Parse channel and message urls with SnowflakeParser#355
Parse channel and message urls with SnowflakeParser#355Rutherther wants to merge 1 commit intoRemora:mainfrom
Conversation
Rutherther
force-pushed
the
feat/snowflake-parse-url
branch
from
6dbef75 to
0785c79
Compare
|
The parser should also handle the subdomains of other client versions (ptb, canary) if this is to be accepted, imo e.g., I use canary, and thus message and channel links are |
|
Hm, I had no idea that is a thing. How do you obtain such urls though? The one I used in the parser is obtained by right clicking and copying a link |
Rutherther
force-pushed
the
feat/snowflake-parse-url
branch
from
0785c79 to
95fedea
Compare
|
By using Discord's beta clients (e.g. for Windows ); ptb is available via the webpage; both can be used in the browser by just prepending |
|
If such urls are to be allowed, I will switch to regex matching with a group, and also check for only numbers, as at that point, it doesn't really matter. I didn't do it just because I haven't used regex for this simple case. |
Oh you mean like if you copy the url from the bar, right |
|
No; the beta clients give links that match; e.g. https://canary.discord.com/channels/789912892426027039/789924168837365810/1389617555861143764 edit: I meant that PTB (Public Test Beta) can be downloaded via Discord's website |
|
Give... how? |
|
Right clicking a message like normal Discord -> Same goes for all links that can be |
|
So Copy link button okay. I also see Discord parses all of them, so I agree with your proposal. I will probably not be able to make it today, though. |
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
Rutherther
force-pushed
the
feat/snowflake-parse-url
branch
from
95fedea to
7c20456
Compare
|
So, as proposed, I added ptb and canary support. I used regex from MessageParser and changed it a little. I made message id optional, and also changed * with + for the individual snowflake parts of the url. Tbh I don't understand why messageparser uses * instead of +, it seems to me that https://discord.com/// would match... But I haven't tried it yet. Let me know what you think. |
| public override ValueTask<Result<Snowflake>> TryParseAsync(string value, CancellationToken ct = default) | ||
| { | ||
| var channelLinkMatch = _channelLinkRegex.Match(value); | ||
| if (channelLinkMatch.Success) |
Check warning
Code scanning / InspectCode
Invert 'if' statement to reduce nesting Warning
There was a problem hiding this comment.
I don't get this tbh.
If user sends a link to a channel, it's not being parsed by Remora. On the other hand the client understands this link and visually there is no difference between a mention and a channel link. Add parsing of a channel links, without checking the whole string, only beginning 'https://discord.com/channels/' is used to check this is a url, and only the last part of the string is then used, all other parts are ignored.
Rutherther
force-pushed
the
feat/snowflake-parse-url
branch
from
7c20456 to
b36653b
Compare
2 participants
Hey, a user of my bot tried to use a link in a command. It wasn't working and it causes bit of a confusion as the mention looks the same a sent link in the client. I will use this in my bot if it won't be accepted here for sure, but wanted to check with you here for feedback / to merge this to Remora.Discord directly if others will appreciate this behavior as well. I can add more validation if needed, ie. to check if there are only 2 or 3 parts after the url (guild, channel, message), and to check if they consist of all numbers if you think that is necessary. But I don't see any risks if wrong links, such as https://discord.com/channels/XYZ/XYZ/XYZ/XYZ/ are accepted.
If user sends a link to a channel, it's not being parsed by Remora. On the other hand the client understands this link and visually there is no difference between a mention and a channel link.
Add parsing of a channel links, without checking the whole string, only beginning 'https://discord.com/channels/' is used to check this is a url, and only the last part of the string is then used, all other parts are ignored.