corrected a couple of missing ':' prefix in an example displays - Deduplicated `RPL_METADATASUBS` in the numerics table - Replaced RPL_ placeholders with their actual numeric value in examples Only thing I didn't touch is [this one](ircv3#501 (comment)) purely because the spec already mentions that `metadata-2` requires `batch` to work, so shouldn't be a problem =] Thank you Herringway for taking the time to point these out!
|
Is there any reason for metadata-notify-2 to use a cap rather than an 005 token? From what I can see you can't subscribe before connection registration anyway, there's no need for immediate behaviour that needs to be performed on seeing it, and you now need to manually sub to any caps you want to see so it doesn't make sense for it to be a cap. |
It's an opt-in feature. There's no standard way of opting into features advertised by ISUPPORT tokens. I don't see the value in having it separate from metadata-2, though |
|
|
Their first argument is the requesting client (like any numeric), followed by the target (their first own argument)
for consistency with other specs
…iple params It simplifies parsing, as it's tokenized by the main parser this way.
As far as I am aware, it addresses all known issues with the current deprecated metadata spec (listed in #339).
It is based on @DanielOaks' latest version, which is itself based on #339. Notable changes since dan's version are:
metadata-2, since it is a breaking change compared to the deprecated specrendered