Notices


Reply
Thread Tools
Posts: 87 | Thanked: 98 times | Joined on Oct 2007 @ Austria
#361
I uploaded a new interim package to my site, source code is in the garage repository. The new features in 0.0.9.5 are:
  • Implement tap-and-hold context menu for open-link-in-new-window functionality
  • Check for connection status on page load and request a connection if necessary
  • Remember zoom level between sessions

And the accumulated list of changes since pronvit's 0.0.9 release:
  • Switch to bundyo's libwebkit package
  • Re-implement kinetic scrolling
  • Make Plugins and Javascript switchable via browser UI
  • Implement simple save-as file download functionality
  • Make install script less intrusive, but require browser-daemon restart or reboot for first install
  • Improve browser engine chooser applet
  • Implement back/forward button popup menu

Unfortunately, I couldn't figure out the stop button crashes so far. If you press the stop button after the first bytes of a page have been read but before the main resource is fully loaded, the engine is pretty likely to crash. I can't reproduce this with Tear, so it must be something with the webkit-eal code. I'm having a hard time debugging browserd in a meaningful way, any help is greatly appreciated.

Apart from the stop button, the code is pretty stable for me now, so I guess the next release will be a more official 0.0.11. Just remember to restart the browser daemon or reboot your tablet if you upgrade from pronvit's package or get any strange behaviour, and let me know how this is working for you.

Last edited by hns; 2009-03-17 at 19:30.
 

The Following 7 Users Say Thank You to hns For This Useful Post:
tso's Avatar
Posts: 4,783 | Thanked: 1,253 times | Joined on Aug 2007 @ norway
#362
the crashes would not be so bad, if it only loaded webkit rather then a partially broken microb (iirc).
 
Posts: 87 | Thanked: 98 times | Joined on Oct 2007 @ Austria
#363
Originally Posted by tso View Post
the crashes would not be so bad, if it only loaded webkit rather then a partially broken microb (iirc).
You mean it uses the microb engine, ignoring the setting in the browser engine chooser? This happens sometimes, restarting the browser daemon (as root) helps:

/etc/init.d/tablet-browser-daemon restart

Another interesting thing I noticed is that webkit is much slower when booting from SD-card on my N800, even a fast Sandisk Extreme III, than when booting from internal flash. No idea why this is so, but I found it a bit surprising.
 
qole's Avatar
Moderator | Posts: 7,109 | Thanked: 8,820 times | Joined on Oct 2007 @ Vancouver, BC, Canada
#364
hns: Thanks for picking up the torch here.
__________________
qole.org --- twitter --- Easy Debian wiki page
Please don't send me a private message, post to the appropriate thread.
Thank you all for your donations!
 
tso's Avatar
Posts: 4,783 | Thanked: 1,253 times | Joined on Aug 2007 @ norway
#365
Originally Posted by hns View Post
You mean it uses the microb engine, ignoring the setting in the browser engine chooser? This happens sometimes, restarting the browser daemon (as root) helps:

/etc/init.d/tablet-browser-daemon restart
no, iirc, sometimes the the webkit engine would crash (say thanks to the stop button issue) but then browserd would recover without the browser window closing.

but then it would use microb rather then webkit, and would have some odd input issues that i do not recall the specifics off, sadly...

interesting, i just tried to follow a link, then stop the browser before a connection was made (the tablet was disconnected from wifi), end result, it stopped fine, the connection came up then it tried to download the page, only to crash the whole browser window...
 
Posts: 542 | Thanked: 117 times | Joined on Sep 2008 @ 52 N, 6 E
#366
I tried the Webkit engine for Microb.

It is faster than microb.
Maybe stupid questiion, does it support cookies ? (Local) sites using cookies do not recognize any cookie.
And opening wikimapia opens a lot faster but I cannot drag on the map. Is this normal ?
 
Posts: 87 | Thanked: 98 times | Joined on Oct 2007 @ Austria
#367
Originally Posted by tso View Post
no, iirc, sometimes the the webkit engine would crash (say thanks to the stop button issue) but then browserd would recover without the browser window closing.

but then it would use microb rather then webkit, and would have some odd input issues that i do not recall the specifics off, sadly...
I was refering to that behaviour, except I didn't notice the input issues.

interesting, i just tried to follow a link, then stop the browser before a connection was made (the tablet was disconnected from wifi), end result, it stopped fine, the connection came up then it tried to download the page, only to crash the whole browser window...
It's amazing what one little character can do. I just built and uploaded a fixed package. Thanks for the find!
 

The Following 2 Users Say Thank You to hns For This Useful Post:
Posts: 542 | Thanked: 117 times | Joined on Sep 2008 @ 52 N, 6 E
#368
Can the developer of this webkit check whether cookies do not work ?
Here an example URL to reproduce, it is an RPN and expression javascript scientific calculator which saves its data (RPN stack and numbers) as a cookie, so a reload of the page retains the numbers on the stack which is similar to the good old 'permanent memory' calculators.
http://tools.velp.info/scicalc.html
This works in ANY browser, even on Nokia phone mobile browsers, except this webkit engine where it always starts with 0, so it is easy to reproduce.
 
Posts: 87 | Thanked: 98 times | Joined on Oct 2007 @ Austria
#369
Originally Posted by skatebiker View Post
Can the developer of this webkit check whether cookies do not work ?
Cookies are working for me, but you are right, they don't seem to work on that calculator page. It may be an issue with script-generated cookies (vs. server generated ones). I can't find a setting for this in WebKit, but I'll look into it.
 
Posts: 542 | Thanked: 117 times | Joined on Sep 2008 @ 52 N, 6 E
#370
Originally Posted by hns View Post
Cookies are working for me, but you are right, they don't seem to work on that calculator page. It may be an issue with script-generated cookies (vs. server generated ones). I can't find a setting for this in WebKit, but I'll look into it.
Here I'll show you the javascript which generates the cookies:

I hope this will help you, e.g. by creating a simple page which just does as listed below.

BTW I use these files also locally (i.e. with file://). I have a scrabble game which uses the same functions for saving the game which works excellent on all browsers except the webkit engine on microb.

Hope this helps you to looking at this problem.

Code:
<!-- saved from url=(0014)about:internet -->
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">

<html>
  <head>
    <title>cookie test</title>
    </head>
<body>
<script type="text/javascript">

function getcookievalue(name)
{
    var ind;
    name += '=';
    str = document.cookie;
    if ((ind = str.indexOf(name)) != -1)
    {
        var end = str.indexOf(';', ind);
        if (end == -1) end = str.length;
        str = str.substr(ind+name.length, end-ind-name.length);
        str = unescape(str);
        return str;
    }
    return '';
}

function setcookievalue(name, value)
{
    name += '=';
    var expDays = 100;
    var exp = new Date();
    exp.setTime(exp.getTime() + (expDays*24*60*60*1000));

    var expire = '; expires=' + exp.toGMTString();
    var str = name + escape(value) + expire;
    document.cookie = str;
}

alert(getcookievalue('test'));
setcookievalue('test', new Date());


</script>
</body>
</html>

Last edited by skatebiker; 2009-03-17 at 11:15.
 
Reply

Tags
browser, webkit


 
Forum Jump


All times are GMT. The time now is 05:17.