Resuming podcast downloads

[Bart Cerneels]

Hi Stewart,

I'm the one that wrote PodcastFetcher. The reason I'm not using KIO (which 
supports resuming) is because KIO does not report redirects to the Amarok. So 
I'm using QHttp to fetch podcastepisodes. Whether or not QHttp supports 
resuming I do not know.
When you figure that out I would suggest you create a partial file in 
PodcastFetcher::slotResponseReceived( 200 ) and rename the partial file in 
PodcastFetcher::slotDone() when completed. That way only one file is created 
and renamed, unlike the ugly hack it is now.

And then there is the conflicting filename problem Ian keeps hammering on. It 
is possible that a podcaster always names his files the same, like 
episode.mp3, in every rss-item. If that's the case a number is appended 
before the extension. Problem is that when resuming a download a new file 
will be created because the partial is already there. A way to prevent that 
might be checking the content size in the enclosure.

From personal experience I can tell podcastfetching is a bitch, so good luck.

Shanachie

On Sunday 09 July 2006 21:47, Stewart Allen wrote:
> Hi,
>
> I've got a very slow dial-up line at home and downloading podcasts takes
> a long time and very often the connection is broken.  This means that
> the download has to start from the begining.  So I'm busy changing the
> way podcasts are downloaded so that if a partial file exists the user
> has the option to resume the download.
>
> Hopefully you guys will find it useful.  I should be able to send a
> patch in the next few days.