containers/attach: Note bug around goroutine leak

[cgwalters]

See:

• disk: Pass /dev/null to containers.attach stdin podman-bootc#59 (comment)
• disk: Switch to just invoking podman run podman-bootc#61

Sorry for not trying to fix this, but I am not aware of a remotely straightforward way to do so.

None

[cgwalters]

I am not aware of a remotely straightforward way to do so.

Searching around a bit https://groups.google.com/g/golang-nuts/c/4ZrDw__VkyI looks like the most related thread. I can imagine doing that, basically drop down to a select/poll syscall on the fd, and have that send a message on a channel when ready; the other goroutine selects on that and the cancellation context. When the select chan is ready we do a single read+write cycle.

In the end, it is just simpler and clearer for us to just fork off podman as a binary which fully papers over this bug, but maybe someone else who wants to fix this can find this PR discussion later and do it.

[Luap99]

Yeah this is really the ugly parts about go that you cannot cancel a read without closing the stream.

Also your tree is outdated please rebase on the latest main

[Luap99]

select/poll syscall on the fd

Yeah that would make it ugly as we have to support the bindings on windows as well. So then we woul dend up with OS specific code...

[Luap99]

LGTM

[openshift-ci]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: cgwalters, Luap99

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [Luap99]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[packit-as-a-service]

Ephemeral COPR build failed. @containers/packit-build please check.

[rhatdan]

/lgtm