The websockets server example works as expected. On browser refresh (e.g. S-F5 with chrome), the websocket disconnects, still working as expected. After refresh, the user has to give name again to connect to the server.
How would you capture the refresh-event and keep the user connected? E.g.
Is this doable only on server side or does the client require modifications as well? Haskell examples or links to such would be nice as well as hints on how to do this!
How would you capture the refresh-event...
There isn't really such a thing as a refresh event to detect (I would love to be proved wrong in this!)
... and keep the user connected...
The refresh, or rather, the leaving of the page before loading it again, causes the websocket to disconnect, and (especially if this is the only page on the site that is open), there isn't really much you can do about it.
So the only thing that can be done, is have some sort of auto-reconnect the next time the page loads. A solution that allows this is one where..
when the name is initially entered, the name is saved somewhere in the browser;
when the page reloads, it checks for a previously saved name;
and if it's found, it connects again using that name.
Local storage is one such place to save this, as in the below example, modified from https://github.com/jaspervdj/websockets/tree/master/example to save/retrieve the name from local storage.
$(document).ready(function () {
var savedUser = sessionStorage.getItem("rejoin-user");
if (savedUser) {
joinChat(savedUser);
}
$('#join-form').submit(function () {
joinChat($('#user').val())
});
function joinChat(user) {
sessionStorage.setItem("rejoin-user", user);
$('#warnings').html('');
var ws = createChatSocket();
ws.onopen = function() {
ws.send('Hi! I am ' + user);
};
ws.onmessage = function(event) {
if(event.data.match('^Welcome! Users: ')) {
/* Calculate the list of initial users */
var str = event.data.replace(/^Welcome! Users: /, '');
if(str != "") {
users = str.split(", ");
refreshUsers();
}
$('#join-section').hide();
$('#chat-section').show();
$('#users-section').show();
ws.onmessage = onMessage;
$('#message-form').submit(function () {
var text = $('#text').val();
ws.send(text);
$('#text').val('');
return false;
});
} else {
$('#warnings').append(event.data);
ws.close();
}
};
$('#join').append('Connecting...');
return false;
};
});
... Is this doable only on server side or does the client require modifications as well?
It definitely needs something done in the client to auto-reconnect. The bare bones version above needs no changes to the server, but if you wanted something fancier, like having the cases of initial connect and auto reconnect handled/shown differently somehow, then the server might need to be modified.
I know that client-side Javascript cannot write data to the local filesystem, for obvious security reasons.
The only way to save data locally with Javascript seems to be with cookies, localStorage, or allow the user to download a file (with a "Save..." dialog box or to the browser's default Download folder).
But is it possible, in the very specific case when the file is accessed locally with an URL like file:///D:/test/index.html (and not through internet) to write data locally ? (without any server language, and even without any server at all: just local browsing of a HTML file)
For example, would it be possible, by clicking on SAVE here:
<div contenteditable="true" style="height:200px;">Content editable - edit me and save!</div>
<button>Save</button>
...that such a HTML file (accessed through file:///D:/test/index.html) is overwritten with its new content ? (i.e. the local HTML file should be updated when SAVE is pressed).
TL;DR: Is this possible to SAVE a file thanks to Javascript, when the HTML page is accessed locally?
Note: I want to be able to silently save, not propose a Download/Save dialog box in which the user has to choose where to download, then "Are you sure to want to overwrite" etc.
EDIT : Why this question? I'm doing an in-browser notepad that I can run locally without any server (no Apache, no PHP). I need to be able to save easily without having to deal with Dialog Box "Where do you want to download the file?" and having to always re-browse to the same folder to overwrite the currently-being-edited file. I would like a simple UX like in any notepad program: CTRL+S done, the current file is saved! (example: MS Word doesn't ask to browse where you want to save the file each time you do "Save": CTRL+S, done!)
