[Ext] SessionSaver .2d: new (rewrite)

[rue]

However, when I shut down, only the top three lines remain in the prefs file.
The critical question: and the "shutdown" pref is still false?
.
(ps: not a hassle. it's kinda fun to track stuff down.)

[Xanager]

Yeah, the shutdown pref is still false. I have two profiles, in the old one I just ran the installer to update the SS. This one now doesn't have the first two lines of autosave and autoload. It only has the shutdown pref which is false.

On the new profile, it has all three prefs, but the shutdown is still false.

[rue]

nightly 3 pass2 is posted.

same as pass1, but using setTimeout instead of eval

[Xanager]

Hate to be the bearer of bad news, but alas. When I shutdown via the file menu, the shutdown setting always goes true and stores the last page. But any other method of shutting down renders the shutdown setting false, and no session storage.

I started a new profile, and also tried updating my old one. All to no avail. Would it make any difference loading it in the browser directory? I wouldn't think so, but just asking.

[johnleemk]

Whoops. Ignore this.

[SteveRzx]

I just started from a clean slate. I installed firebird then installed SS for browser instead of profile. Nothing else changed, haven't even set my homepage yet. And the same exact thing still happens.

I'm confused. But I don't do this kind of programming either

Good luck. Any other tests you need me to try for ya, I'll be happy to.

[rue]

d1 nightly 4 is posted: <a href="http://adblock.mozdev.org/sessionsaver/sessionsaver-02-dev.xpi">Installer</a>, and <a href="http://adblock.mozdev.org/sessionsaver/sessionsaver.jar">jar-file</a>.

observer-object is really copied local to the hidden-window

--
Steve:
You posted one minute before I uploaded :P
.
The problem looked simple, but presented no easy solution: copy an object from one scope to another. I broke the onload function in two, intending the observer bridge them when loaded. But the observer refused to really copy to the Hidden Window -- no matter what I tried. After.. uh, a long time, I hit on defining it as a function and exporting to string. Passing the string through a setTimeout and calling new on it constructed the copy in the second scope.

[SteveRzx]

lol, you aren't going to believe this. I just tried your new one.

Doesn't seem to work for me at all now.

The only prefs that are getting saved are the 2 checkbox options.

And when I start firebird I get this msg in the console:

Error: uncaught exception: Permission denied to get property UnnamedClass.classes

This was on a clean install btw. Installed for user, not browser.

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031007 Firebird/0.7

[dorando]

Would you change:

Code: Select all

         tabHistoryAll.push(savedHistory); // ..and add it to master-array
         
         // catch the current-tab
         if (Tabbrowser.browsers[i] == Tabbrowser.mCurrentBrowser)
            selectedtab = i;

to

Code: Select all

         tempI = Tabbrowser.mTabContainer.childNodes[i].ordinal || i;
         tabHistoryAll[tempI] = savedHistory; // ..and add it to master-array

         // catch the current-tab
         if (Tabbrowser.browsers[i] == Tabbrowser.mCurrentBrowser)
            selectedtab = tempI;

please?
This should make it work with tabs reordered by miniT (perhaps also with TBE?) without breaking anything (at least I hope so).

[rue]

Steve:
Was that the whole error? No line-number?
.
It's rather odd. I'm having no trouble at all under moz1.3.1. I'll continue poking at things, but I honestly don't see anything wrong; yet...

--
dorando:
The rest is fine, but are you sure you want this: tabHistoryAll[tempI] = savedHistory; ? If there's any gaps in the tab-numbering, it will insert empty cells in the array. It would be better, I think, to just add the ordinals straight into the array, and sort by them later.
.
Btw, is the latest build working for you?

[SteveRzx]

I copied the error exactly. No line #s.

[dorando]

If there's any gaps in the tab-numbering, it will insert empty cells in the array.

There shouldn't be any gaps with miniT (if it happens something got wrong) and if I read the source correct also none without it... Reordering later might be more failsafe tough.

Btw, is the latest build working for you?

Nope, same error as SteveRzx.

[rue]

Could someone with the latest build installed, run <a href="http://forums.mozillazine.org/viewtopic.php?t=47184&start=15#342955">this test</a>, followed by <a href="http://forums.mozillazine.org/viewtopic.php?t=47184&postdays=0&postorder=asc&start=30#345163">this one</a>?

[Xanager]

Okay, I'm getting the same "Error: uncaught exeption" as steve.

On the first test, I'm getting a window full of code.
On the second test, I'm now getting an "undefined" instead of a "1" as I did on previous builds.

Hope this helps.

[SteveRzx]

Could someone with the latest build installed, run <a href="http://forums.mozillazine.org/viewtopic.php?t=47184&start=15#342955">this test</a>, followed by <a href="http://forums.mozillazine.org/viewtopic.php?t=47184&postdays=0&postorder=asc&start=30#345163">this one</a>?

First test gave me <a href="foo">this</a>.

The second test resulted in 'undefined'