clicking links in emails not working properly

[gucki]

Is there any special reason why not navigating to the "raw" url but wrapping it in request_uri instead? This is causing troubles when trying to follow absolte urls (with domain and/ or port).

    def click_email_link_matching(regex, email = current_email)
      url = links_in_email(email).detect { |link| link =~ regex }
      raise "No link found matching #{regex.inspect} in #{email.default_part_body}" unless url
      visit request_uri(url)
    end

    def click_first_link_in_email(email = current_email)
      link = links_in_email(email).first
      visit request_uri(link)
    end

[bmabey]

I think the original reason was due to some initial limitations of webrat. I have pull request #46 that allows for using the absolute URL but I haven't looked at it in a bit.. If you can verify that the patch in #46 fixes your problem then I'd feel better about merging it in.

[gucki]

I'd really prefer if we could check is this "hack" is still needed. As far as I can tell I never had any problems with any valid url I passed to visit. Can we give it a try?

[glebm]

email_spec at the moment breaks subdomain routing

My fix is this

# Working around issue here: https://github.com/bmabey/email-spec/issues/108
def request_uri(link)
  link && URI::parse(link)
end

def parse_email_for_anchor_text_link(email, link_text)
  if textify_images(email.default_part_body) =~ %r{<a[^>]*href=['"]?([^'"]*)['"]?[^>]*?>[^<]*?#{link_text}[^<]*?</a>}
    CGI::unescape_html $1
  end
end

[Papipo]

It's not just subdomains. I am using capybara and custom domains (full domains, not just subdomains), and email_spec uses relative paths, so I can't hit specific domains from emails.

[vidaica]

links_in_email seems not to function properly. In my case, the returned urls contain an extra single quote ( ' ) at the end if the urls in the links are wrapped with single quotes.

[janxious]

It also doesn't seemingly work properly if you anchor tag has anything in it other than an href attribute.

[StephenHogsten]

parse_email_for_anchor_text_link always joins with a ? - basically assume the path has only a query. If there is a fragment it treats it like a query