Geyser: "given scan urls are empty. set urls in the configuration" on startup

[DoctorMacc]

Describe the bug
On starting up plugin versions, the message [org.geysermc.platform.<PLATFORM>.shaded.reflections.Reflections] given scan urls are empty. set urls in the configuration is given. This is a harmless error that affects nothing.

To Reproduce
Start up a plugin version of Geyser.

Geyser Version
As of build 377.

[anoxic]

this might have something to do with the Reflections library: https://github.com/ronmamo/reflections/issues/186

[rtm516]

It is, and an issue with the version we use but if we upgrade it then it breaks android support.

[anoxic]

Maybe this stackoverflow will help: https://stackoverflow.com/questions/41932635/scanning-classpath-modulepath-in-runtime-in-java-9/45612376#45612376

[DoctorMacc]

At a cursory glance this isn't specifically a Java 9 thing - I develop with Java 8 and get this all the time.

[MichaIng]

Seeing this on Java 11 as well.

[NotStirred]

I'm seeing issues with it taking more than 3 minutes to get through it. Whether it's this or something else I'm not sure

[10:34:52 WARN]: [org.geysermc.platform.spigot.shaded.reflections.Reflections] given scan urls are empty. set urls in the configuration
[10:38:20 WARN]: [org.geysermc.platform.spigot.shaded.reflections.Reflections] given scan urls are empty. set urls in the configuration

[MichaIng]

Yes those messages are appearing with a significant delay (heavily depending on the power of the host machine of course, in my case it was something around 5-10 seconds) between each other. But it cannot be derived that the underlying issue is the reason for the delay, but maybe there is a lot of other processing which at some distant points produce the warnings.

[DoctorMacc]

I would expect that the delay is not because of those errors.

[soshimee]

It is, and an issue with the version we use but if we upgrade it then it breaks android support.

d...owngrade maybe?

[Chew]

I believe you misunderstand. The version we are on gives that warning, but IF we upgrade, Android support will no longer work. No need to downgrade, it won't solve anything.

[Advewave]

Any update on this bug? It's still going on for me.

[yehudahrrs]

Yes. This bug is part of a library we use, you can ignore it and it will not break anything.

[Advewave]

Yes. This bug is part of a library we use, you can ignore it and it will not break anything.

thank you!

[soshimee]

I don't know why people say it takes minutes to go through that stage, it only takes 3~5 secs with a small server.

[MichaIng]

I tested it on some VMs (old i3 3110M CPU) where it took a significant time (minutes range all together). A native x86 server will likely have no issues, so practically it's more the SBC servers, which however become increasingly relevant (and also capable) in general, I guess.

[Heath123]

What about using a fork of Reflections? Or is it better to wait for them to fix Android?

[Tim203]

It's a harmless message, so I don't think that we need to fork it.

[Heath123]

Yes, but it's confusing and with a fork you could pull in some of the new changes probably

[Heath123]

0.9.11 vs 0.9.12
https://github.com/ronmamo/reflections/compare/c95609a4e83e1f3fc71e4b2fc58b17352a16eef7..7741e8f5d13f739d644ef5e91dde055fbe88ca57

[jackson-57]

It's a harmless message, so I don't think that we need to fork it.

Agreed, although over in the Discord, we still get a ton of people asking about it. I think that forking it wouldn't be too hard to do, but if we don't want to go that route, could we possibly add our own message to the top/bottom of it, telling people to ignore?