Fix (at least) windows problem.

[lookfwd]

When you run scrapyd on Windows, paths like C:\blah are generated and then the "C" is assumed to be the storage type resulting in errors ("Unknown feed storage scheme: c" - from feedexport.py). Adding "file://" should be enough to fix this problem.

[pablohoffman]

@lookfwd thanks for helping to improve Scrapyd on windows, it's a platform traditionally not well supported on Scrapyd (unlike Scrapy), so expect more simple fixes like this.

I would like to merge this change but at the same time have it tested to make sure it won't break again, any ides on how to do that?. Maybe we could port the extras/text-scrapyd.sh to windows.

[lookfwd]

Thanks, actually I do some cross-testing between Windows and Linux etc for the book. The only other limitation I found is the length of Arguments on the exec of the spider which is limited in comparison to linux. It's probably just a documentation issue. I will have a look on extras/text-scrapyd.sh

[pablohoffman]

@lookfwd I decided to merged this one, but I look forward to improving the test suite (esp. on windows).