jdx/mise · error

task {} cache command input must not be empty: {command:?}

Error message

task {} cache command input must not be empty: {command:?}

What it means

When computing/verifying a task's cache, mise re-derives command inputs from the task's cache.command_inputs entries. Each entry is treated as a command string; a blank or whitespace-only entry cannot produce a meaningful program/args pair, so mise fails the task naming the offending entry rather than silently ignoring it.

Source

Thrown at src/task/task_executor.rs:1423

            sandbox.filter_env(resolved_env)
        } else {
            resolved_env.clone()
        };
        let timeout = task
            .timeout
            .as_ref()
            .and_then(|value| match duration::parse_duration(value) {
                Ok(timeout) => Some(timeout),
                Err(err) => {
                    warn!("invalid timeout {:?} for task {}: {err}", value, task.name);
                    None
                }
            })
            .unwrap_or(COMMAND_INPUT_TIMEOUT);
        let mut inputs = Vec::with_capacity(cache.command_inputs.len());
        for command in &cache.command_inputs {
            if command.trim().is_empty() {
                eyre::bail!(
                    "task {} cache command input must not be empty: {command:?}",
                    task.name
                );
            }
            let (program, args, cmd_verbatim) =
                self.get_cmd_program_and_args(command, task, &[])?;
            // The same refusal as in `exec_program`, and it matters more here: these commands feed
            // the cache key, so running them from C:\Windows would hash the wrong directory's
            // answer. Measured on 2026.8.6 with the project on a UNC share — a `command_inputs`
            // entry reading a file that exists in the project fails, while the same config on a
            // local path succeeds. `--dry-run` does not reach here (see the cache branch in
            // `run_task`), so there is nothing to exempt.
            #[cfg(windows)]
            if cmd_shell_cannot_use_dir(&program, &root) {
                eyre::bail!("{}", unc_working_dir_error(&root));
            }
            #[cfg(not(windows))]
            let _ = cmd_verbatim;

View on GitHub (pinned to afd2eddd3a)

Solutions

1. Remove the empty entry from cache.command_inputs in the task definition
2. Fix the variable/template inside the entry that resolves to an empty string
3. Filter blank strings at generation time if the list is built programmatically
4. Re-run the task after cleaning the config

Example fix

# before
[tasks.build.cache]
command_inputs = ["", "Cargo.lock"]
# after
[tasks.build.cache]
command_inputs = ["Cargo.lock"]

const cmds = (config.cache?.command_inputs ?? []).map(c => c.trim());
const bad = cmds.findIndex(c => c === "");
if (bad !== -1) throw new Error(`command_inputs[${bad}] is empty`);

fn non_blank(s: &&str) -> bool { !s.trim().is_empty() }
// let valid: Vec<_> = command_inputs.iter().filter(non_blank).collect();

try {
  await miseRun(["build"]);
} catch (e) {
  if (String(e).includes("cache command input must not be empty")) {
    console.error("Fix cache.command_inputs in mise.toml: remove blank entries");
  }
  throw e;
}

When it happens

Trigger: A task defines cache.command_inputs containing "", " ", or a tabs/newlines-only string; running the task (not under --dry-run) reaches task_executor.rs:1423 and bails on that entry.

Common situations: TOML like command_inputs = ["", "Cargo.lock"] from copy-paste; a template/variable that evaluates to an empty string inside an entry; a programmatically built list with an empty element.

Understand the failure class

Background: "must not be empty", "cannot be empty" — required-field validation errors across open-source libraries — this error's family across 41 libraries.

Related errors

• action prediction payload is too large
• task action manifest has an invalid identity
• task action manifest contains duplicate predictions
• remote action manifest keys must use blake3
• cached rustc output set is incomplete

AI-assisted analysis of jdx/mise@afd2eddd3a (2026-09-09). Data as JSON: /api/errors/6fe6f35bf57793f2. Report an issue: GitHub.