Arggh, ignore this Issue, I was being a massive tool and had not uploaded the right files to our dev server when testing on IE. Seems to work fine on IE11. I'll delete this Issue if I find out how...

As described by RFC 2045, all US-ASCII characters except SPACE, CTLs or tspecials are allowed in the content type.

The regular expression detecting data url usage only checks for words (\w) and dashes (\-).

How to reproduce:
Use a data url with a content type containing a dot (.) character.

For example:
data:application/vnd.openxmlformats-officedocument.spreadsheetml.sheet;base64,<BASE64_ENCODED_DATA>

I don't know how to describe it properly, and neither do I have a MRE.
But it seems that if the main function is called "download", that creates an error as described in the title.

I have renamed your function to something else (e.g. "downloadit") and then it worked.

No idea what might be triggering this error (I am using OS X version 10.9.5 and Chrome Version 64.0.3282.167), but I hope this might be of help somehow.

This is a great lib for downloading, event though it is supported in modern browsers some people are in older browser for them i want to show which file is download or not download, to do so expose the limitations via isChromeSupoort,isIESupport like that variables.

We are using download.js to download CSV/TSV/etc. data from query application, see source code here: https://github.com/wikimedia/wikidata-query-gui/blob/master/wikibase/queryService/ui/App.js#L515

With 1.4.4, it worked just fine. With 1.4.6+, however, the data is downloaded broken - non-ASCII characters are mangled. Looks like some changes in 1.4.6 broke something in non-ASCII download.

The resulting files for comparison:
1.4.4: https://gist.github.com/smalyshev/e4c02c94f1433111f7100836496d7eff
1.4.6+: https://gist.github.com/smalyshev/379b46d1a0b8bec7953f26a7c4261800

I've tested in Chrome, but other browsers are reported by users to produce similar results.

Would love an NPM package for those of us that use it!

I have been working with this library for last week now. In the beginning the file download message was coming at the bottom of the browser like this

I have seen it change to opening a save as window

and Today it just opens the file in jpeg, doesn't give a window message or show the file download bottom of the file, just simply opens it in photos,

Do we know what drives that behavior?

Thanks

Hi, thanks for this awesome library. I just have a couple questions.

1.) is the download function asynchronous?
2.) is there a way notify the calling code when it is done?

My motivation for asking is that I want to close the parent window when the download starts or completes.
I know this isn't really an issue with the code, just wondering if I can make it do what I want:)

This is quite an old bug but it looks like it has not been resolved yet. Basically when I try to use your plugin to download a base64 that is bigger than several MB (from a canvas), I get "network error" instead of the proper file.
Probably you should switch to Blob if the base64 string is too long.

Hello

When i download something whatever it is the file only contains this "[object Object]"
how to solve this

I am learning javascript

thanks

Using base64 encoded content, I can download text and images. I can also download PDFs in chrome.

However, I'm getting an error when downloading the same PDF in Internet Explorer 11.

index.txt

How can I fix this?

Hi,

I have few PDFs available on a URL, and I want to allow user to download them. I tried passing PDF as URL, but it did't work.

In example available on http://pagedemos.com/yvvmxbjrwq7u/2 , I replaced code for image download from:
download( "/diff6.png" );
to
download( "www.example.com/diff6.pdf" );

but it did't work. Do we need to do any changes for PDF to get downloaded successfully?

Thanks.

Is download.js licensed under MIT, Apache, CC-BY 4.0, or CCBY2?

It is really a great library, but I have a comment on source code. The local variable names like a, _D, _z are really hard to read. I know the short name will make the code smaller, but we already have JS minify tool. So please use full name in source code!

Thank you for this great resource. I'm using it to download files returned from a request that uses a proxy and the link to the documents cannot simply be placed in an <a href>

I can successfully download all file types with the exception of the microsoft office docx and xlsx types. When attempting to open the downloaded files with MS Office I get an error message stating that they are corrupt and cannot be repaired.

I have tried including the correct mime type
application/vnd.openxmlformats-officedocument.wordprocessingml.document

AND

leaving the mime type blank (and defaulting to application/octet-stream)

Both result in corrupted file.

