Some people think that the autoscroll version in AiO is "better" and should be the basis for the "standard" one in FireBird. See http://forums.mozillazine.org/viewtopic.php?t=14971&postdays=0&postorder=asc&start=10 As the author if AiO Autoscroll enhancements, I ported the AiO autoscroll in browser.xml. See attachment in next post. Enhancements: - smooth scrolling speed changes (so called pseudo-logarithmic). - better "something to scroll" detection: ability to scroll in divs with overflow: auto, textareas, etc.

AiO = All-in-One Extension: http://perso.wanadoo.fr/marc.boullet/ext/extensions-en.html

Comment on attachment 127429 [details] [diff] [review] Patch against 07-06 browser.xml Blake, can you take a look at this?

--> enhancement

This also fixes bug 212072. I think this bug is more than just an enhancement because it fixes two other bugs.

Steffen: That's right. -> OS: Windows Tweaking summary to be a meta bug.

The patch is bit-rotted because Blake modified browser.xml twice since then. I wonder why he's fixing the bugs of the native autoscroll instead of just switching to the AiO one.

We are still investigating to switch AiO. Please wait for a while. Thanks.

checked in by noririty between 03:34 and 04:21 today.

but why isn't it marked FIXED then?

Because it's a meta bug. Fixing this depends on all the dependant bugs being fixed.

backing out AiO because of performance and some event handling issues.

When/if AiO's implementation gets ported, we should make sure that bug 213250 is fixed in that extension's code, too. I first noticed it in the extension, and found it again in a recent nightly. See my comment at http://bugzilla.mozilla.org/show_bug.cgi?id=213250#c7 on how to fix it (for the current non-AiO impl).

We also have to do something about the conflicting nature of autoscroll and contentLoadURL in Linux, particularly since both are enabled by default. Right now it's simply a usability nightmare as they fight each other. In bug 216899 I suggest a solution that will allow both features to co-exist and not conflict. How difficult this will be to code I'm not sure, but it does involve giving precedence to contentLoadURL rather than autoscroll (autoscroll currently seems to have precedence in most cases).

Bug 242621 has been filed for moving the autoscroll icon out of the DOM, which would solve at least five of the above bugs.

Robert, would this include not being able to scroll with wheel or autoscroll AT ALL in e.g. DIVs with overflow=scroll? Or is this a separate bug?

*** Bug 291134 has been marked as a duplicate of this bug. ***

Test it

With bug 242621 fixed, we have a "better" autoscroll. A couple deps for this bug are still open, but it might be ok to resolve this metabug.