You can just use the Blob function:
function save() {
var htmlContent = ["your-content-here"];
var bl = new Blob(htmlContent, {type: "text/html"});
var a = document.createElement("a");
a.href = URL.createObjectURL(bl);
a.download = "your-download-name-here.html";
a.hidden = true;
document.body.appendChild(a);
a.innerHTML = "something random - nobody will see this, it doesn't matter what you put here";
a.click();
}
and your file will save.
The canonical answer, from the W3C File API Standard:
User agents should provide an API exposed to script that exposes the features above. The user is notified by UI anytime interaction with the file system takes place, giving the user full ability to cancel or abort the transaction. The user is notified of any file selections, and can cancel these. No invocations to these APIs occur silently without user intervention.
Basically, because of security settings, any time you download a file, the browser will make sure the user actually wants to save the file. Browsers don't really differentiate JavaScript on your computer and JavaScript from a web server. The only difference is how the browser accesses the file, so storing the page locally will not make a difference.
Workarounds:
However, you could just store the innerHTML of the <div> in a cookie. When the user gets back, you can load it back from the cookie. Although it isn't exactly saving the file to the user's computer, it should have the same effect as overwriting the file. When the user gets back, they will see what they entered the last time. The disadvantage is that, if the user clears their website data, their information will be lost. Since ignoring a user's request to clear local storage is also a security problem, there really is no way around it.
However, you could also do the following:
Use a Java applet
Use some other kind of applet
Create a desktop (non-Web based) application
Just remember to save the file when you clear your website data. You can create an alert that pops up and reminds you, or even opens the save window for you, when you exit the page.
Using cookies: You can use JavaScript cookies on a local page. Just put this in a file and open it in your browser:
<!DOCTYPE html>
<html>
<head>
</head>
<body>
<p id="timesVisited"></p>
<script type="text/javascript">
var timesVisited = parseInt(document.cookie.split("=")[1]);
if (isNaN(timesVisited)) timesVisited = 0;
timesVisited++;
document.cookie = "timesVisited=" + timesVisited;
document.getElementById("timesVisited").innerHTML = "You ran this snippet " + timesVisited + " times.";
</script>
</body>
</html>
Chromium's File System Access API (introduced in 2019)
There's a relatively new, non-standard File System Access API (not to be confused with the earlier File and Directory Entries API or the File System API). It looks like it was introduced in 2019/2020 in Chromium/Chrome, and doesn't have support in Firefox or Safari.
When using this API, a locally opened page can open/save other local files and use the files' data in the page. It does require initial permission to save, but while the user is on the page, subsequent saves of specific files do so 'silently'. A user can also grant permission to a specific directory, in which subsequent reads and writes to that directory don't require approval. Approval is needed again after the user closes all the tabs to the web page and reopens the page.
You can read more about this newish API at https://web.dev/file-system-access/. It's meant to be used to make more powerful web applications.
A few things to note about it:
By default, it requires a secure context to run. Running it on https, localhost, or through file:// should work.
You can get a file handle from dragging and dropping a file by using DataTransferItem.getAsFileSystemHandle
Initially reading or saving a file requires user approval and can only be initiated via a user interaction. After that, subsequent reads and saves don't need approval, until the site is opened again.
Handles to files can be saved in the page (so if you were editing 'path/to/file.html', and reload the page, it would be able to have a reference to the file). They can't seemingly be stringified, so are stored through something like IndexedDB (see this answer for more info). Using stored handles to read/write requires user interaction and user approval.
Here are some simple examples. They don't seem to run in a cross-domain iframe, so you probably need to save them as an html file and open them up in Chrome/Chromium.
