Fork me on GitHub
#xtdb
<
2021-03-20
>
tianshu06:03:58

Hi, I just found crux.api/project-many does not respect the order of eids, is it expected?

refset11:03:11

Hi 🙂 looking at our tests I see we return sets https://github.com/juxt/crux/blob/9fe64e36e1a7c03646afc93945665d11871a6cf4/crux-core/test/crux/pull_test.clj#L122 Whereas DataScript returns vectors that respects the input order https://github.com/tonsky/datascript/blob/60feefdd2b2f0d354d4def55b84bed35288e3c0a/test/datascript/test/pull_api.cljc#L72 I'm not sure what might have prompted our choice to return sets but I will investigate! Thanks for mentioning it. Hopefully you're okay simply running pull multiple times as a workaround for now

tianshu12:03:29

Thanks! Just saw Crux have a complete support for pull, great job!

🙏 6
tianshu12:03:06

Can I create an issue for this? In case anyone who have the same confusion.

refset13:03:55

I can't take any credit for the implementation - but I'm glad you approve! Certainly, please feel free to open an issue, that will make 100% sure we get some eyes on it this week 🙂

tianshu14:03:48

Hmm, I think I understand why it returns set, because pull-many is just one more q , and the query itself can't guarantee the result have the same order with eids. It make sense for current implementation. Thanks for the help!

lispyclouds11:03:37

Hello! Im seeing a couple of errors come up when trying out JDK 16:

java.lang.reflect.InaccessibleObjectException: Unable to make field transient volatile java.util.concurrent.ConcurrentHashMap$Node[] java.util.concurrent.ConcurrentHashMap.table accessible: module java.base does not "opens java.util.concurrent"
at crux.cache.second_chance.ConcurrentHashMapTableAccess
and
java.lang.NoClassDefFoundError: Could not initialize class crux.cache.second_chance.ConcurrentHashMapTableAccess
Is this known?

👋 3
lispyclouds11:03:39

Ah found the issue: https://github.com/juxt/crux/issues/1462

🙂 3
👍 3