stz is e280's standard library of environment-agnostic typescript tools. zero dependencies.
ergonomic event emitters
import {pub, sub} from "@e280/stz"- make a publisher fn
// create a pub fn const sendMessage = pub<[string]>() // subscribe to it sendMessage.subscribe(m => console.log(m)) // publish to it sendMessage("hello")
- make a subscriber fn β it's just like pub, except it's flipsy-reversey!
// create a sub fn const onMessage = sub<[string]>() // subscribe to it onMessage(m => console.log(m)) // publish to it onMessage.publish("hello")
- pub and sub both have the same facilities
.publish.subscribe.on.next.clear
- i seem to use
submore often
- publish actually returns a promise, to wait for all async subscribers
await onMessage.publish("hello")
- subscribe returns a fn to unsubscribe
const unsubscribe = onMessage(() => {}) unsubscribe()
.clear()to wipe all subscribed listenersonMessage.clear()
.next(fn?)is a better way to do .once..- you can use it like a .once:
onMessage.next(message => {})
- but it also gives you a promise like this:
const [message] = await onMessage.next()
- of course the promise can be used like this:
onMessage.next().then(([message]) => {})
- you can use it like a .once:
defer the resolve/reject of a promise to the outside
import {defer} from "@e280/stz"
const deferred = defer()- resolve the deferred promise
deferred.resolve()
- reject the deferred promise
deferred.reject(new Error("fail"))
- await the promise
await deferred.promise
sleep for some milliseconds
import {nap} from "@e280/stz"
await nap(900)
// wait for 900 millisecondseasy trash management
import {disposer} from "@e280/stz"- create a disposer
const dispose = disposer()
- schedule something for cleanup
dispose.schedule(() => console.log("disposed!"))
- schedule multiple things at once
dispose.schedule( () => console.log("disposed thing 1"), () => console.log("disposed thing 2"), () => ev(window, {keydown: () => console.log("keydown")}), )
- schedule is chainable if you prefer that vibe
dispose .schedule(() => console.log("disposed thing 1")) .schedule(() => console.log("disposed thing 2")) .schedule(() => ev(window, {keydown: () => console.log("keydown")}))
- dispose of all that garbage
dispose()
extended js data types
extended js Map
- many are saying it's "The Deluxe Mapping Experience"
import {GMap} from "@e280/stz" const map = new GMap<number, string>([ [1, "hello"], [2, "world"], ])
map.require(key)β returns the value for key.. if missing, throw an errorconst value = map.require(1) // "hello"
map.guarantee(key, makeFn)β returns the value forkey.. if missing, runmakeFnto set and return the valueconst value = map.guarantee(3, () => "rofl") // "rofl"
extended js Set
new GSet<T>()set.adds(item1, item2, item3)β add multiple items without a for-loopset.deletes(item1, item2, item3)β add multiple items without a for-loop
extended js WeakMap
new GWeakMap<K, V>()weakMap.require(key)β returns value for key.. if missing, throw an errorweakMap.guarantee(key, makeFn)β returns the value for key.. if missing, runmakeFnto set and return the value
execute calls in sequence (not concurrent)
import {queue, nap} from "@e280/stz"
const fn = queue(async() => nap(100))
fn()
fn()
await fn() // waits for the previous calls (sequentially)ensure a fn is only executed one time
import {once} from "@e280/stz"
let count = 0
const fn = once(() => count++)
console.log(count) // 0
fn()
console.log(count) // 1
fn()
console.log(count) // 1throws an error if the async function takes too long
import {deadline} from "@e280/stz"
const fn = deadline(100, async() => {
// example deliberately takes too long
await nap(200)
})
await fn()
// DeadlineError: deadline exceeded (0.1 seconds)wait some time before actually executing the fn (absorbing redundant calls)
we use debounce a lot in ui code, like on a user's keyboard input in a form field, but rendering the form input can actually be slow enough that it causes problems when they type fast β to eliminate the jank, we debounce with like 400 ms, so we wait for the user to finish typing for a moment before actually running the validation.