Opening and Saving, with Drag and Drop (no external libraries):
<body>
<div><button id="open">Open</button><button id="save">Save</button></div>
<textarea id="editor" rows=10 cols=40></textarea>
<script>
let openButton = document.getElementById('open');
let saveButton = document.getElementById('save');
let editor = document.getElementById('editor');
let fileHandle;
async function openFile() {
try {
[fileHandle] = await window.showOpenFilePicker();
await restoreFromFile(fileHandle);
} catch (e) {
// might be user canceled
}
}
async function restoreFromFile() {
let file = await fileHandle.getFile();
let text = await file.text();
editor.value = text;
}
async function saveFile() {
var saveValue = editor.value;
if (!fileHandle) {
try {
fileHandle = await window.showSaveFilePicker();
} catch (e) {
// might be user canceled
}
}
if (!fileHandle || !await verifyPermissions(fileHandle)) {
return;
}
let writableStream = await fileHandle.createWritable();
await writableStream.write(saveValue);
await writableStream.close();
}
async function verifyPermissions(handle) {
if (await handle.queryPermission({ mode: 'readwrite' }) === 'granted') {
return true;
}
if (await handle.requestPermission({ mode: 'readwrite' }) === 'granted') {
return true;
}
return false;
}
document.body.addEventListener('dragover', function (e) {
e.preventDefault();
});
document.body.addEventListener('drop', async function (e) {
e.preventDefault();
for (const item of e.dataTransfer.items) {
if (item.kind === 'file') {
let entry = await item.getAsFileSystemHandle();
if (entry.kind === 'file') {
fileHandle = entry;
restoreFromFile();
} else if (entry.kind === 'directory') {
// handle directory
}
}
}
});
openButton.addEventListener('click', openFile);
saveButton.addEventListener('click', saveFile);
</script>
</body>
Storing and Retrieving a File Handle using idb-keyval:
Storing file handles can be tricky, since they can't be unstringified, though apparently they can be used with IndexedDB and mostly with history.state. For this example we'll use idb-keyval to access IndexedDB to store a file handle. To see it work, open or save a file, and then reload the page and press the 'Restore' button. This example uses some code from https://stackoverflow.com/a/65938910/.
<body>
<script src="https://unpkg.com/idb-keyval#6.1.0/dist/umd.js"></script>
<div><button id="restore" style="display:none">Restore</button><button id="open">Open</button><button id="save">Save</button></div>
<textarea id="editor" rows=10 cols=40></textarea>
<script>
let restoreButton = document.getElementById('restore');
let openButton = document.getElementById('open');
let saveButton = document.getElementById('save');
let editor = document.getElementById('editor');
let fileHandle;
async function openFile() {
try {
[fileHandle] = await window.showOpenFilePicker();
await restoreFromFile(fileHandle);
} catch (e) {
// might be user canceled
}
}
async function restoreFromFile() {
let file = await fileHandle.getFile();
let text = await file.text();
await idbKeyval.set('file', fileHandle);
editor.value = text;
restoreButton.style.display = 'none';
}
async function saveFile() {
var saveValue = editor.value;
if (!fileHandle) {
try {
fileHandle = await window.showSaveFilePicker();
await idbKeyval.set('file', fileHandle);
} catch (e) {
// might be user canceled
}
}
if (!fileHandle || !await verifyPermissions(fileHandle)) {
return;
}
let writableStream = await fileHandle.createWritable();
await writableStream.write(saveValue);
await writableStream.close();
restoreButton.style.display = 'none';
}
async function verifyPermissions(handle) {
if (await handle.queryPermission({ mode: 'readwrite' }) === 'granted') {
return true;
}
if (await handle.requestPermission({ mode: 'readwrite' }) === 'granted') {
return true;
}
return false;
}
async function init() {
var previousFileHandle = await idbKeyval.get('file');
if (previousFileHandle) {
restoreButton.style.display = 'inline-block';
restoreButton.addEventListener('click', async function (e) {
if (await verifyPermissions(previousFileHandle)) {
fileHandle = previousFileHandle;
await restoreFromFile();
}
});
}
document.body.addEventListener('dragover', function (e) {
e.preventDefault();
});
document.body.addEventListener('drop', async function (e) {
e.preventDefault();
for (const item of e.dataTransfer.items) {
console.log(item);
if (item.kind === 'file') {
let entry = await item.getAsFileSystemHandle();
if (entry.kind === 'file') {
fileHandle = entry;
restoreFromFile();
} else if (entry.kind === 'directory') {
// handle directory
}
}
}
});
openButton.addEventListener('click', openFile);
saveButton.addEventListener('click', saveFile);
}
init();
</script>
</body>
Additional Notes
Firefox and Safari support seems to be unlikely, at least in the near term. See https://github.com/mozilla/standards-positions/issues/154 and https://lists.webkit.org/pipermail/webkit-dev/2020-August/031362.html
Yes, it's possible.
