[Patch] JOSM doesn't work well when not maximized

[GerdP]

What steps will reproduce the problem?

1. start JOSM with clean home dir (JOSM opens in a non-maximized window)
2. Open BING background layer

What is the expected result?

Empty data layer and side windows for tags, relations etc are visible with reasonable geometry

What happens instead?

The side windows are far too small, barely visible and a new user probably has no idea what to do.

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

Maybe this doesn't happen always. I noticed it first opening a BING layer in latest r18934, tried again with older versions and then tried with CTRL+N for a new data layer. Always saw the same problem.
After creating this ticket I am no longer able to reproduce with CTRL+N, but BING still works, so I changed the description. Hope you can reproduce.

Build-Date:2024-01-05 09:09:15
Revision:18934
Is-Local-Build:true

Identification: JOSM/1.5 (18934 SVN de) Windows 10 64-Bit
OS Build number: Windows 10 Home 2009 (19045)
Memory Usage: 284 MB / 1753 MB (54 MB allocated, but free)
Java version: 1.8.0_272-b10, AdoptOpenJDK, OpenJDK 64-Bit Server VM
Look and Feel: com.sun.java.swing.plaf.windows.WindowsLookAndFeel
Screen: \Display0 1920×1080 (scaling 1.00×1.00)
Maximum Screen Size: 1920×1080
Best cursor sizes: 16×16→32×32, 32×32→32×32
System property file.encoding: Cp1252
System property sun.jnu.encoding: Cp1252
Locale info: de_DE
Numbers with default locale: 1234567890 -> 1234567890
VM arguments: [-Djosm.dir.name=23362]
Dataset consistency test: No problems found

[stoecker]

Probably we should restrict the minimum size of the sidebar. Something like 100px or so.

[GerdP]

That could be an easy work around but I think there must be a timing problem. My guess is that the code which calculates the geometry of the side bar runs parallel to the code which opens the background layer, and both change the layout of the windows. Unfortunately my Eclipse installation refuses to compile JOSM since some months, so I cannot debug anything right now.

Besides that: JOSM seems to open the background layer for lat=0,lon=0 at a low zoom, resulting in an a black background with an error message "Keine Kacheln auf dieser Zoomstufe". I think this is be very frustrating for newbies. Maybe JOSM could use a larger zoom or a "neutral" point on earth which shows some (interesting) data?

[anonymous]

a "neutral" point on earth which shows some (interesting) data?

?? doesn't that conflict?

whoever used lat=0,lon=0, intentionally or not, already picked a rather neutral spot..

[GerdP]

Yes, any other point on earth would probably cause trouble. My idea was to show an unpopulated area like desert that looks beautiful so that anybody understands that this is not his home area.
A black screen with an error messages doesn't say that and doesn't invite to continue.

[taylor.smock]

Replying to stoecker:

Probably we should restrict the minimum size of the sidebar. Something like 100px or so.

This would be "easy" to do. See source:trunk/src/org/openstreetmap/josm/gui/MapFrame.java@18934:239#L237​ .

Replying to GerdP

A black screen with an error messages doesn't say that and doesn't invite to continue.

Maybe geolocation? If the OS supports it anyway.

Replying to GerdP

Unfortunately my Eclipse installation refuses to compile JOSM since some months

I'm willing to help you troubleshoot, but I bet it will work after you run ant dist on the command line. It is probably the generated sources (I'm betting MapCSS) that are at issue, and the Eclipse project isn't "smart" enough to know that some packages are generated. I think this can be fixed via #8269.
EDIT: Poke me on #josm IRC so we don't spam this ticket.

[GerdP]

Reg. my Eclipse installation: I updated Eclipse and started to immport JOSM project from scratch and got it working again. I think I did that ten times before but probably didn't do ir right. I am now able to compile and debug (at least the java 8 version).
I'll try again to debug this ticket tomorrow.

[GerdP]

OK, reg. the too small side windows:
When I comment this line in AddImageryLayerAction

                AlignImageryPanel.addNagPanelIfNeeded(infoToAdd);

the layout is OK. Also, when I first add a new data layer and then a background image.
So, the additional panel at the top somehow corrupts the calculation of the layout. No idea where to fix that.

[GerdP]

possible fix for the ill-sized Nag panel

[taylor.smock]

Here is something else that works (with the default 1000w x 740h).

src/org/openstreetmap/josm/gui/layer/AlignImageryPanel.java

@@ -91 +91 @@
         MapFrame map = MainApplication.getMap();
         if (MainApplication.isDisplayingMapView() && showAgain.get() && !infoToAdd.isGeoreferenceValid()
                 && map.getTopPanel(AlignImageryPanel.class) == null) {
-            double w = GuiHelper.getScreenSize().getWidth();
+            double w = map.getWidth();
             map.addTopPanel(new AlignImageryPanel(w > 1300, showAgain, infoToAdd));
         }
     }

[GerdP]

I tried that also and it didn't help on my PC.

[taylor.smock]

That's funny.
I hate it when things work on one OS and not on another. What happens if you set w to 0.0?

[GerdP]

map.getwidth() returns 400 on my machine. That value is hardcoded, so I wonder how you can see a different result.

[taylor.smock]

That isn't hardcoded. I think we just set it to 400 during initialization, and it hasn't updated yet.

Either way, I saw different behavior when I ensured that the w variable was less than 1300.

[GerdP]

Different yes (two line vs. one line), but not good on Windows

[GerdP]

use popup width, not screen width

[gaben]

Ahh, I saw this behaviour many times but never was able to reproduce.

[taylor.smock]

@GerdP: Do you want to apply the patch?

[GerdP]

Will do it tomorrow.

[GerdP]

In 19147/josm:

fix #23392: JOSM doesn't work well when not maximized

• use SwingUtilities.invokeLater() for the AlignImageryPanel

[gaben]

I don't know if it's related but I recently noticed the tab pressing could not always hide the toolbar :/

[GerdP]

Any idea why the changes in r19147 could be related?