request: error handling without exceptions about publicsuffix-ruby

[weppos]

The purpose of the Public Suffix Library is to parse a valid domain into its parts. For this reason, an invalid input should be considered an exceptional event.

After all, this is the same interface exposed by the standard URI library.

If you simply want to validate the URL, you can use valid?

An other alternative is the one-line rescue.

tld = PublicSuffixService.parse(domain_name).tld rescue nil

Honestly, I would prefer the #parse method to raise an exception when it can't handle the input.

[gregwebs]

I guess I am wondering how you are using this library- I assume like me it is in a web applications where users enter domain names- receiving a domain name entered in an invalid way is a normal event. Certainly there could be another library to deal with this normal event. That library would have to parse the domain name also. Either that library would need a way to pass a parsed domain name to Public Suffix Service or it will have to be parsed twice and the domain parsing logic will be duplicated. Here is the code I am using now:

if (s = URI.parse(domain_name.to_s).scheme).present?
  errors.add :domain_name, "must not include the scheme: #{s}"
  return # make sure there is no scheme
end 
# URI needs the scheme to figure out what is the path!
if domain_name.present? and (p = URI.parse('http://' + domain_name).path).present?
  errors.add :domain_name, "must not include the path: #{p}"
  return # make sure there is no path
end 

# down to just a domain name now - which is what PublicSuffixService wants
unless PublicSuffixService.valid? domain_name 
  tld = begin PublicSuffixService.parse(domain_name.to_s).tld
  rescue PublicSuffixService::DomainInvalid
    nil 
  end 
  errors.add :domain_name,
    tld ? "has an invalid public suffix: #{tld}" : "invalid"
  return
end 

How are you validating domain names?