When typing quickly, can ignore the previous useless request?

[yezige]

When typing quickly, can ignore the previous useless request?

eg. I search forhains, and will search for ha,hai, andhains respectively.

[yezige]

I have an idea that I don't know if it can be realized, as follows:

var timeout
function oninput(s) {
  timeout && clearTimeout(timeout)
  timeout = setTimeout(() => {
    doquery(s)
  }, 500)
}
function doquery(s) {
  console.log(s)
}
oninput('ha')
oninput('hai')
oninput('hains')

setTimeout(() => {
  oninput('hains point')
}, 1000)

output:

hains
hains point

It can be seen that when typing quickly, only the real requirements:hains, hains point are successfully executed

[amrbashir]

We do debounce search queries, see

meilisearch-docsearch/src/DocSearchModal.tsx

Lines 203 to 227 in 97cb6a2

	const debouncedSearch = utils.debounce(search, 100);
	if (initialQuery) {
	onMount(() => {
	setQuery(initialQuery);
	search(initialQuery);
	});
	}
	function onInput(
	e: InputEvent & {
	currentTarget: HTMLInputElement;
	target: Element;
	},
	) {
	const query = e.currentTarget?.value;
	setQuery(query);
	if (!query) {
	onReset();
	return;
	}
	debouncedSearch(query);
	}

it is 100ms though, not 500ms

[yezige]

Cool, maybe 100ms is too short.

[amrbashir]

I will probably make this configurable instead.

[yezige]

Great, but I think 100ms is indeed too low, 500ms will be better, it seems that there is no need to change it to configurable, because everyone's needs are not much different.

[amrbashir]

from my previous testing, 500ms was a bit slow and noticeable specially when selecting some text on page then hitting the search hotkey.

[yezige]

Hmm, you need to decide the specific data. this time should be greater than the speed of normal people typing two letters.

[yezige]

加油老哥

[yezige]

太秀了，神速