In your example, you are already using ContentEditable and most of tutorials for that attribute have some sort of localStrorage example, ie. http://www.html5tuts.co.uk/demos/localstorage/
On page load, script should check localStorage for data and if true, populate element. Any changes in content could be saved in localStorage when clicking save button (or automatically, in linked example, using blur and focus). Additionally you can use this snippet to check weather user is online or offline and based on state modify your logic:
// check if online/offline
// http://www.kirupa.com/html5/check_if_internet_connection_exists_in_javascript.htm
function doesConnectionExist() {
var xhr = new XMLHttpRequest();
var file = "http://www.yoursite.com/somefile.png";
var randomNum = Math.round(Math.random() * 10000);
xhr.open('HEAD', file + "?rand=" + randomNum, false);
try {
xhr.send();
if (xhr.status >= 200 && xhr.status < 304) {
return true;
} else {
return false;
}
} catch (e) {
return false;
}
}
EDIT: More advance version of localStorage is Mozilla localForage which allows storing other types of data besides strings.
You could save files, and make it persistent using the FileSystem-API and webkit. You would have to use a chrome browser and it is not a standards technology but I think it does exactly what you want it to do. Here is a great tutorial to show how to do just that http://www.noupe.com/design/html5-filesystem-api-create-files-store-locally-using-javascript-webkit.html
And to show that its on topic it starts off showing you how to make the file save persistent...
window.webkitRequestFileSystem(window.PERSISTENT , 1024*1024, SaveDatFileBro);
Convert your HTML content to a data uri string, and set as href attribute of an anchor element. Don't forget to specify a filename as download attribute.
Here's a simple example:
<a>click to download</a>
<script>
var anchor = document.querySelector('a');
anchor.setAttribute('download', 'example.html');
anchor.setAttribute('href', 'data:text/html;charset=UTF-8,<p>asdf</p>');
</script>
Just try it in your browser, no server required.
Have a look into this :)
Download File Using Javascript/jQuery
there should be everything you need. If you still need help or it's not the solution you need, tell me ;)
Yes, it is possible. Proof by example:
TiddlyFox: allows modification of local files via an add-on. (source code) (extension page):
TiddlyFox is an extension for Mozilla Firefox that enables TiddlyWiki
to save changes directly to the file system.
Todo.html: An HTML file that saves edits to itself. Currently, it only works in Internet Explorer and you have to confirm some security dialogs when first opening the file. (source code) (functional demo).
Steps to confirm todo.html actually saves changes to itself locally:
Save todo.html to local harddrive
Open with Internet Explorer. Accept all the security dialogs.
Type command todo add TEST (todo.html emulates the command-line interface of todo.txt-CLI)
Inspect todo.html file for addition of 'TEST'
Caveats: there is no cross-platform method. I'm not sure how much longer these methods will exist. When I first started my todo.html project, there was a jQuery plugin called twFile that allowed cross-browser loading/saving of local files using four different methods (ActiveX, Mozilla XUL, Java applet, Java Live Connect). Except for ActiveX, browsers have disallowed all these methods due to security concerns.
If you are fine with your code running outside of the scope of your default browser, and you are fine with windows only support, HTAs meet the silently save without prompts requirement easily.
