Conversation
Signed-off-by: Richard Brežák <richard@numtide.com>
Signed-off-by: Richard Brežák <richard@numtide.com>
Signed-off-by: Richard Brežák <richard@numtide.com>
|
Line 68 in 4b96629 shouldn't we be opening the socket? that seems a bit more portable |
|
Line 107 in 4b96629 if im reading this correctly, any single daemon connection can only be open for up to 5 minutes by default. That's not a lot and any number you could put in the timeout won't make much sense. Individual builds can take a long time. Also I'm not sure we need the timeout at all. WSs run over TCP which already handles system level timeouts and handling daemon protocol level timeouts would require parsing and reasoning about the data being passed back&forth imo. |
Signed-off-by: Richard Brežák <richard@numtide.com>
Signed-off-by: Richard Brežák <richard@numtide.com>
|
https://github.com/numtide/nix-relay/blob/feedback/src/auth.rs#L226 the documentation for this function says:
is this turned off for a reason? if so we should document that reason with a comment |
|
Also, the CI didn't catch a failing check. 4b96629 introduced a bug which caused the |
No description provided.