Any ideas how to support the download of those file types correctly?

my code:

var x = new XMLHttpRequest();
 x.open("GET", attachUrl, true);
 x.responseType = 'blob';
// with correct mime type
//x.onload = function (e) { download(x.response, name, file.ContentType); }
// without mime type
x.onload = function (e) { download(x.response, name); }
console.log(file.ContentType);
x.send();

and the error message in MS Office

Safari Version 9.0 (11601.1.56)

When using the data URI, it tries to navigate to something like application/octet-stream,... which Safari doesn't recognize

This does not seem to work on safari 9+

http://pagedemos.com/sxks39b72aqb/1
http://pagedemos.com/sxks39b72aqb/2

generate nothing

http://pagedemos.com/sxks39b72aqb/3

Failed to load resource: Frame load interrupted

How can I set a specific folder to download the file, instead of the default folder "Downloads"?

I got an Exception about saying payload.charCodeAt is not a function at line 78 when I was testing this library with an Uint8Array input.

To download from a URL, currently I have to provide a single argument, which is the URL itself. The filename will be automatically determined from the URL. If I added the second filename argument, downloadjs will treat it like downloading blob url.

I think it should be possible to still treat it like downloading from URL, and use the second argument as filename instead of determining from the URL. Something like:

download('https://example.com/1234-5678-9012.pdf', 'mine.pdf')

Hi!
This script is very useful. Thanks for sharing.

I've tried with:
download('http://i.imgur.com/7fypLpI.jpg', 'new.jpg');
download('http://i.imgur.com/7fypLpI.jpg', 'new.jpg', 'image/jpg');
download('http://i.imgur.com/7fypLpI.jpg', 'new.jpg', 'image/jpeg');

And I can not open the image because it contains errors in the header.

How to solve it?

Please help to check, this is my code:

  var fileUrl = form.vals().download;
  var fileName = fileUrl ? fileUrl.replace(/^.*[\\\/]/, '') : '';
  download(fileUrl, fileName, "application/pdf");

After download, that file can not open:

I need some help to use download(), passing the first parameter (data) to download a .docx file.
Thank you!

I've tried setting up a simple string download which works fine in Chrome and Firefox but not in Safari 7.1. After triggering the download I can see the data being inserted into the current page real quick and then it disappears.

download('This is some string', filename, 'text/plain');

I saw that for safari the image is opened in the tab and the user should save it manually which is acceptable. My problem is that when i have multiple images and i call download function many times one after the other , the confirm popup appears as many times as images i have and i have to press ok for every each of them, but the main problem is that I can see only the last image, all the others are lost.

Is it possible to hide the message of download (silent mode)!

Hi,

maybe it's a simple question, but I can't make it myself:

how has the function to look alike if I want to download an existing file?

like the filename is "image.jpg"?

thank you!

When I download an mp4 with only a single parameter all is good:
download(highestQuality);
If I add a filename parameter it downloads a text file with the URL as the contents:

downloadName = videoName + '.mp4';
download(highestQuality,downloadName);

I tried using a static string for the name and added the mimetype, but neither helped.

Am I misunderstanding something, or is this a limitation?
Thanks-
Matt

Here is a fix, taken from bpampuch/pdfmake#294:

line 82:
try {
blob = payload instanceof myBlob ?
payload :
new myBlob([payload], {type: mimeType}) ;
}
catch (e) {
// Old browser, need to use blob builder
window.BlobBuilder = window.BlobBuilder || window.WebKitBlobBuilder || window.MozBlobBuilder || window.MSBlobBuilder;
if (window.BlobBuilder) {
var bb = new BlobBuilder();
bb.append(payload);
blob = bb.getBlob(mimeType);
}
}

In the saver() function where an invisible anchor element is used to start the download, clicks are allowed to bubble up to the document root.

It would be great if the call to anchor.click() would stop clicks from propagating to the body element.

Hi,
in my app i was using just this piece of js:

var link = document.createElement("a"); link.download = dataItem.Filename; link.href = data; link.click();

where data is the data uri with mime type etc...
this is working perfectly on chrome, but not in ie
so i used your library, and it's perfect for ie
the only issue is that is not working with documents like docx, ods...

Hi do you think you could add some download progerss in % in console.log as file downloads?

Thanks for Anwsering and Best Regards

How can I specify path? For example if I want to download to my Desktop in my local computer?

Hello,
Your lib is very cool, you should publish it into the bower repository, so we will include easily in our project.

good job 🍺

For all other request cross origin issue doesn't come as cors is already implemented in node side but when i use download(url). It throws cors error?

