support for resuming downloads

[sascha_silbe]

Today Sugar crashed while downloading files. It would be great if downloading these could be resumed.

All metadata necessary to do this should be available. We might want to store the URL in a separate property so we don't need to parse the description, esp. since the user might modify the latter.

[RafaelOrtiz]

Chrome and Firefox provide ways to do this, we should be also having that on browse or surf when available.

[icarito]

Interesting challenge. Since the metadata is available in the journal, in theory Sugar could continue the download. Or we could probably wait until Browse is opened again. This is a design question.

[humitos]

Here are two things to keep in mind:

1. Do we want to continue the download after closing browse? If yes, we should use Sugar to get the file instead webkit. Get Book uses gobject.idle_add to handle the download. Is there a Sugar Feature that already handle this?
2. Do we want to stop the download after closing and resume it next time the browse starts? If yes, I think we can keep using webkit but we should research about how webkit handle by itself the resume option. I guess it could be "automagically" :)