Expand use of failOnFailedSeed option

[ldko]

I was looking at where --failOnFailedSeed was added in #300 and am wondering if it would generally make sense (it does in my mind for my use case) for that to also apply when a seed isn't parsed successfully. Right now if I crawl with a seed list (generated automatically from URLs extracted from a document) containing this particular URL https://doi.org10.1901/jaba.2012.45-85 that is missing a slash after the .org and is presumably read as having a TLD of .1901, the crawl aborts. Perhaps if failOnFailedSeed is false, such a problem could be logged lower than fatal, so the crawl doesn't abort?

[tw4l]

Perhaps we could have the following behavior for seeds that aren't parsed successfully:

• If it's the only seed or --failOnFailedSeed is set, fail the crawl
• If there are multiple seeds, not all of them are invalid, and --failOnFailedSeed is not set, log an error and continue the crawl

[tw4l]

In simpler terms, if --failOnFailedSeed is not set, we'll want to check that we have at least one valid seed and continue on if so or fail if not.

[ldko]

That logic makes sense to me. :)