someone should post it on mantis and we will try to include it in 3.9 (modulo that this is working).

Stef

On 21 juil. 05, at 9:58, Matej Kosik wrote:

cottonsqueak@earthlink.net wrote:

...

from preamble: "Change Set: ExplorerSpeedup-efc Date: 21 July 2005 Author: Eddie Cottongim Makes scrolling smooth in SimpleHierarchicalLists faster by using a binary search to only draw visible elements. TransformMorph has also been given this ability through an extension. IndentingListItemMorph has a minor change to prevent an unnecessary copy of its owners submorph collection. The object explorer remains slow to open initially. Try it out with (1 to: 10000) explore. "!

---

---

Hi,

I didn't read the source code, though after mere installation, scrolling indeed works better. Thank you. What will be the future of this changeset? Would it be incorporated to main stream? How? When? Who is in charge? -- mk

Matej Kosik wrote:

Suggestion:

Scrolling done directly by moving the scrollbar works (thanks to your changeset) faster. But when I try to scroll things up and down with mouse scroll wheel, things are as slow as before. I am not sure if this could be fixed with the similar trick. I guess that yes.

Thanks for the feedback. You are right; the mouse wheel is going down another path thats got a linear scan of submorphs. I'll look at it tonight.

Thanks again, Eddie

Matej Kosik wrote:

Suggestion:

If it had sense it would be advantageous to do the same enhancement also for the horizontal scrolling.

I think that the one set of changes should improve horizontal scrolling as well - the basic problem is more with drawing and event handling than the scrolling mechanism. Unless you mean scrolling a horizontally oriented list, which is something I have thought about but not tried. Does that sound right? If I've missed a case I'll be happy to look at it.

The changes to the event handling were more invasive than I first thought so I will be giving them a good test drive before I send a new version. In particular I've refactored MorphicEventHandler, which I think is probably a worthwhile change in its own right. There were three nearly identical submorph scans, which I've managed to consolidate to one method. Then its easy to override the submorph scan strategy if a morph has a better way to do it.

Eddie