The below code doesn't use many HTA specific features but it does still use microsoft specific stuff like ActiveXObject("Scripting.FileSystemObject").
<html>
<head>
<title>Simple Notepad</title>
<meta http-equiv="X-UA-Compatible" content="IE=9">
<script>
document.addEventListener('keydown', function (event) {
if (event.ctrlKey) {
if (event.key == 's') {
var FSo = new ActiveXObject("Scripting.FileSystemObject");
//see https://learn.microsoft.com/en-us/office/vba/language/reference/user-interface-help/opentextfile-method
var thisFile = FSo.OpenTextFile(window.location.pathname, 2, true, -1);
thisFile.Write(document.getElementsByTagName("html")[0].outerHTML);
thisFile.Close();
// Comment out the below alert to get truly silent saving.
alert('Saved Successfully');
if (event.preventDefault) event.preventDefault();
return false;
}
}
}, false);
</script>
</head>
<body contentEditable="true">
<h1>Press <kbd>CTRL + S</kbd> To Save</h1>
</body>
</html>
It also isn't a very rich editing experience but that can be fixed with some more buttons or keyboard shortcuts I think. Like CTRL + B to embolden selected text. It doesn't have any safety checks as of yet, but binding an event handler to beforeunload should prevent any data loss caused by accidentally closing the program.
HTA's do have other disadvantages too. They don't support ES6 (though transpiling is an option).
Although it is a bit dated, If you're not trying to use modern web features, I think you'll agree that it is very functional and usable.
Edit
I forgot to mention, but HTAs have to be saved with the .hta file extension for mshta.exe to be registered as their file type handler. Which is needed so that you can double click it in windows explorer to open it easily.
See also
Introduction to HTML Applications on MSDN
HTML Applications reference on MSDN
I think it's important to clarify the difference between server and client in this context.
Client/server is a program relationship in which one program (the client) requests a service or resource from another program (the server).
Source: http://searchnetworking.techtarget.com/definition/client-server
I'm not sure you'll find too many advanced applications that don't have at least one server/client relationship. It is somewhat misleading to ask to achieve this without any server, because any time your program speaks to another program, it is a client/server relationship, with the requester being the client and the response coming from the server. This is even if you are working locally. When you want to do something outside of the scope of the browser, you need a hook in a server.
Now, that does not mean you can't achieve this without a server-side specific language. For example, this solution uses NodeJS for the server. WinJS has WinJS.xhr, which uses XmlHttpRequest to serve data to the server.
AJAX seeks to offer the same sort of functions. The point here is that whether you have a program or there is some sort of hook pre-built, when you issue a command like "save file" and the file actually gets saved, there is a program on the other side that is parsing it, whether it's a server-side language or something else, meaning you can't possibly have something like this function without a server to receive the request.
Just use https://github.com/firebase/firepad — See it in action
This doesn’t need a server on your computer, it will reach out and save the data remotely.
Use jsPDF -> https://github.com/MrRio/jsPDF
<div id="content">
<h3>Hello, this is a H3 tag</h3>
<p>a pararaph</p>
</div>
<div id="editor"></div>
<button id="cmd">generate PDF</button>
Javascript
var doc = new jsPDF();
var specialElementHandlers = {
'#editor': function (element, renderer) {
return true;
}
};
$('#cmd').click(function () {
doc.fromHTML($('#content').html(), 15, 15, {
'width': 170,
'elementHandlers': specialElementHandlers
});
doc.save('sample-file.pdf');
});
This is an example for those who want to know how to use the localStorage.
<div id="divInput" contenteditable="true" style="height:200px;border: 2px solid blue">
Content editable - edit me and save!
