Struct MotionResult 

pub struct MotionResult {
    pub target: Position,
    pub linewise: bool,
    pub exclusive: Option<bool>,
    pub notice: Option<MotionNotice>,
    pub curswant: Option<Curswant>,
}

What a motion’s evaluator returned.

Fields

target: Position

Where the cursor lands.

linewise: bool

true if the motion is linewise (ranges expand to whole lines on resolution).

exclusive: Option<bool>

VM.3c: override MotionSpec::exclusive for THIS invocation. None — the overwhelming default — means “use the spec’s flag”.

Why the axis had to move

linewise has always travelled with the RESULT and exclusive with the SPEC, and nothing needed the asymmetry resolved until a motion existed whose exclusivity is not knowable until it runs. ; is that motion: it repeats whatever f / F / t / T came last, and vim gives it that motion’s exclusivity — f and t are inclusive, F and T are not. A single flag on the ; spec has to be wrong half the time, and it was: d, after an f deleted one character too many, because , acts as F while the spec said inclusive.

Not on the WIT boundary. A plugin declares exclusivity on its MotionSpec, which is the right place for every motion that knows its own answer, so from_wit decodes this as None deliberately rather than for want of a field to read.

notice: Option<MotionNotice>

VM.3d-2: a notice for the user, which the dispatcher echoes alongside the motion’s effect — the search wrap. Not on the WIT boundary: a plugin motion reports none.

curswant: Option<Curswant>

VM.3g-3: override the goal column for THIS invocation, the way Self::exclusive overrides the spec’s flag. None — the overwhelming default — means “apply the spec’s CurswantEffect”.

Why a motion has to be able to say it

CurswantEffect can express “keep the goal”, “pin it to the line end” and “take it from where I landed”, and for every motion in vim except two that is the whole story. gj / gk are the exception: they aim at a screen column that the reached display row may be too short to hold, and vim records the aim, not the landing.

Measured in vim 9.2 (vimcheck_scroll_curswant.vim), wrap at 80 over a 240-char line then a 100-char line:

gj -> line 2 row 1            col=80   curswant=80
gj -> line 2 row 2 (CLAMPED)  col=100  curswant=160

160 is the column the motion wanted — next_start + goal, before the clamp to the row’s 100 characters. Recording the landing instead loses the aim permanently, so a following j onto a long line returns to 100 where vim returns to 160.

SetFromTarget cannot express this because the target IS the clamped landing, and Keep cannot because the goal genuinely changes on every gj (it tracks the screen column down the wrapped rows). Only the motion knows the unclamped aim, so only the motion can report it.

Not on the WIT boundary, for the same reason exclusive is not: a plugin motion declares its effect on the spec, which is the right place for every motion that knows its own answer. from_wit decodes this as None deliberately rather than for want of a field to read.

Trait Implementations

impl Clone for MotionResult

fn clone(&self) -> MotionResult

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for MotionResult

impl Debug for MotionResult

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for MotionResult

fn default() -> MotionResult

Returns the “default value” for a type.