For me, I was looking into how the contrib checks can work in the new GitHub CI setup. I don't think it's super easy to make it exist how it does today. I think Infra would have to bless a GitHub App or bot user to vote using status checks. A comment may be possible, so I am planning to try that in my staging setup. I don't want it to block the conversion though. It is an optional check after all. I also revived my LLM-assisted port of CH2++ now that the license for CH2++ itself has been clarified. Claude got a little too ambitious and basically ended up porting NumPy and Python's PRNGs to match the exact generated data with a given seed. I haven't made up my mind yet as to whether or not this is a wise idea. It is very interesting though, and clean in a way. Test-only code like that certainly allows more freedom. I haven't quite understood how to properly note the license though, given that those clearly have to be based on their respective implementations (which both have compatible licenses, so it should be OK)
- Ian On Fri, Oct 23, 2026 at 10:12 AM Calvin Dani <[email protected]> wrote: > > Calvin > - Working on VTree Speed up and Scale up Experiment > - Preparing schema inference patch > > On Fri, Oct 23, 2026 at 8:42 AM Ian Maxon <[email protected]> wrote: > > > Hi everyone, > > Looking forward to hearing what everyone has been upto this week in > > AsterixDB. > > In case you aren't in ASF slack or didn't catch the message, I had to > > temporarily restrict Gerrit to only allow known safe IPs in for now > > due to abuse. Anubis was able to save us, but a particularly vicious > > swarm of scrapers came on Friday and performed a DoS on the CI > > infrastructure. I plan to focus my energy on the planned migration to > > GHA rather than devising a new method to repel these attacks. If you > > want access to Gerrit, please email me and I will add your IP to the > > allowlist. > > > > Toplevel status: > > - GitHub Actions migration 1st step up on Gerrit > > > > Issue roundup: > > New: > > > > - 3958 (Retry failed files when doing parallel S3 downloads) > > - 3959 (ExceptionUtils.matchingCause only checks the first cause) > > - 3960 (HyracksConnection.close() leaks threads) > > - 3961 (Advance netty to 4.2.19.Final) > > - 3962 (S3SyncDownloader.close() never shuts down its thread pool) > > - 3963 (Cloud retries treat socket timeouts as interrupts) > > - 3964 (GCS parallel file download ignores failed downloads) > > - 3965 (Failed cloud file download leaves a remnant file) > > - 3966 (Outer variable referenced after GROUP BY binds to an arbitrary > > row or wont compile) > > - 3967 (Port CH2++ Datagen to Java) > > - 3969 (Migrate CI and CR to GitHub) > > - 3970 (Migrate all Jenkins jobs to GitHub Actions Workflows) > > - 3971 (Iceberg scan fails when the pinned snapshot records no schema ID) > > - 3972 (Iceberg variant pruning skips files whose decimals read back > > equal to the literal) > > - 3973 (Iceberg delete-file cache can exceed its total size budget) > > > > Fixed/Implemented: > > - n/a > > > > Open APEs: > > Schema inference: Starting again? > > > > Codegen: > > Drafting/sketching - License Compatibility Inspection > > > > Sampling cursor: > > APE exists in cwiki but needs vote thread on dev@ > > > > Sandboxed UDFs: > > Needs more discussion > > > > External functions: > > Discussion > > > > Dynamic memory > > No discussion for a while > > > > MCP > > Awaiting > > > > Columnar schemas for unbounded object keys (APE 37) > > Draft in cwiki > > > > Blocked-ish patches: > > - Geometry type for Columnar (still, maybe for next release) > > - Docs improvements (still, but hopefully for 0.9.10/1.0.0) > > - Metadata function introspection (probably next release?) > >
