On Sun, Aug 5, 2012 at 6:25 AM, C Anthony Risinger <[email protected]> wrote: > On Fri, Aug 3, 2012 at 4:38 PM, C Anthony Risinger <[email protected]> wrote: >> On Fri, Aug 3, 2012 at 5:44 AM, C Anthony Risinger <[email protected]> wrote: >>> this is brief because i need to go to bed, but i've got the gobject >>> backed running near flawlessly for simple examples, and only a couple >>> edge cases to get the heavier stuff running. >>> >>> still need to add important bits, like XMLHTTP, but i'll be working on >>> it more this weekend. the good news is DOM handling is near 100% >>> perfect, and stuff like alerts are working as well. > > DOM handling is 100% now, and save one or two very small quirks > (namely, window.location needs a small custom object) i am able to run > KitchenSink 100%! hooray! > > i'll be working on XMLHTTP today/tomorrow (whichever == Sunday in your > head/timezone) ... after that it should serve as a full replacement > for any other runtime. > > after XMLHTTP ill add Timer support -- SVG is not currently enabled in > GObject generated bindings, but by this point i'm pretty familiar with > the code so i may look into enabling it down the line. > > once we are relatively stable i plan to fix the upstream patch, > hopefully before next release, and also dig into Firebreath as the > End-All-Be-All-Solution.
well it was super-incredible-too-much-effort, but i've created a PPA for the Ubuntu crew: https://launchpad.net/~xtfxme/+archive/pyjs ... add the PPA and update, or if you don't have libwebkit3 already, install: libwebkitgtk-3.0-0 ... then checkout `feature/giwebkit`. i got slightly distracted fixing some hacky crap withe the way these loaders are initialized (will be pushing an update tomorrow that 1/2 eliminates the "need" for bootstrap.py), and i've unfortunately discovered that Luke sprinkled PyWebkitGTK hack all over the place instead of using the override system proper ... ugh .. will be remedying that ASAP. regardless, i expect to merge the `feature/giwebkit` sometime this weekend or early/mid next week at latest, once XMLHttpRequest is finished. there is currently a problem with history (triggered properly from app perspective but not actually doing the browser part), and no Timers, but both of those will be corrected soon after, and not significant enough to block what is already a remarkably stable backend. there is also once or two bugs in pyjs that the backend exposes but i've yet to fix. so test it out, and report back bugs, because this will be replacing both hulahop and pywebkitgtk ASAP. -- C Anthony --