import {debounce} from "@e280/stz"
const fn = debounce(100, async() => {
await coolAction()
})
// each fn() call resets the timer
fn()
fn()
fn()
// coolAction is only called once here, other calls are redundantexecute a function over and over again, back to back
import {cycle} from "@e280/stz"
let ticks = 0
const stop = cycle(async() => {
// use a nap to add a delay between each execution
await nap(200)
ticks++
})
// stop repeating whenever you want
stop()convert to/from utf8 string format
txt.fromBytes(bytes)β bytes to stringtxt.toBytes(string)β string to bytes
utilities for dealing with Uint8Array
bytes.eq(bytesA, bytesB)β check if two byte arrays are equalbytes.random(32)β generate crypto-random bytes
convert binary data to/from various encodings
import {hex, base58, base64} from "@e280/stz"all BaseX utilities have these methods
hex.fromBytes(u8array)β encode bytes to stringhex.toBytes(str)β decode string to byteshex.toInteger(string)β decode string as js integerhex.fromInteger(n)β encode js integer as a stringhex.random(32)β generate random encoded string (32 bytes)
hexbase2base36base58base62base64base64url
- you can provide a
lexiconto produce your own BaseX codecconst myHex = new BaseX({characters: "0123456789abcdef"})
- fun fact: you can make insanely compact timestamp strings like this:
base62.fromInteger(Date.now() / 1000) // "1uK3au"
1748388028base10 epoch seconds (10 chars)1uK3aubase62 epoch seconds (6 chars)- nice
friendly string encoding for binary data
a bytename looks like "midsen.picmyn.widrep.baclut dotreg.filtyp.nosnus.siptev". that's 16 bytes. each byte maps to a three-letter triplet
the bytename parser (bytename.toBytes) ignores all non-alphabetic characters. thus midsen.picmyn, midsenpicmyn, and mid@sen$pic@myn are all equal.
import {bytename} from "@e280/stz"-
bytename.fromBytes(new Uint8Array([0xDE, 0xAD, 0xBE, 0xEF])) // "ribmug.hilmun"
-
bytename.toBytes("ribmug.hilmun") // Uint8Array, 4 bytes
-
const data = new Uint8Array([ 0xDE, 0xAD, 0xBE, 0xEF, 0xDE, 0xAD, 0xBE, 0xEF, ]) bytename.fromBytes(data, { groupSize: 2, // default is 4 groupSeparator: " ", wordSeparator: ".", }) // "ribmug.hilmun ribmug.hilmun"
hybrid of bytename and base58 to make binary data more human-friendly
- looks like
nodlyn.fasrep.habbud.ralwel.Avo7gFmdWMRHkwsD149mcaBoZdS69iXuJ - the idea is that the first parts are in bytename format, so it's easy for humans to recognize
- and the remaining data is shown in base58
thumbprint.fromBytes(u8array)β encode bytes to thumbprint stringthumbprint.toBytes(thumbstring)β decode thumbprint string to bytesthumbprint.fromHex(hexstring)β convert a hex string into a thumbprintthumbprint.toHex(thumbstring)β convert a thumbprint into a hex string
tar-like binary file format for efficiently packing multiple files together
import {toq, txt} from "@e280/stz"- 4 magic bytes
"TOQ\x01" - for each file (little endian)
name length1 byte (u8)namex bytes (max 255 B)data length4 bytes (u32)datax bytes (max 4 GB)
- toq.pack β accepts any iterable of file entries
const pack: Uint8Array = toq.pack([ ["hello.txt", txt.toBytes("hello world")], ["deadbeef.data", new Uint8Array([0xDE, 0xAD, 0xBE, 0xEF])], ])
- toq.is β check if a file is a toq pack or not
toq.is(pack) // true
- toq.unpack β generator fn yields file entries
for (const [name, data] of toq.unpack(pack)) console.log(name, data.length)
- pack a map of files
const files = new Map<string, Uint8Array>() files.set("hello.txt", txt.toBytes("hello world")) files.set("deadbeef.data", new Uint8Array([0xDE, 0xAD, 0xBE, 0xEF])) const pack = toq.pack(files)
- unpack into a new map
const files = new Map(toq.unpack(pack))
reward us with github stars
build with us at https://e280.org/ but only if you're cool