Aleksa Sarai

No, in the past I only ever built `linux/amd64`. I'll add some extra binaries to `make release`.

I suspect that the best solution would actually be to entirely remove the idmapping logic from C and instead implement it in Go (if we can use `move_mount(2)` entirely without...

It's been a while since I reworked the user code -- do we use the `$HOME` detection at all? Or is absolutely everything ignored by runc? I wouldn't be surprised...

> For [os.LookupId()](https://pkg.go.dev/os/user#LookupId), I wonder if that would work, because we need to lookup the HOME based on the container environment, not the "current" environment? I was wondering the same...

I'm not sure if the cgroup configuration for `procHooks` is actually necessary -- we already set up the cgroups far earlier. IIRC the reason for this being added in the...

The problem is that some of the recent mount rework is only tested through `mount_sshfs` because that was the way the old behaviour was tested. Disabling the tests will also...

I still think #3985 is needed for 1.2.0. There are several issues with mount propagation (not to mention the serious restriction of idmapped mounts) that are fixed with #3985. We...

IMHO this is ultimately a Go stdlib bug -- they assume that only the Go process itself will change its own rlimits (and thus use a cached version of the...

Copying my comment from https://github.com/opencontainers/runc/pull/4234#issuecomment-2029874782, which explains my current thinking on the issue. > I think the issue is more likely related to the thread-local storage used by pthreads. From...

> > It seems that there is no error for `fork`. > > https://public-inbox.org/libc-alpha/[email protected]/ For those reading at home, this is https://github.com/bminor/glibc/commit/c579f48edba88380635ab98cb612030e3ed8691e. So, they removed the PID cache entirely and...