Touch Browser-Download of epub Files

[escape]

Im working on a standalone browser library for Kobo Touch users.
I have a big problem about the download function. When ever i open a Link to an .epub-File, the browser does not download it. It just opens the file cryptik inside the browser that freezes after that. First i thought: there is no way it could be solved. But gutenberg.org does have a solution for that. Also it works inside the calibre library that you reach through your webbrowser.

How do they get the files to download? Or easier: How do they bring up the box that asks if file should be downloadet or not?

Sorry for my bad english

[Aydan]

Have a look at the mime type sent by the server. This should control how the browser reacts.

Regards
Aydan

[escape]

<a type="application/octet-stream" href="xxx.epub">
or
<a type="application/epub+zip" href="xxx.epub">
or
<a type="application/force-download" href="xxx.epub">
did not work.

The header manipulation inside the file did also not work.

Any other hints?

[Aydan]

Check the HTTP response header of your response.
This is what project gutenberg spits out as an example:

Status: HTTP/1.1 200 OK
Server: Apache
Content-Disposition: attachment; filename=pg38934-images.epub
X-Rate-Limiter: php
Cache-Control: max-age=86400
X-Frame-Options: sameorigin
Content-Type: application/epub+zip
Content-Length: 281011
X-Powered-By: 3
X-Hits: 1
Date: Tue, 21 Feb 2012 13:48:47 GMT
X-Varnish: 1635092337 1635083511
Age: 323
Via: 1.1 varnish
Connection: close

I got this info via web-sniffer.net

Regards
Aydan

[escape]

Status: HTTP/1.1 200 OK
Date: Tue, 21 Feb 2012 13:55:45 GMT
Server: Apache/2.2.9 (Debian) PHP/5.2.6-1+lenny13 with Suhosin-Patch mod_python/3.3.1 Python/2.5.2 mod_ssl/2.2.9 OpenSSL/0.9.8g
Last-Modified: Sat, 21 Jan 2012 09:44:11 GMT
ETag: "23b000a-49800-4b706a3b3d0c0"
Accept-Ranges: bytes
Content-Length: 301056
Connection: close
Content-Type: text/plain

So thats the fault.

But how can i change it without touching the apache settings?

[Aydan]

You'll have to tell apache that epub is not text/plain.
look here

Regards
Aydan

[escape]

thanks for the answer.. so now i see that my hoster does not make the right settings for epubs.

[rashkae]

The real culprit here is the damnabe "mime magic" plugin that Apache installs by default. Without this plugin, any file with an extension that Apache doesn't recognized gets passed to the browser as binary-bin. At worst case, that means the browser will at least let you download/save the file (without requiring a right-click save link as...) I've seen that mime magic cause *all* kinds of mischief by mis-identifying files as text.

[escape]

I asked my hoster but he won't change the settings on the Apache config.

Is there any easy and working MIME-Type configuration inside the webpage possible? Without htaccess?

[voom]

If you have PHP available sth. like this should do the trick:

http://www.ryboe.com/tutorials/php-h...force-download

[escape]

The problem is:

The KoboBrowser does not open URLs which begin with header manipulation...

Dont ask me why.