Python bindings for WebKit
 

I just got pywebkitgtk, the Python bindings for WebkitGTK, running on my tablet. As you can guess I built it against Bundyo's webkit package, so if you have Tear and Python installed you have everything you need. Binary and source packages over here.

Here's a screenshot of the demo browser, which is running pretty nicely (although largely unhildonized of course):

http://www.henso.com/files/pywebkitgtk/screenshot00.png

The original package is from the Gwibber team launchpad archive, and Gwibber is what I'll try to get running next. Should be fun :)

Re: Python bindings for WebKit
 

Yeah, I've been having a fun time hackin' on this stuff for the last few weeks. The individual progress bars come in very handy. And I really like the convenience of compiling pywebkitgtk right on the tablet. Still so much to do though; lots more fun to be had. Oh, and before today's kludges it ran unmodified on my Kubuntu machines.

Attachment 3257

Attachment 3258

Re: Python bindings for WebKit
 

im guess its the same old osk input issues (i think someone suspected that it was a deep webkit issue).

Re: Python bindings for WebKit
 

Mh, I wonder if i's possible to run this in a simple dialog/popup. It would be great for some kind of preview mode for the bookmark manager.

Re: Python bindings for WebKit
 

I'll claim ignorance 'cause I have zero osk issues, but there are some known GET and POST issues with gtk.Notebook. Bundyo has the webkit somehow compiled and running against Nokia's GTK so all hildonization works as expected. PyGTK and GTK aren't on exactly the same page, as an example I had to extend the ImageMenuItem to get a label in a popup menu. And there are a few other little things here and there, but this is the only browser I use for browsing (I haven't implemented downloads yet). It's really nice opening a new tab/link without leaving the current window; I've wanted that ability for over a year.

But to be fair and balanced, webkit GTK is in a bleeding-edge state at the moment; the API changes with the weather. A moving target like this is for hackers only IMO and access to the DOM is in major flux. The pywebkitgtk guy yanked the gjs stuff recently, I'm gonna try and put it back in over the next few weeks. I'm not sure without it, unless there's a way through ATK objects, that webkit can know export it's exact context. I think that's part of the reason why you can scroll while in a text box.

Re: Python bindings for WebKit
 

try opening the full screen osk and hit enter.

i have also seen space become deactivated in the stylus osk...

Re: Python bindings for WebKit
 

Problem #1: My ignorance is getting worse by the minute; I can't for the life of me remember how to bring up the fullscreen keyboard.

Yes, there are many, many bugs, holes and other assorted issues. But like qgil (I think it was qgil) said in another thread, you have to pick a few problems and go after them. And think about how sketchy my situation is: I've got 4 libraries that are all slightly out of sink. Not just 2. Regardless, it's my primary browser and it should only get better. A big usability feature I implemented early on was domain-based user content css files. I wouldn't be using it otherwise. A big thanks to Benson and Brontide for that one.

Re: Python bindings for WebKit
 

either center dpad or using a finger on the input area...

tho im not sure about dpad if one is using a N810...

Re: Python bindings for WebKit
 

I'll plug my n800 in later and check it out (currently waiting on a new battery). But I couldn't get it to work on my n810.

And, of course, endless thanks to Bundyo.

Re: Python bindings for WebKit
 

I verified that this is not working.