Review Request: Rewrite Google's tracking URLs in search results

[Thomas Fischer]

> On Dec. 23, 2012, 12:57 p.m., Anders Lund wrote:
> > Wouldn't it be better to improve the userscripts plugin for KHTML? I have  auserscript that removes the google tracking URLS in khtml, and there are probably similar scripts eg for facebook and apart from that a lot of other usefull scripts in userscripts.org.
> > 
> > I do not understand the rationale behind targeting one specific website this way! Just my 2c :)
> 
> Thomas Fischer wrote:
>     > userscripts plugin for KHTML
>     Do you mean this one here?
>     http://kde-apps.org/content/show.php?content=140676
>     It says it is no longer maintained. I will have a look ...
>     My code is fairly simple and more likely (I assume) to get accepted than a "large" solution like userscript.
>     
>     > rationale behind targeting one specific website this way
>     It was my itch to scratch. Google is just the start.
>     As I stated in the code as a TODO comment: more cases to add!
>

Hello Anders, your comment on my posting two days ago isn't here, but I'll answer it here. I agree that hardcoding replacement patterns is inflexible regarding future changes on Google's (or anyone else's) homepage.
I already fetched the latest sources from KHTML-userscript and imported them into my git scratch. I'll have a closer look during the next few days.


- Thomas


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://git.reviewboard.kde.org/r/107867/#review23899
-----------------------------------------------------------


On Dec. 23, 2012, 11:09 a.m., Thomas Fischer wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://git.reviewboard.kde.org/r/107867/
> -----------------------------------------------------------
> 
> (Updated Dec. 23, 2012, 11:09 a.m.)
> 
> 
> Review request for kdelibs.
> 
> 
> Description
> -------
> 
> This patch adds the feature to KHTML to rewrite URLs that are used to track users. Right now, only tracking URLs from Google's search result are supported, but the list can be expanded (hard-coded right now).
> Example: A search for "KDE" may result in a list of links, including a link like
> http://www.google.com/url?q=http://www.kde.org/&sa=U&ei=YsYFfgOqAZzBQBC&ved=GEFANYNoNG&usg=Y8BfN6qj0QYNHYJQQBEB
> When you follow this link, Google will transparently redirect you to http://www.kde.org, but still record your behaviour.
> The patch rewrites such links already in the HTML parsing phase, i.e. you never see the tracking URL, but instead the final URL only.
> 
> The rewrite feature can be disabled through a setting, but there is no GUI for that yet.
> 
> I was thinking about automatically detecting tracking URLs through a regular expression, but I guess running a regular expression check for every URL would be too time-consuming.
> 
> I wrote the patch for 4.9.3 as this is the version I am using on the testing machine. I assume the affected classes haven't changed much in recent months, so it should be fairly simple to port to HEAD or future 4.11.
> 
> 
> Diffs
> -----
> 
>   khtml/khtml_settings.h 0faec6d 
>   khtml/khtml_settings.cpp b5693b4 
>   khtml/xml/dom_docimpl.cpp bb65a89 
> 
> Diff: http://git.reviewboard.kde.org/r/107867/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Thomas Fischer
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-core-devel/attachments/20121225/a361ef8f/attachment.htm>