-
-
Notifications
You must be signed in to change notification settings - Fork 3.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Revert "Set socket family of connector to AF_INET" #9870
Conversation
842f216
to
035e26b
Compare
This change was made since Discord doesn't support IPv6, and there were concerns about clients with DNS64 enabled without NAT64. However, this breaks hosts who don't have v4 connectivity and are _actually_ running NAT64. Having DNS64 without NAT64 is really an issue on the client's end. It would break far more than just discord.py, so I don't think we should be concerned about those cases.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
I'm not a fan of this solution. This should probably be a |
Users can (Actually, this appears to have been removed, see now linked PR) pass in a connector already, having this be set to AF_INET by default causes the error in the more complex networking case to at least make sense, the error users would get without this in the failing case was not useful, see error from the original issue, maybe this should just be solved with additional documentation that if you have NAT64 without ipv4 you need to also have DNS64 and pass in your own connector. |
This provides paths for users to handle two entirely seperate issues - Alternative fix for Rapptz#9870 - Allows handling of windows sslcontext issues without a global truststore.inject_into_ssl() use
This provides paths for users to handle two entirely seperate issues - Alternative fix for Rapptz#9870 - Allows handling of windows sslcontext issues without a global truststore.inject_into_ssl() use
This provides paths for users to handle two entirely seperate issues - Alternative fix for Rapptz#9870 - Allows handling of windows sslcontext issues without a global truststore.inject_into_ssl() use
This provides paths for users to handle two entirely seperate issues - Alternative fix for Rapptz#9870 - Allows handling of windows sslcontext issues without a global truststore.inject_into_ssl() use
This provides paths for users to handle two entirely seperate issues - Alternative fix for #9870 - Allows handling of windows sslcontext issues without a global truststore.inject_into_ssl() use
I've given this some more thought and I'm just going to go ahead and allow it. |
* Fix documentation for abc.GuildChannel.move to be more clear * Fix Polls limiting duration at 7 days * Fix _get_command_error improperly handling some error messages * Fix introduced potential TypeError with _get_command_error * Re-add client connector param This provides paths for users to handle two entirely seperate issues - Alternative fix for Rapptz#9870 - Allows handling of windows sslcontext issues without a global truststore.inject_into_ssl() use * Fix Poll.duration rounding error * Add support for getting the attachment's title * Fix default_avatar for team user and webhook * [docs] Remove pack_id attribute from Sticker [docs] Remove unnecessary pack_id * Fix url for GIF StickerItem * [docs] correct hyperlink to discord docs * [docs] Fix typehint for Embed.set_(image,thumbnail) * Remove stale documentation in Embed.set_thumbnail * Add Guild.fetch_role * Add approximate_user_install_count to AppInfo * Revert "Set socket family of connector to AF_INET" This change was made since Discord doesn't support IPv6, and there were concerns about clients with DNS64 enabled without NAT64. However, this breaks hosts who don't have v4 connectivity and are _actually_ running NAT64. Having DNS64 without NAT64 is really an issue on the client's end. It would break far more than just discord.py, so I don't think we should be concerned about those cases. * Add support for get sticker pack * Fix stacklevel for Message.interaction deprecation warning * Use fallback audioop package for Python v3.13 or higher Fix Rapptz#9477 * Add DummyCookieJar to client owned ClientSession * Add missing guild incident fields Co-authored-by: owocado <[email protected]> Co-authored-by: Danny <[email protected]> * Add support Member.fetch_voice * Add support for application emojis Co-authored-by: DA344 <[email protected]> Co-authored-by: Danny <[email protected]> * Add a warning if interaction endpoint URL is set on login * Remove aiodns from being used on Windows --------- Co-authored-by: Rapptz <[email protected]> Co-authored-by: DA344 <[email protected]> Co-authored-by: Pipythonmc <[email protected]> Co-authored-by: Michael H <[email protected]> Co-authored-by: ambdroid <[email protected]> Co-authored-by: Soheab <[email protected]> Co-authored-by: owocado <[email protected]> Co-authored-by: Andrin <[email protected]> Co-authored-by: Gaurav <[email protected]> Co-authored-by: fretgfr <[email protected]> Co-authored-by: lmaotrigine <[email protected]> Co-authored-by: Alex Nørgaard <[email protected]> Co-authored-by: Oliver Ni <[email protected]> Co-authored-by: Jakub Kuczys <[email protected]> Co-authored-by: Deep Jain <[email protected]> Co-authored-by: Danny <[email protected]>
This change was made since Discord doesn't support IPv6, and there were concerns about clients with DNS64 enabled without NAT64.
However, this breaks hosts who don't have v4 connectivity and are actually running NAT64 (me!)
Having DNS64 without NAT64 is really an issue on the client's end. It would break far more than just discord.py, so I don't think we should be concerned about those cases. I'm proposing this change be reverted.