I'm building a library to "nodeize" the HTML5 File Api (currently in alpha)[0], to make it work with binary contents and don't have problems with charsets, I'm using Buffer[1] utility.
But the HTML5 File API uses Blob native object. Actualy I'm using the type 'application/octet-stream', and 'binary' from Buffer encoding. But, I want to use base64 in order to prevent any problem:
CoFS.prototype.writeFile = function (fileName, data, encoding, callback) {
var self = this;
if (Buffer.isBuffer(data)) {
callback = encoding;
encoding = undefined;
} else {
data = new Buffer(data, encoding);
}
this._ifready(function () {
self.getFileEntry(fileName, {create: true, exclusive: true}, function (err, fileEntry) {
if (err) return callback(new Error("Error getting file access " + err.message));
fileEntry.createWriter(function(fileWriter) {
fileWriter.onwriteend = function () {
callback(null);
};
fileWriter.onerror = function(e) {
callback(new Error(e.toString()));
};
var blob = new Blob([data.toString('binary')], {type: 'application/octet-stream'});
fileWriter.write(blob);
}, function () {
callback(new Error('Error writing ' + fileName));
});
});
});
};
Exacts on:
var blob = new Blob([data.toString('binary')], {type: 'application/octet-stream'});
I red the MDN page[2] but I didn't see anything about the encoding.
Is there any way to accomplish something like that?:
var blob = new Blob([data.toString('base64')], {type: 'application/octet-stream', encoding: 'base64'});
Thankyou.
I realized the Buffer object is extended from Int8Array, and both are compatible.
My code now is:
var blob = new Blob(
[data], // Data is a buffer!
{
type: 'application/octet-stream'
}
);
fileWriter.write(blob);
And for read, I use readAsArrayBuffer method, you can do:
var d = new Int8Array(arr); // When arr is arrayBuffer
var buf = new Buffer(d); // And works!
This solve my problem. For convert base64 encoded content and use in a blob, with Buffer, can be:
var blob = new Blob(
[new Buffer(cleanData, 'base64')],
{
type: 'application/octet-stream'
}
);