FireSoup is a Firefox extension that allows posting to the soup.io tumblelogging platform. It behaves similar to the official soup.io bookmarklet and even opens up the same service page, the actual opening of the post window (including the communication of necessary page information) happens by means of chrome instead of Javascript injected into the currently open page, the latter which leads to major issues if you happen to use NoScript or probably similar JavaScript blockers.

In case you come across any bugs in this extension, don't hesitate to report them.

The Git repository of the extension can be found at Github.

FireSoup is hosted at addons.mozilla.org. addons.mozilla.org is currently experiencing some issues that lead to several addons (this one among them) not being installable via the addon page (Error -261, invalid file hash). However, installation via my user profile there seems to still work, so please use this until the issues are fixed.

Alternatively, you can install the latest version (0.3 as of 2010/05/01) from here.

FireSoup provides a new toolbar button you have to first add into Firefox. Simply right click on a toolbar, select “Customize…” and in the opening toolbar palette window scroll down until you find the FireSoup button and drag it into the toolbar of your choice.

After you added the button, you can use it to open the soup posting dialog you might already know from the soup.io bookmarklet. It's used identically, e.g. text on the page selected prior to opening the posting dialog will be prefilled as a quote etc.

FireSoup was created after I got fed up with the issues of the soup.io bookmarklet with pages on which JavaScript is disabled (e.g. by NoScript). The bookmarklet does some script injection into the current page to retrieve the selected text, site title, images and so on, which then gets posted (as in by HTTP Post) into the opened bookmarklet frame. This of course fails if the current page does not allow the execution of JavaScript. The result is a bookmarklet frame hanging on the “The bookmarklet is loading” screen with not a chance to ever work properly.

I circumvented this issue by doing the opening of the bookmarklet frame and the feeding of the necessary information via a HTTP Post from within the extension instead of by JS injection. Thus, all functionality from the bookmarklet is preserved, only the means of calling it is changed.

My initial goal was to write a real soup.io extension that does not use the bookmarklet service pages. I sadly had to give up on that plan though thanks to the non-existing soup.io API. While I came quite far by reverse engineering the HTTP communication between bookmarklet page and the soup.io servers, there were simply some pieces of information¹⁾ not available without a direct way of communication with the server or page scraping, which in my book comes way too close to a hack and is too easily broken.