aFeedSlot - URLS in Description become links automatically

[johnnyoffline]

Tom,

I am marking this critical only because it's related to a client project.

We need an anchor tag aware scrubber that looks through incoming RSS feed posts for URLS that are -not- links and transforms them INTO links.

Example below is a twitter feed. We get shortened URLS from a plethora of sites but they aren't clickable:

[boutell]

The weird part here is that RSS message bodies usually contain HTML, yet these are not already linkified in the HTML. Fixing that without breaking RSS message bodies that *do* contain HTML is the tricky thing, we can't just call aHtml::textToHtml().

[boutell]

This can be done at the app level and we can migrate the benefits into 1.5 later

[wjohnald]

This appears to be a non-issue as Symfony's TextHelper?'s auto_link_text already creates links from urls while ignoring existing links. After performing some unit tests, I rearranged the aHtml::textToHtml() function from:

return auto_link_text(simple_format_text(htmlentities($text, ENT_COMPAT, 'UTF-8')));

to

   	return simple_format_text(htmlentities(auto_link_text($text), ENT_COMPAT, 'UTF-8'));

I believe the problem was related to htmlentities replacing the characters needed by Symfony's regular expression before it had the chance to create the links. This seems to work fine, and I will commit my unit tests tomorrow.

[wjohnald]

Even better:

return htmlentities(simple_format_text(auto_link_text($text), ENT_COMPAT, 'UTF-8'));

[tboutell]

Looks like good detective work. I am a little surprised because I could've sworn the feed slot expected the post body to be HTML already and therefore probably wouldn't be calling textToHtml() on it; the issue as I understood it was that Twitter wasn't bothering to put HTML in their feed...? Is this solution safe for content that is already HTML, as feed items often are in other RSS feeds?

[wjohnald]

My unit tests are now in my trunk of the repository.

I have run a test that uses a mixed string (an HTML link "<a href="*">*</a>" and just a regular old text URL.

If passed a block of html code, it will linkify plain urls and add some extra formatting elements "<p>"s and "<br/>"s.

If auto_link_text is called alone, it will only linkify plain urls.

[wjohnald]

Resolved in commit [2107]