[Patch] Replace Geometry: Denies possible replacement with a wrong warning

[skyper]

What steps will reproduce the problem?

1. Have a closed way (id:0) with four nodes where three already have an id greater than zero and one is new (id:0).
2. Draw a new, closed way with four nodes.
3. Replace Geometry

What is the expected result?

The older way is replaced by the new one and all three node ids are reused.

What happens instead?

Warning: Please select one way that exists in the database and one new way with correct geometry"

Please provide any additional information below. Attach a screenshot if possible.

If I delete the new node from the first way the action performs fine.

Repository Root: http://josm.openstreetmap.de/svn
Build-Date: 2014-07-03 01:36:33
Last Changed Author: Don-vip
Revision: 7283
Repository UUID: 0c6e7542-c601-0410-84e7-c038aed88b3b
URL: http://josm.openstreetmap.de/svn/trunk
Last Changed Date: 2014-07-02 20:36:46 +0200 (Wed, 02 Jul 2014)
Last Changed Rev: 7283

Identification: JOSM/1.5 (7283 en) Linux Debian GNU/Linux 7.5 (wheezy)
Memory Usage: 547 MB / 882 MB (268 MB allocated, but free)
Java version: 1.7.0_55, Oracle Corporation, OpenJDK 64-Bit Server VM
Java package: openjdk-7-jre:amd64-7u55-2.4.7-1~deb7u1
Dataset consistency test: No problems found

Plugin: OpeningHoursEditor (30416)
Plugin: conflation (0.1.7)
Plugin: imagery-xml-bounds (30495)
Plugin: jts (30416)
Plugin: mirrored_download (30495)
Plugin: notes (v0.9.2)
Plugin: photoadjust (30428)
Plugin: reverter (30436)
Plugin: terracer (30416)
Plugin: todo (29154)
Plugin: undelete (30416)
Plugin: utilsplugin2 (30460)
Plugin: wikipedia (30449)

[simon04]

For unclear cases (also two new ways being selected) we could show a dialog similar to ticket:10730#comment:4 (split way action). Any thoughts?

[GerdP]

Possible solution

[GerdP]

Sorry, while writing this comment I found out that the attached patch will probably not always work.
Replying to simon04:

For unclear cases (also two new ways being selected) we could show a dialog similar to ticket:10730#comment:4 (split way action). Any thoughts?

Yes, such a dilaog would help, but we might display two ways with id 0 unless the user has configured JOSM to show negative ids.
I think in some cases it is possible to compute the right way if both are new, working on it again.

[GerdP]

sample data, search for "type:way new"

[GerdP]

The new patch fixes the problem. I've created a another example similar to the one described above where existing code fails. This one also adds a new (shared) node to both ways. The patched version also works with this case.
For now I've only changed the message to "Both ways are new and have new nodes, cannot decide which one has the correct geometry."
The patch also improves javadoc and simplifies a try/catch block.

Please suggest a better message, else I'll commit this tomorrow.

[GerdP]

fixed with [o34844:34845]