dat-ecosystem-archive / dat-js Goto Github PK

The Dat website says that you can just reference
<script type="text/javascript" src="https://bundle.run/dat-js@7"></script>
and you get
a Dat prototype on the window object
...is this still valid with the updates? Mind showing a gist of how to instantiate? It it something like
var dat = window.Dat()? (Since I can't seem to get it to work :/) Thanks!

As per the conversation on IRC, we don't currently prefix the file names per archive, so if a user passes in something like random-access-idb into the Dat-js constructor, it'll cause issues with archives overwriting each other's files.

In order to fix this we should:

• Wrap the calls to storage inside the repo to prefix files with the key, similar to what I did before in dat-polyfill
• In order to do that we need to create keys for new archives before initializing hyperdrive. Similar to what I did in dat-archive-web
• Make a working example of using persistent storage through random-access-idb

I looked through the documentation and I can't seem to find any information about how to listen to changes like you can with node.js API. Is that possible in the browser version as well?

var DEFAULT_SIGNALHUBS = 'https://signalhub.mafintosh.com'
?so need a server?

Is it possible to interact with something DAT in a webapp (some static files) over HTTP? Not just be shown a data replication, but actually interact, like one or several people updating a a hyperDB, living ini indexedDB.

The reason for the question is that not many people use Beaker Browser, and not many people will install a plugin for Firefox / Chrome. Kind of "How to get the future to everyone, now" 😄

Hi I'm really interested in dat protocol, in particular I'm interested in dat-js, browser version.
I'd like to develop little app for the browser but:

• I can't find source of dat-js, althought I've found only minimized version
• I like to know if browser version cache file of others clients, like nodejs version:
  I explain me better:
  for example: I run 10 browser with simple data shared, the there're all online, well after a while one of this browser will shut down.
  Are data of the last client even available, even if only 3 or 4 of all have browsed content.
  thanks

The first thing this repo says is it's out of date. Is this approach permanently discontinued? Or is there an alternative elsewhere I should look at?

Would love to start playing with Dat!

make it work in the browser

This repo is confusing some users! Could we explain in the readme why it's out of date, and archive it?

EDIT: rewrote to be less abrupt.

First of all: awesome work!

Second, the issue: repo.archive.readFile doesn't call the callback in some case.

Reproduction

• Put the code below in an html file.
• Uncomment all comments (don't forget the comment at the end //)}).
• Open the file in the browser. You'll see the log correctly.
• Re-comment the code & open the file in the browser. The problem: nothing gets logged.

<title>Check Chrome Console</title>
<script type="text/javascript" src="https://bundle.run/[email protected]"></script>
<script type="text/javascript" src="https://bundle.run/[email protected]"></script>

<script>
var Dat = datJs
/*var db = randomAccessIdb('dats')

var dat = new Dat()

var repo = dat.create({
  db: db
})

repo.archive.writeFile('/example.txt', 'Hello World!', () => {
  // Save it for later
  localStorage.setItem('My_Repo', repo.url)*/
  var db2 = randomAccessIdb('dats')
  var dat2 = new Dat()
  var repo2 = dat2.get(localStorage.getItem('My_Repo'), {
    db: db2
  })

  repo2.archive.readFile('/example.txt', 'utf-8', (err, data) => {
    console.log(`It's still there: ${data}`)
  })
//})
</script>

Hi, I'm getting the following error with dat-js

Version: 8

 ERROR  Failed to compile with 1 errors                               8:14:59 PM

 error  in ./node_modules/@sammacbeth/random-access-idb-mutable-file/src/random-access-idb-mutable-file.js

Module parse failed: Unexpected token (4:12)
You may need an appropriate loader to handle this file type.
| 
| import RandomAccess from "random-access-storage"
> import type {
|   Stat,
|   Request,

 @ ./node_modules/random-access-web/index.js 38:24-77
 @ ./node_modules/dat-js/index.js
 @ ./src/boot/dat.js
 @ ./.quasar/client-entry.js
 @ multi (webpack)-dev-server/client?http://0.0.0.0:8080 (webpack)/hot/dev-server.js ./.quasar/client-entry.js

package.json

  "name": "qdat",
  "version": "0.0.1",
  "scripts": {
    "test": "echo \"No test specified\" && exit 0"
  },
  "dependencies": {
    "@quasar/extras": "^1.0.0",
    "dat-js": "^8.0.0",
    "quasar": "^1.0.0-beta.0"
  },
  "devDependencies": {
    "@quasar/app": "^1.0.0-beta.0",
    "devtron": "1.4.0",
    "electron": "4.0.5",
    "electron-debug": "2.1.0",
    "electron-devtools-installer": "2.2.4",
    "strip-ansi": "=3.0.1"
  },
  "engines": {
    "node": ">= 8.9.0",
    "npm": ">= 5.6.0",
    "yarn": ">= 1.6.0"
  },
  "browserslist": [
    "last 1 version, not dead, ie >= 11"
  ],
  "resolutions": {
    "ajv": "6.8.1"
  }
}

I don't get the same issue with dat-js v 7.1.1

This issue is to suggest removing Dat class and to expose Repo directly.

Arguments I see for keeping it:

1. To make it possible to push default opts that apply to each newly-created Repo.
2. Having one close function.
3. dat-js is meant to be this one interface for the developer that has everything, such as this one close.

