Expose URL Parsing to Library Users about safebrowsing HOT 3 CLOSED

Exposing parseUrl (or a wrapper around it) sounds reasonable. It's a bit of duplicate work for you, parsing then doing a lookup. It's also heavier than the regex solution from the xurls package.

Another solution (which was talked about in another issue) was creating exposed errors/error types that clients could match on easier. In this case, maybe it would return a ParseError which you could check via a type assertion. The problem with that for LookupURLs is that it takes a slice of URLs, but can return a single error. So you'd have to un-batch to get these benefits.

What do you think?

from safebrowsing.

I'd prefer the first solution, because we're parsing the URLs anyway to extract various components, so the rework isn't too much of an issue, whereas the latter would be a bit messy codewise, and we'd have to some rather strange error cases to handle such as what to do if one link in a message containing 3 links is not properly formatted.

from safebrowsing.

Sorry for the delayed response! I've created a pull request that exposes a "IsURL" function. It doesn't necessarily eliminate the need to do some basic string splitting/regex work to find URLs (as explained in the commit), but would prevent getting a parse error when calling LookupURLs.

from safebrowsing.