This does not solve the captcha issue but allows the viewer to manually
solve the captcha by displaying a form and using the response from the
viewer. Maybe a first step to automated captcha solving?
This process relies on the use of a PHP session for storing captcha
details so that the user cannot submit anything else than the response
to the captcha. Response is filtered before being forwarded, also. Once
captcha is solved we get a page ready to be parsed, as usual.
Combined to some kind of OCR, this could automatically solve the
captcha, but currently if only automates the process of retrieving the
challenge and submitting the response.
Correction de la page d'accueuil pour qu'elle soit conforme aux standards du W3C.
Correction de la regex de listage des fichiers pour qu'elle ignore les sauvegardes.
Ajout d'un nettoyeur HTML, par défaut.
Currently emoticons are retrived in textual form eg <i><u>smile
emoticon</u></i> which is not really visual... so let's convert them back
as ASCII emoticons eg ':)'. This works using a hardcoded table mapping
emoticon names to their visual representation, and the regular expression
match the two words because eg in french facebook will display
<i><u>émoticône smile</u></i> so we need to test both. Unknown emoticon
descriptions will be left as is.
Retrieve author display name from page title
Build short readable title using name and content
Convert relative links into absolute links
Remove attributes from </a> tags (may cause issues)
Remove onclick and onmouseover (javascript code)
Retrieve url of first image of post as thumnail
Use author avatar as thumnail if no first image
Use display name in feed name: name - Fb bridge
Minor code indent fixes and use simple quotes
