fix(emulator): diagnose a spawn failure instead of blaming the toolchain

[zackees]

Closes #1390.

The message was wrong for every case it fired on

Every failure to spawn the emulator went through build_linux_macos_qemu_hint:

The cached QEMU toolchain may be incomplete or corrupt. On Linux/macOS,
also ensure runtime deps are installed: libgcrypt, glib2, pixman, SDL2,
and libslirp.

That advice fits exactly one failure mode — a missing shared library — and
that mode cannot reach this code path. The dynamic loader runs after a
successful execve, so a missing .so surfaces as exit code 127, which the
monitor already detects thirty lines below and already words correctly:

} else if status.code() == Some(127) {
    MonitorOutcome::Error(format!(
        "{} failed to start (exit code 127): a required shared library is missing.\n{}",
        label, build_linux_macos_qemu_hint("")

So on the spawn path the runtime-deps text was never the right answer for
anything. A reader whose real problem was a missing binary got sent to
reinstall five libraries that were never involved.

Not hypothetical — it cost a debugging session in this repo. A CI failure read
failed to launch QEMU at sh: No such file or directory followed by advice
about pixman and SDL2; the actual defect was one bare program name in a test
helper (#1389), and nothing about QEMU or system libraries was wrong.

What it says now

The spawn site branches on io::Error::kind():

kind	message
NotFound	names the path tried, says nothing ran, and says outright that reinstalling runtime libraries will not help
PermissionDenied	present but not executable — check the exec bit and noexec mounts
anything else	the path plus the OS error verbatim, rather than a guess

build_linux_macos_qemu_hint is untouched and still used by the exit-127 path,
which is the one place it is correct.

Tests

Three, and the first one asserts a negative — that a missing binary's message
contains none of libgcrypt, glib2, pixman, SDL2, libslirp. That is the
actual regression to guard: the failure here was never a crash, it was
confidently wrong advice, and only an assertion about what the message must
not say can catch it coming back.

soldr cargo test -p fbuild-daemon --lib — 246 passed. clippy -D warnings
clean.

[coderabbitai]

Warning

Review limit reached

Next included review available in 59 minutes.

View limit details

Limit details: You’ve used the included review currently available.

You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository.

Learn how review limits work.

Review configuration:

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro Plus

Run ID: 8fba53bb-3aab-4c1c-beab-d54a750f3206

📥 Commits

Reviewing files that changed from the base of the PR and between 218caf1 and 86cc220.

📒 Files selected for processing (1)

• crates/fbuild-daemon/src/handlers/emulator/shared.rs

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}