Klaus Post
Klaus Post
There are literally millions of uses of this per day, so the chance of you doing concurrent writes is a bit bigger than an unseen bug, that hasn't been caught...
Without a reproducer it is quite hard to get further. Would it be feasible to wrap the encoder in a `defer func() { if r == recover(); r != nil...
I know the issue. My hunch is that if this is triggered, simply slapping a "if level == 0 { something}" will just cause something else to crash - or...
@cfif1982 Hiding your comment as off-topic, as it is in unrelated code in another package. It safe, since the `!e.longTableShardDirty[i]` will always be true. It is however inefficient, so it...
Closed as it seems to be an isolated incident.
@zdz1715 If you submit a PR, please include a regression test.
Pro: Convenient Cons: Seems a bit strange to offer this as a single config to abstract. Undecided.
> This is mainly for fix the behavior of this API for certain section of applications like alluxio underfs. Okie, but how does it work across a cluster?
Conflicts and ``` 2024-08-09T10:16:09.0411135Z === RUN TestListMultipartUploadsHandler 2024-08-09T10:16:09.0412005Z bucket-handlers_test.go:490: Test 2: ErasureSD: Expected the response status to be `404`, but instead found `200` 2024-08-09T10:16:09.0413146Z bucket-handlers_test.go:507: Test 2: ErasureSD: Expected the...
@marktheunissen If you have time to look at this, that would be awesome. Without testing, I think something like this is the serverside fix: ```diff diff --git a/cmd/bucket-handlers.go b/cmd/bucket-handlers.go index...