DownloadStream wrongly encodes filename?

hey there,

not sure if this is really true, but I expected `DownloadStream#getContentDispositionFilename(String)` to encode spaces in filenames with `%20` rather than `+`. However, this is of course related because of using `java.net.URLEncoder` as it does a `application/x-www-form-urlencoded` which says: *Space characters are replaced by `+'*

The effect is, that at least Firefox browser suggests to download the file `file with spaces.txt` with name `file+with+spaces.txt`. 

Unfortunately, I cannot set the `Content-Disposition` parameter by hand as `StreamResource#getStream()` always returns a new instance which makes 

treamResource streamResource = new StreamResource(source, filename);
String cd = "attachment; " + DownloadStream.getContentDispositionFilename(filename);
streamResource.getStream().setParameter(DownloadStream.CONTENT_DISPOSITION, cd.replace("+", "%20"));

senseless.

Do you have any ideas?

Some links from my research:

• http://www.w3.org/TR/html401/interact/forms.html#h-17.13.4.1
• https://vaadin.com/api/com/vaadin/server/DownloadStream.html#getContentDispositionFilename(java.lang.String)
• https://vaadin.com/api/com/vaadin/server/DownloadStream.html#getContentDispositionFilename(java.lang.String)

Note: I do not have problems with this: http://kb.mozillazine.org/Filenames_with_spaces_are_truncated_upon_download

The only way I found to cope with that issue is to override the handleConnectorRequest method and in there to call my own getContentDispositionFilename-method instead of the original method of class DownloadStream.

The method looks as follows:

public String getContentDispositionFilename() { String filename = onDemandStreamResource.getFilename(); String filenameNoBlanks = filename.replace(" ", "%20"); // replace blanks with %20 for firefox, chrome and ie don't need that return String.format("filename=\"%s\"; filename*=utf-8\'\'%s", new Object[]{filename, filenameNoBlanks}); }

Instead of url encoding the filename as done in the original method here only blanks are replaced by "%20" for the "filename*=utf-8"-part. In chrome and IE it seems to work well even with blanks, however, Firefox can not read the filename if it contains blanks.