Thanks, Ken. It sure does appear to act as you suggested. Here's a potentially interesting data point. Since I'm lucky enough to have two Macs both running 5.10 but on different platforms, I tried a simple A/B test. On my Studio (current, new machine - acting weirdly), I started up JMRI and opted to create a new profile called "foo". After clicking OK, it opens up the preferences window as well as the startup PanelPro dialog. Both are "skinny" (see the original post below). Without doing anything else, I went into the profile folder and "vi"ed the user-interface.xml file. Then I went to the iMac (older but working system) and did the same thing. Here are the two xml files:
<?xml version="1.0" encoding="UTF-8"?>
<auxiliary-configuration xmlns="http://www.netbeans.org/ns/auxiliary-configuration/1">
? ? <classPreferences xmlns="http://jmri.org/xml/schema/auxiliary-configuration/class-preferences-4-3-5.xsd">
? ? ? ? ? ?(deleted - they're identical between the two files
? ? </classPreferences>
? ? <windowDetails xmlns="http://jmri.org/xml/schema/auxiliary-configuration/window-details-4-3-5.xsd">
? ? ? ? <window xmlns="" class="apps.gui3.tabbedpreferences.TabbedPreferencesFrame:Preferences" height="1440.0" locX="-654.0" locY="0.0" width="80.0"/>
? ? ? ? <window xmlns="" class="PanelPro" height="237.0" locX="48.0" locY="583.0" width="80.0"/>
? ? ? ? <window xmlns="" class="jmri.jmrit.roster.Roster"/>
? ? </windowDetails>
</auxiliary-configuration>
?
Now, the .xml file from the old, working system:
<?xml version="1.0" encoding="UTF-8"?>
<auxiliary-configuration xmlns="http://www.netbeans.org/ns/auxiliary-configuration/1">
? ? <classPreferences xmlns="http://jmri.org/xml/schema/auxiliary-configuration/class-preferences-4-3-5.xsd">
? ? ? ? ? ? ? (again deleted - same as new file)
? ? </classPreferences>
? ? <windowDetails xmlns="http://jmri.org/xml/schema/auxiliary-configuration/window-details-4-3-5.xsd">
? ? ? ? <window xmlns="" class="jmri.jmrit.roster.Roster"/>
? ? ? ? <window xmlns="" class="apps.gui3.tabbedpreferences.TabbedPreferencesFrame:Preferences" height="1440.0" locX="147.0" locY="0.0" width="1018.0"/>
? ? ? ? <window xmlns="" class="PanelPro" height="237.0" locX="1025.0" locY="601.0" width="510.0"/>
? ? </windowDetails>
</auxiliary-configuration>
?
Notice that in the first, incorrect file, the width for both the Preferences window and the opening PanelPro window tags is "80.0". For the working system, the Preferences window width is set to "1018.0" and the width for the PanelPro window is "510.0". On both systems, the width in the .xml file appear to correctly match the windows that appear on the screen. In fact, it looks like all "skinny" windows are 80 [pixels] wide. (When they open, I line them up with each other to check it out.) Not sure where that 80 comes from. One random thought I had was that 80 is frequently the default width (in characters, not pixels) of a terminal window (24x80). Maybe just a meaningless coincidence, but it did come to mind.
Maybe another thing worth noting is the locX property in the "bad" file. The Preferences? window tag has a locX property of -654.0 and the PanelPro has a locX of 48.0. On the screen, the PanelPro pane is on my main monitor. The Preference pane is painted on the secondary monitor. I have the Mac configured so that the main monitor is on the right side of my desk and the secondary is on the left. This then looks as though the coordinate system is based on the main monitor and the negative locX for the Preferences panel reflects the fact that it is to the left of the main monitor (whose left screen edge would represent x = 0).
A couple of other things that come to mind. The new system (with the problem) is a Mac Studio with two monitors. The old, working system is an iMac with a 5K display. Prior to getting the Studio, I used this same Thunderbolt monitor (located on the left as it is now with the new system). I don't remember ever having "skinny window" problems with that system. Perhaps there is a difference between how Apple reports screen geometries between a headless system. with two monitors and an all-in-one computer with one monitor.
The iMac now has no second monitor. So, any tests I run at this point will not be an exact apples-to-apples comparison (no pun intended...although I like it? :-0 ) And, as we appear to be seeing now, even the new system (Studio) works with only one monitor attached.
Please feel free if there are some tests you would like to see run. I am eager to help if I can.
Vince