</div>
<button onclick="onSave()">Save</button>
<button onclick="onLoad()">Load</button>
<script>
config = {
localStorageItemName: "demo",
datetimeFormat: {
year: 'numeric',
month: '2-digit',
day: '2-digit',
hour: '2-digit',
hour12: false,
minute: '2-digit',
second: '2-digit'
}
}
function Now() {
return new Date().toLocaleString("zh-TW", config.datetimeFormat)
}
const errMap = {
IsEmptyError: new Error('is empty'),
LengthError: new Error('length = 0')
}
/**
* #param {string} itemName
* #return {Object}
* */
function getLocalStorageItem(itemName) {
const dbDataString = localStorage.getItem(itemName)
if (dbDataString === null) {
throw errMap.IsEmptyError
}
const db = JSON.parse(dbDataString)
if (Object.keys(db).length === 0) {
throw errMap.LengthError
}
return db
}
function onSave() {
const inputValue = document.querySelector(`#divInput`).textContent
try {
const db = getLocalStorageItem(config.localStorageItemName)
db.msg = inputValue
db.lastModTime = Now()
localStorage.setItem(config.localStorageItemName, JSON.stringify(db))
console.log("save OK!")
} catch (err) {
switch (err) {
case errMap.IsEmptyError:
console.info("new localStorageItemName")
localStorage.setItem(config.localStorageItemName,
JSON.stringify({msg: inputValue, createTime: Now()})
)
break
/*
case ...
break
*/
default:
console.error(err.message)
}
}
}
function onLoad(e) {
try {
const db = getLocalStorageItem(config.localStorageItemName)
console.log("load")
document.querySelector(`#divInput`).textContent = db.msg
} catch (err) {
return
}
}
(()=>{
window.onload = () => (
onLoad()
)
})()
</script>
It is written in pure javascript with no dependencies.
I was asked to take a look at what should be a simple problem with one of our web pages for a small dashboard web app. This app just shows some basic state info for underlying backend apps which I work heavily on. The issues is as follows:
On a page where a user can input parameters and request to view a report with the given user input, a button invokes a JS function which opens a new page in the browser to show the rendered report. The code looks like this:
$('#btnShowReport').click(function () {
document.getElementById("Error").innerHTML = "";
var exists = CheckSession();
if (exists) {
window.open('<%=Url.Content("~/Reports/Launch.aspx?Report=Short&Area=1") %>');
}
});
The page that is then opened has the following code which is called from Page_Load:
rptViewer.ProcessingMode = ProcessingMode.Remote
rptViewer.AsyncRendering = True
rptViewer.ServerReport.Timeout = CInt(WebConfigurationManager.AppSettings("ReportTimeout")) * 60000
rptViewer.ServerReport.ReportServerUrl = New Uri(My.Settings.ReportURL)
rptViewer.ServerReport.ReportPath = "/" & My.Settings.ReportPath & "/" & Request("Report")
'Set the report to use the credentials from web.config
rptViewer.ServerReport.ReportServerCredentials = New SQLReportCredentials(My.Settings.ReportServerUser, My.Settings.ReportServerPassword, My.Settings.ReportServerDomain)
Dim myCredentials As New Microsoft.Reporting.WebForms.DataSourceCredentials
myCredentials.Name = My.Settings.ReportDataSource
myCredentials.UserId = My.Settings.DatabaseUser
myCredentials.Password = My.Settings.DatabasePassword
rptViewer.ServerReport.SetDataSourceCredentials(New Microsoft.Reporting.WebForms.DataSourceCredentials(0) {myCredentials})
rptViewer.ServerReport.SetParameters(parameters)
rptViewer.ServerReport.Refresh()
I have omitted some code which builds up the parameters for the report, but I doubt any of that is relevant.
The problem is that, when the user clicks the show report button, and this new page opens up, depending on the types of parameters they use the report could take quite some time to render, and in the mean time, the original page becomes completely unresponsive. The moment the report page actually renders, the main page begins functioning again. Where should I start (google keywords, ReportViewer properties, etc) if I want to fix this behavior such that the other page can load asynchronously without affecting the main page?
