-
Notifications
You must be signed in to change notification settings - Fork 657
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
Failure to process URL where path is nothing but "//" #948
Comments
I am currently hitting this issue in Spring Cloud gateway.
(I am testing hoverfly in front of spring cloud gateway btw, https://hoverfly.io/, may help you reproduce the issue) I thought I would try to see if it is fixed in 0.9.5 but unfortunately, there is still an NPE. Seems like you re-introduce the bug two days ago, via 327e9b3 @violetagg Here is the current stack trace:
|
@qcastel can you provide more info? I did the following: Spring Gateway
Hoverfly started in capture mode
Then Spring Gateway requested like below and the response is received
Running with latest Reactor Netty 0.9.5.BUILD-SNAPSHOT |
Ok so indeed, that wasn't that obvious to reproduce the issue. Did manage to reproduce it in isolation at the end, I shared a demo repo here: https://github.com/qcastel/reproduce-netty-npe Essentially two factors are triggering this issue: Thanks for looking at this issue btw! |
@qcastel I can reproduce it |
@violetagg Just tried on https://github.com/qcastel/reproduce-netty-npe and think the NPE is resolved now! Although I do have another error:
I also simplified the reproduction steps to not have hoverfly and directly hit Spring cloud gateway. I guess I will raise a new ticket on Spring cloud gateway this time! |
@qcastel so the new failure is not in Reactor Netty? |
Hard to tell, I don't get any logs this time. It could still be the netty rejecting the 'CONNECT' requests. |
After upgrading to Dysprosium snapshots, there is one test failure in the Spring Framework in a scenario where the path is nothing but a double slash, e.g.
"http://localhost:8080//"
.It results in the following:
This looks like it might be a side effect of refactoring for #926 in dc3fb4b.
The text was updated successfully, but these errors were encountered: