Problem
`internal/containers/manager.go:332` (and the equivalent line in any future Exec variant):
```go
_ = m.cli.ContainerExecResize(ctx, execResp.ID, container.ResizeOptions{
Height: uint(size[1]),
Width: uint(size[0]),
})
```
Error is dropped on the floor. Silent failures hide real problems (stale exec, network hiccup, permission issue, daemon restart mid-session). Operators have no telemetry to correlate 'zellij didn't resize' with a Docker API error.
Proposal
Log at `slog.Warn` with `component: "exec-cleanup"` (or a new `exec-resize` component) — exec ID, target size, and error. Keep the non-returning behavior: resize is best-effort, a single failed resize shouldn't kill the session.
Motivation
Surfaced during root-cause investigation for #3. Not the primary bug fixed there, but adjacent and cheap to address separately.
Effort
Trivial — one-line change plus a unit or integration assertion that the log line fires.
Problem
`internal/containers/manager.go:332` (and the equivalent line in any future Exec variant):
```go
_ = m.cli.ContainerExecResize(ctx, execResp.ID, container.ResizeOptions{
Height: uint(size[1]),
Width: uint(size[0]),
})
```
Error is dropped on the floor. Silent failures hide real problems (stale exec, network hiccup, permission issue, daemon restart mid-session). Operators have no telemetry to correlate 'zellij didn't resize' with a Docker API error.
Proposal
Log at `slog.Warn` with `component: "exec-cleanup"` (or a new `exec-resize` component) — exec ID, target size, and error. Keep the non-returning behavior: resize is best-effort, a single failed resize shouldn't kill the session.
Motivation
Surfaced during root-cause investigation for #3. Not the primary bug fixed there, but adjacent and cheap to address separately.
Effort
Trivial — one-line change plus a unit or integration assertion that the log line fires.