Edit -
I tried doing the follow, which was in a linked answer in a comment here:
$.ajax({
context: document.body,
async: true, //NOTE THIS
success: function () {
window.open(Address);
}
});
this replaced the window.open call. This seems to work, but when I check out the documentation, trying to understand what this is doing I found this:
The .context property was deprecated in jQuery 1.10 and is only maintained to the extent needed for supporting .live() in the jQuery Migrate plugin. It may be removed without notice in a future version.
I removed the context property entirely and it didnt seem to affect the code at all... Is it ok to use this ajax call in this way to open up the other window, or is there a better approach?
Using a timeout should open the window without blocking your main page
$('#btnShowReport').click(function () {
document.getElementById("Error").innerHTML = "";
var exists = CheckSession();
if (exists) {
setTimeout(function() {
window.open('<%=Url.Content("~/Reports/Launch.aspx?Report=Short&Area=1") %>');
}, 0);
}
});
This is a long shot, but have you tried opening the window with a blank URL first, and subsequently changing the location?
$("#btnShowReport").click(function(){
If (CheckSession()) {
var pop = window.open ('', 'showReport');
pop = window.open ('<%=Url.Content("~/Reports/Launch.aspx?Report=Short&Area=1") %>', 'showReport');
}
})
use
`$('#btnShowReport').click(function () {
document.getElementById("Error").innerHTML = "";
var exists = CheckSession();
if (exists) {
window.location.href='<%=Url.Content("~/Reports/Launch.aspx?Report=Short&Area=1") %>';
}
});`
it will work.
I have written a simple download manager for Windows and I would like to create an addon for Firefox that when enabled intercepts new downloads in Firefox and sends them to the download manager.
I have already done this for Google Chrome using:
chrome.downloads.onCreated.addListener(function(details) {
// stop the download
chrome.downloads.cancel(details.id, null);
}
The question is how can I achieve something similar using the Firefox add-on SDK.
I see there is a way of intercepting page loads to view the content / headers which might be helpful but then I won't know if the request will turn into a download or not.
Firefox add-on SDK: Get http response headers
I could perhaps look for a content type that is not text/html or check for a content disposition header but that could cause problems if I don't correctly handle all cases.
Is there no way of accessing the download manager using the JS SDK or some way of knowing when a download has been started / being started and stop it?
The http-on-examine-response observer that the linked question discusses is the wrong way to go. It concerns all requests not just downloads.
Instead use the Downloads.jsm to observe new downloads, then cancel them, and so on.
To load Downloads.jsm in the SDK use:
const {Cu} = require("chrome");
Cu.import("resource://gre/modules/Downloads.jsm");
Cu.import("resource://gre/modules/Task.jsm");
Then you can add your listener.
let view = {
onDownloadAdded: function(download) {
console.log("Added", download);
},
onDownloadChanged: function(download) {
console.log("Changed", download);
},
onDownloadRemoved: function(download) {
console.log("Removed", download);
}
};
Task.spawn(function() {
try {
let list = yield Downloads.getList(Downloads.ALL);
yield list.addView(view);
} catch (ex) {
console.error(ex);
}
});
The linked MDN docs have more information and samples.
Since your add-on is a restartless SDK add-on, you'll need to remove the listener again using .removeView on unload, or else there will be a memory leak.
Here's the JSM way.
