The OCaml Planet RSS

NPY to Zarr has been completed for v1 and v1.1; day10 revdeps; Tessera tools packaged and failures in bench.ci.dev.

Fast Tessera inference that brings one Xeon node level with a GPU; reverse dependencies in day10; Windows, macOS and FreeBSD worker updates and a new ad hoc pipeline engine.

The icechunk to zarr conversion finished, although not when I said it had; benchmarking Tessera inference on Intel’s AMX silicon; and day10 spreading to four architectures while finding real bugs.

Headline activities have included resolving the scraping attacks, fixing obuilder bugs, and Tessera AMX optimisations.

Last week I updated the OpenBSD workers to OCaml 5.5.0 and took that opportunity to deploy OpenBSD 7.8. Shortly after, issue#1061 was opened as jobs randomly failed with an opam parse error.

The FreeBSD CI workers get slower over time. Is this a build-up on ZFS snapshots or something else?

This post explores two more sustainable solutions to the crawler bots issue affecting opam-repo-ci

On the base image builder, we have recently seen a significant increase in the number of times we hit the Docker rate limits on manifest pushes. Most of these work on the second attempt, but this week, a new class of permanent manifest push failures has emerged.

Scaleway provide the RISC-V workers for OCaml CI, and they have been down for about a week with no real evidence that they’ll be back anytime soon. I can’t provision any new ones as they are “temporarily out of stock”.

By now, you’ll know that I’ve been looking at OCaml CI more than I ever had before and was confused by the solver jobs, which seemed to hang around in the job queue.