Arguments I see for removing it:

1. (1) could be dangerous as turned out in #20.
2. For (2), the design style of the dat ecosystem seems to prefer doing just one thing really well in each module/npm library.
3. For (3), actually, the developer ends up mostly using Repo APIs anyway. create and get could be static methods on Repo (like DatArchive).
4. As long as there's no shared state/variables between the repos, I think providing Dat adds complexity, and may actually encourage us to add shared state/variables when there could've been none.

So, my opinion would be to remove it.

What do you think?

dat-js is currently making use of a gateway and signalhubws instance that I set up on a server I have lying around.

I'm totally OK with this setup, but I think that it would be better going forward if it was using instances that were under the datproject domain to make it more obvious to users what it is and where to go if it's down. (Also in case I get run over by a bus).

Steps to get this done:

• Register subdomains on datproject
• Set up a VM with the services
• Set up a letsencrypt certificate with auto-renewal (I did this with nginx)
• Update dat-js and docs to point to the new domains

cc: @Karissa @joehand

With [email protected], now when I go to do

var archive = dat.create({
   persist: true
})

I get an uncaught TypeError:

Code on https://glitch.com/~dat-js-disco-penalty

i plan build a app with dat-js ,i know dat-node ,but most user not know node ,and i hope user just open and use it ,so,,just hope enhance dat-js

There's a lot of vulnerabilities that needs to be fixed, mostly related to zuul that hasn't been maintained for two years.

recently we started using the cache instead of indexeddb to store large images, the performance is much better https://github.com/digidem/cache-blob-store

wonder if we could create a random-access-cache (or if one exists) for that..

Hi, I'm noticing an issue with how connected peers are being managed. I'm attempting to log the number of connected peers for a given archive, and am getting some unexpected results. It looks like peers are being added twice as duplicates. Furthermore, they are only removed once, leaving a copy still in the peer list. The following reproducible code snippet illustrates what I'm seeing:

const dat1 = new Dat()
const dat2 = new Dat()

const archive = dat1.create()

archive.ready(() => {
  archive.metadata.on('peer-add', onPeerAdded)
  archive.metadata.on('peer-remove', onPeerRemoved)
  
  dat2.get(archive.url) // trigger peer-add event
})

// Here's the weird thing. This handler gets called twice when dat2 joins.
// The first time it'll log an array of one peer as expected.
// The second time it'll log an array of two duplicate peers
function onPeerAdded() {
  console.log('peers after adding', archive.metadata.peers)

  dat2.close() // trigger peer-remove event
}

// Only called once as expected, however the log will show that only one
// of the duplicate peers is removed. The other remains in the list
function onPeerRemoved() {
  console.log('peers after removing', archive.metadata.peers)
}

Here is a codesandbox running the snippet: https://codesandbox.io/s/nervous-turing-wppi3

Please let me know if I can be of any more help.
As an aside, this project is super cool, and I am excited to see it mature.

The dat-node package has a sample for dat creation and replication. It would be nice if this package had something similar.

Can you include an example? Thanks!

I'm actually really liking this API - I feel it makes sense as it turns dat into a manager of multiple hyperdrive archives that are exposed over hyperdiscovery. Simple and to the point.

I think the only change I'd make would be to rename dat.add to dat.create / dat.createRepo as it's closer to the hyperdrive terminology of drive.createArchive. - I'd use this repo 😁

Not sure how crucial this is (everything still seems to work), but when I create a new Dat

const dat = new Dat({
          persist: true, 
          gateway:  'ws://gateway.mauve.moe:3000'
})

I keep getting a websocket error in the browser Dev Console:

WebSocket connection to 'ws://localhost:3472/' failed: 
Error in connection establishment: 
net::ERR_CONNECTION_REFUSED

Tried to dig through the code to find where this is called, but can't seem to find it. I'm using Glitch.com, but not sure why the Dat-JS is calling localhost for a websocket.

hi, test exmlpe for dat-js and understend why not work.

test.html have code:

<script type="text/javascript" src="https://bundle.run/dat-js@7"></script>
<script>
var dat= new datJs({
  gateway: 'ws://gateway.mauve.moe:3000'
})
var url = 'dat://13f1d1ba5be8723edb4b9499c97a16fbb061d5fc72afe5e7e107d020cb7b34f5'
var repo = dat.get(url)
repo.ready(() => {
  console.log('ready')
  repo.archive.readFile('/1.jpg', function (err, data) {
    console.log(data.toString(), err) // prints 'world'
  })
  var readStream = repo.archive.createReadStream('1.jpg')
  readStream.on('data', console.log)
})

console.log(repo.archive)
</script>

in console chrome no error logs. event .ready succesfull emited, but readFile callback no.

At computer succesfull start

Exiting the Dat program...nemozar@nemozar-PC:~/Pictures/test$ dat share ./
dat v13.12.2
dat://13f1d1ba5be8723edb4b9499c97a16fbb061d5fc72afe5e7e107d020cb7b34f5

I followed the example on https://docs.dat.foundation/docs/browser-intro but i get TypeError: path is undefined, using this code:

var dat = new Dat()
var archive = dat.create({
    persist: true
})
archive.writeFile('/example.txt', 'Hello World!', () => {
    localStorage.setItem('My_Repo', archive.url)
})

Removing .posix in hyperdrive seemed to do the trick but I'm not sure if anything else needs to be updated as well.

I get this error when importing node-js like this: import Dat from 'dat-js'. I am using this in a React app.