Gabriel de Perthuis

There's probably a way to use serde(flatten) to stuff unrecognised fields into a json::Value field while keeping the standard fields.

I have a branch that adds async (draft PR #28), in a not too intrusive way (adds 93 lines net under src/). It's on top of my other two PRs...

The current API goes either through receive or through receive_find_map. receive_find_map callers do the request-response association, but a client that calls receive in parallel and discards some messages could starve...

I have replaced receive_find_map with a new send_get_reply in my branch (), now everything is plugged properly so that receive() only gets non-replies and wakes send_get_reply() futures, while send_get_reply() polls...

The issue was that bcachefs superblock detection can silently fail due to an outdated blkid dependency. That's now detected at format time (#347) but it needs to be in a...

@thememika I get logs from journalctl or pstore (look at /usr/lib/tmpfiles.d/systemd-pstore.conf to enable), then I pipe them through `scripts/decode_stacktrace.sh vmlinux` from the kernel source tree.

I couldn't reproduce the bug, on the same kernel build (of a6c90193bc06c071ac00989791a284cf94def0e5 merged with 6.12.1) the upgrade went through the next time I tried (also with fsck,fix_errors). Nothing particular in...

The first time with the crash, I was mounting two filesystems at the same time, and the upgrades (from 1.13 to 1.18) were eating up memory. Next time, I was...

Link to another report of the same crash, also during the 1.13 to 1.18 upgrade: https://gist.github.com/noradtux/cb4f4a0234d2cd91871cb56714add986/

Tried to reproduce, downgrade went well, upgrade failed to mount with this error reported in dmesg: ``` 12:30:56.916511: bcachefs (fff…): extents_to_backpointers: 12%, done 14783/116807 nodes, at extents:134351211:619088:U32_MAX 12:31:06.696502: bcachefs (fff…):...