Detect only well formed URLs by running them against a valid URL regex

[shubhamgoyal]

To reproduce the issue that prompted this pull request, try including a URL preceded immediately by a character. For example,

ahttps://blog.glycoleap.com/herbal-chicken-brown-rice-congee-ultimate-asian-healing-food/

Notice the a before the URL. But Autolinker's URL parsing doesn't seem to be perfect and so, it treats the whole string as a URL and if you call match.getUrl(), it also adds the protocol before it. This leads to http://stackoverflow.com/questions/38099684/no-suitable-url-request-handler-found rather than the promise being gracefully rejected.

To solve this, I add a well recognised regex https://gist.github.com/dperini/729294 and use it to invalidate malformed URLs.

[ospfranco]

Well, now I'm wondering, if it were to re-validate URLs why use autolinker, might as well do something like string splitting and validate each piece against the regex.

[shubhamgoyal]

@ospfranco I agree with you there. It seems to be a waste to use autolinker now, would you like me to push a version without autolinker?

[ospfranco]

Already on it, will publish to npm in a few minutes.

[ospfranco]

Published under v1.1.0 release.