I am using REST API @ back and it provides the byteStream of the excel file and I am trying to generate file @ client Side.
File is downloaded but the content stores as the same the server responds to the front end.

Hello

When i download something whatever it is the file only contains this "[object Object]"
how to solve this

I am learning javascript

thanks

Hi there,

Thank you for the nice work, I have been using your download script over 2 years. I recently bumped into an issue in chrome with download.js. When loading from script tag:

chrome seems to wait forever, as a result, page dependent on it will never load unless they are loaded async. Firefox seems to have a cached version in my pc. I cannot seem to connect to your website from chrome either. Are you changing files on the server or is this new behavior intended?

Thank you for your work once again,

The library breaks PDF file and produces invalid PDF which opens with blank pages.
PDF Example: http://partners.adobe.com/public/developer/en/xml/AdobeXMLFormsSamples.pdf

Code:

$.ajax({
        url: '/AdobeXMLFormsSamples.pdf',
        success: download.bind(true, "application/pdf", 'AdobeXMLFormsSamples.pdf'), 
        complete: function(){...}
    });

Tested on browsers:

• Chromium Linux Version 44.0.2403.61 (64-bit)
• Firefox Linux 38.0.5

Hi,
When using download.js, some characters became dashes.
http:// became http--- for exemple.

tested codepen with debug mod.

Thanks for your support !

Hi @rndme Any chances you could tag a release version?

I tried every single demo but neither works on Chrome on iPhone, working on solution for iPhone/iPad download and considering your lib as a potential fix.

		function dataUrlToBlob(strUrl) {
			var parts= strUrl.split(/[:;,]/),
			type= parts[1],
			decoder= parts[2] == "base64" ? atob : decodeURIComponent,
			...
		}

https://github.com/rndme/download/blob/master/download.js#L90

This one fails, it is generated with google first hit from text:

data:text/plain;charset=utf-8;base64,w4FydsOtenTFsXLFkXTDvGvDtnJmw7p0w7Nnw6lwLg==

The problem that the charset is given, so the second part is the charset rather than the encoding. You should parse the heading part of the URI instead of making presumptions on its structure.

Got an audio file coming in through ajax and when ajax receives it the file is the right size with correct headers but once its run through download it almost doubles in size and is not readable by mp3 players.

I'm trying to download a image file from a JSON object:

{"type":"Buffer","data":[255,216,255,224,0,16,74,70,7...]}

download(new Blob(buffer.data, {type: "image/jpeg"}), "image.jpeg", "image/jpeg");

but the downloaded file is not valid image. Do i have to perform previous convertion with the Buffer object? any help?

EDIT: my bad.
The `"data":[255,216,255,224,0,16,74,70,7...]}`` from the JSON object is not a actually Buffer, just a String representation of it. Changed my API to return a Base64 String of the file.

change code

if(url && url.length< 2048){ // if no filename and no mime, assume a url was passed as the only argument
			fileName = url.split("/").pop().split("?")[0];
			anchor.href = url; // assign href prop to temp anchor
		  	if(anchor.href.indexOf(url) !== -1){ // if the browser determines that it's a potentially valid url path:
        		var ajax=new XMLHttpRequest();
        		ajax.open( "GET", url, true);
        		ajax.responseType = 'blob';
        		ajax.onload= function(e){ 
				  download(e.target.response, fileName, defaultMime);
				};
        		setTimeout(function(){ ajax.send();}, 0); // allows setting custom ajax headers using the return:
			    return ajax;
			} // end if valid url?
		} // end if url?

fileName = url.split("/").pop().split("?")[0];
url = encodeURI(decodeURI(url));
anchor.href = url

can support Chinese filename

Hello everyone!

I want to customize the behaviour when the browser is Safari. Everybody knows that the Safari isn't support the a[download] attribute yet and, i want to handle the behaviour more customizable. For that, I implemented a simple solution for, just passing an option parameter to the download function but i'm in doubt about the flexibility and customizability for that solution.

I'll send a PR but discussion is opened!
Best regards.

When I used it in my project ,I got an issue as the picture says.

Safari still refuses to download... tries to open up the file in a new tab or window. Any way to workaround this? I need to be able to save a client side code generated file in blob or objectURL format to a user's downloads folder.