[OSM-dev] [patch] JOSM 0.5 smart add node tool

Robert Hart bathterror at gmail.com
Thu Sep 27 23:06:55 BST 2007


On 27/09/2007, Frederik Ramm <frederik at remote.org> wrote:
> Hi Robert,
>
>    I haven't looked at your stuff yet. How does it relate to the
> improvements of the "add node and connect" mode I have made a few
> weeks ago (shift=re-use existing node, ctrl=insert node into segment)
> - does it make these obsolete?

Yes. I only discovered that those facilaties existed by reading the
source. Would have saved me a lot of frustration. My approach is the
other way around... reuse exisinting or insert if possible, unless
specifically told not to (by holding ctrl).

I think it would be worthwhile adding a tool status section to the
status bar at the bottom - with hints like "Reuse existing node [ctrl
to disable]" etc. I know there are tool tips, but as I always use the
keyboard short cuts I'd never see them...

> Have you read my longish http://josm.openstreetmap.de/ticket/11
> and what do you think of it? Are your changes a step in that general
> direction, or is it rather a small thing that makes life easier for
> the time until we drop all those modes?

I've not read that yet. I just scanned it, and it sounds like we are
both thinking the same sort of things though. I 100% agree that delete
is not a "mode" it's just a one off action, and then you go back to
whatever you are doing.

I'm really keen to get everything in place to make JOSM really
powerful and easy to use. I'll work on this some more. Maybe by the
time you've finished moving the API over to 0.5 I'll have it ready.

>
> I have fixed a strange bug in the "findNearestWaySegment" method
> yesterday that led to malfunctions in the "insert node into segment"
> method. If you're using the findNearest... as well then you should
> recompile!

Ah, I think I may have noticed something along those lines. Perhaps it
was the intersection snap, where you call findNearestWaySegment with a
collection...

Rob


-- 
-- 
Robert Hart




More information about the dev mailing list