Struct Output 

pub struct Output { /* private fields */ }

The output channels for one command run.

Cloning is cheap: clones share the same writers.

Implementations

impl Output

pub fn stdio() -> Self

Output bound to the process stdout and stderr, with progress indicators enabled (they still hide themselves when stderr is not a TTY).

pub fn capture() -> (Self, Captured)

Output bound to in-memory buffers, with progress indicators disabled.

Returns the output and a handle that reads what was written.

pub fn capture_merged() -> (Self, Captured)

Output whose two channels write to one in-memory buffer, in the order of the writes, with progress indicators disabled.

The TUI uses it: its output pane shows data and diagnostics as one stream, as a terminal does. Captured::out returns the whole stream and Captured::err returns the same text. Captured::data returns only what went to the data channel.

pub fn teed(&self, limit: usize) -> (Self, Captured)

Output that still writes to this output’s writers and also keeps a copy of the text, as Output::capture_merged does.

The TUI uses it for a command that runs outside it: the user sees the text live on the terminal, and the TUI shows the copy in its scrollback after it returns. Progress indicators stay as they are here (they draw on the real terminal and are not copied). The copy keeps at most about limit bytes, the end of the text; whole lines are dropped from its start. Captured::trimmed says when that happened.

The result is a TUI output (Output::is_tui): its hints are read in the TUI scrollback.

pub fn progress_enabled(&self) -> bool

True when animated progress indicators may draw on the terminal.

pub fn is_tui(&self) -> bool

True when this output belongs to the TUI. Error text then gives TUI next steps, not CLI flags.

pub fn progress_hook(&self) -> Option<ProgressHook>

The progress record of this run, when the output has one. The TUI’s output has one; the plain CLI output does not.

pub fn series_hook(&self) -> Option<SeriesHook>

The series of a repeating run, when the output has one: the TUI’s pane output has one; the plain CLI output does not.

pub fn record_run(&self, metrics: &[Metric], record: RunRecord) -> bool

Hands one run to the screen’s series. Returns false when there is no series (the CLI), so the caller prints the run instead.

pub fn update_progress(&self, change: impl FnOnce(&mut RunProgress))

Changes the progress record. Does nothing when there is no hook.

pub fn write_out(&self, args: Arguments<'_>) -> Result<()>

Writes formatted data to the data channel. Use outln! / out!.

Errors

Returns the writer’s error, for example BrokenPipe when a reader closes the pipe early or a storage error when stdout is a full disk.

pub fn write_err(&self, args: Arguments<'_>) -> Result<()>

Writes formatted text to the diagnostic channel. Use errln! / err!.

Errors

Returns the writer’s error.

pub fn note(&self, args: Arguments<'_>)

Writes a status line to the diagnostic channel. Use note!. A failed write is dropped: see the module policy.

pub fn note_saved(&self, what: &str, path: impl Display)

The one “saved to” line: <what> saved to: <where>, as a note.

Trait Implementations

impl Clone for Output

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.