-
Notifications
You must be signed in to change notification settings - Fork 2.5k
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
Activate camel-core integration tests via -Dnative #2352
Conversation
Hm... the Windows build failed in opentracing which has nothing to do with any of the current changes. |
Build_Native_Linux is passing for me locally without |
The
But it is passing for me
What makes it pass here and fail there? |
When I add Any other ideas what makes it pass here and fail there? |
It could be a timing difference. |
What kind of timing? |
Like a case where a race condition determines the value of a field, which in turn causes something to get initialized or not initialized. Essentially it would mean there is a bug, but it might not manifest all the time. |
Thanks. That hypothetical bug would be in SVM, right? I am just reading this article that explains Do you happen to know if a similar issue was seen on Quarkus before (and how it was solved)? |
In this case, no, probably not. Something causes an
This kind of thing happens fairly frequently. You need to follow the reference graph that is produced in order to deduce where it is coming from. There are reports generated by the build that could be helpful for this purpose. |
Thanks for the explanation, @dmlloyd !
|
Exactly; it'll only show up on the failing system. |
I don't know how you ended up with an extra Windows Build. What I do know is that I don't have enough permissions to delete it off of there. 🤷♂ |
Interesting. I am not aware of doing anything to create it. |
I was able to get a call graph of a failing build - see the |
quarkus/extensions/jaxb/runtime/src/main/java/io/quarkus/jaxb/runtime/graal/JAXBSubstitutions.java Lines 175 to 179 in c5da691
Injector from being initialized at build time (at least via this particular call path)? If yes, I do not see how the error may occur.
|
…e.map on class com.sun.beans.WeakCache
Before this change -Dnative-camel was needed and the camel-core integration tests were not run on Quarkus CI.
If that's the case then the only conclusion I can think of is that substitution is sometimes not working, or that there's a conflicting substitution which is sometimes taking priority and SubstrateVM is not complaining about it. |
I just found the magic switch |
Fwiw, in theory disabling xml and/or jaxb should not be a condition to have native support for Camel. Those flags were added as a way to have lightweight camel native applications by using more dead code elimination. |
Yes, I also rather see the magic switches as a workaround. I'll file a new issue for the Injector failure on monday. Is the current PR good to merge? |
Reported #2417 |
@gnodet could you please review?