Components.utils.import("resource://gre/modules/Downloads.jsm");
Components.utils.import("resource://gre/modules/Task.jsm");
Components.utils.import("resource://gre/modules/FileUtils.jsm");
var view = {
onDownloadChanged: function (download) {
console.log(download, 'Changed');
if (download.succeeded) {
var file = new FileUtils.File(this.target.path);
console.log('file', file);
}
}
};
var list;
Task.spawn(function () {
list = yield Downloads.getList(Downloads.ALL);
list.addView(view);
}).then(null, Components.utils.reportError);
Remember to removeView to stop listening. Can do this anywhere, like in shutdown function or whatever, doesn't have to be within that Task.spawn so list must be global var.
list.removeView(view); //to stop listening
Here's the old way, which seems to still work. Although I thought they said they're going to take out the old downloadManager:
var observerService = Components.classes["#mozilla.org/download-manager;1"].getService(Components.interfaces.nsIDownloadManager);
observerService.addListener({
onDownloadStateChange: function (state, dl) {
console.log('dl=', dl);
console.log('state=', state);
console.log('targetFile', dl.targetFile);
if (state == 7 && dl.targetFile.leafName.substr(-4) == ".txt") {
//guys just downloaded (succesfully) a .txt file
}
}
});
Heres a mozillazine with some more on this: http://forums.mozillazine.org/viewtopic.php?f=19&t=2792021
I'm sort of shooting in the dark here; I have no knowledge how to do this so some pointers and/or links to helpful tutorials would be great:
I have a website that I want to display a text file (server log). Probably embedded. The problem is, this file is updated whenever events happen in the server (faster than half a second usually). How can I make it so the webpage displays the file in real time, meaning showing a live feed of the file?
My guess is that it would use javascript and AJAX but my knowledge on both are pretty limited.
Any pointers and help would be appreciated :)
My answer uses PHP and Ajax though changing to ASP or any other language wont be hard.
In the head
<script type="text/javascript">
function Ajax()
{
var
$http,
$self = arguments.callee;
if (window.XMLHttpRequest) {
$http = new XMLHttpRequest();
} else if (window.ActiveXObject) {
try {
$http = new ActiveXObject('Msxml2.XMLHTTP');
} catch(e) {
$http = new ActiveXObject('Microsoft.XMLHTTP');
}
}
if ($http) {
$http.onreadystatechange = function()
{
if (/4|^complete$/.test($http.readyState)) {
document.getElementById('ReloadThis').innerHTML = $http.responseText;
setTimeout(function(){$self();}, 1000);
}
};
$http.open('GET', 'loadtxt.php' + '?' + new Date().getTime(), true);
$http.send(null);
}
}
</script>
In the Body
<script type="text/javascript">
setTimeout(function() {Ajax();}, 1000);
</script>
<div id="ReloadThis">Default text</div>
</body>
Now using loadtxt.php read the values of the text file
<?php
$file = "error.txt";
$f = fopen($file, "r");
while ( $line = fgets($f, 1000) ) {
print $line;
}
?>
Using jQuery, you could do the following
setInterval(function() {
$('#element').load('/url/to/file');
}, 1000);
Would refresh the div with ID element with the file contents every 1 second
You could use jQuery .get to get the file every few seconds and update the page to show the contents.
Others have talked about loading the log file every refresh but depending on the size of the file this migth be a bad idea. You might want to create a server side page that will read the log file and keep track of how much of it has already been given to you and only give you the new bits. If its a 10k file it would be annoying (and potentially laggy) to have this transferred to you every second.
Otherwise other people seem to have covered most of the client side stuff.
There are various ways of doing this...
You could look into long polling.
Stick a meta refresh tag to refresh the page every X seconds.
tail -f /path/to/log.log in terminal will open a live preview of the last few lines of that file - this is what I do if I need to read the error logs as I debug.
Or simply refresh the page manually as you go, it might be annoying having the page change it's contents automatically.
As you have said your file is very large, I would use the PHP file() function to just grab the first X amount of lines from a file to keep bandwith down and readability up!
use this
setInterval(function() {
jQuery.get('file.txt', function(data) {
alert(data);
//process text file line by line
$('#div').html(data.replace('n','
'));
});
}, 1000);
https://www.sitepoint.com/jquery-read-text-file/
Finally, yes the script and the div id="ReloadThis" work fine together ! It also works to display info from a PHP file which queries the text file so the incoming text can be formatted before being displayed in the div element.