Even with the use of dedicated upscalers or tweaked settings with fancy
AI image processing algorithms doing a regular upscaling of a bitmap
image is never going to be as crisp as an actual higher resolution
bitmap image. This is why you use vector/SVG images or at least provide
bitmaps at various densities to prevent bad (up)scaling.
Yep, I’m aware of all that. Still, 3.3.0 should have crisper images on
HiDPI displays than 3.2.17 did if the images start out large enough to
support it.
I was just
wondering if you added support for providing higher resolution images,
since if you did I would’ve wanted to take advantage of it.
V4 will have that support for multiresolution images. I’m just not
interested in putting all that additional work in to V3 at this point.
I reread this just now and realized that I don’t understand what you’re seeing. The splitter problem I’m familiar with but don’t know how to reproduce reliably is that the splitter starts dragged all the way to the bottom of the Player window. What you’re saying is that there is no splitter at all displayed? I don’t understand how that’s even possible. What are you clicking on the toolbar to make the splitter show up?
I spent a few hours poking at the ComponentSplitter. Try 3.3.0-svn9354-splitter_test. Does that behave better?
I changed is the preferred size of the component in the top half of the Player window, and also reduced the initial number of rows in the chat pane. That gives the map significantly more vertical space when I start games.
I think this is the same issue that I have seen for a long time in Vassal. Generally, when I open a log file, I have to manipulate the window to get the slider and then use the vertical slider to see the map. As already said, it is common and confuses people new to Vassal.
I have noticed different behaviours with log file opening. Sometimes the map is displayed and sometimes I get a minimised window on the bottom left of my screen. When I maximise this, the map is not displayed but the slider is. However, no map and no slider is the common behaviour.
I just discovered that I can add saved game folders to modules in the main Vassal screen. When I open a log file in this way, it seems to behave better. In the 3 modules/log files I tested, 2 opened with the map displayed and one didn’t display the map but displayed the slider. As an aside, it would be useful it the contents of a saved game folder could be listed in descending date modified order.
3.3.0-svn9354 on MacBook Pro (with Retina) MacOS Catalina 10.15.4
Updated:-
deck getting deleted in set up on C&C Napoleonics 3.42 - as you know, still a problem
zoom bug when using the “saved map as PNG file” - confirming, fixed.
different screen resolutions - I think that I noticed this with Churchill module, in that I could not work with the map (scroll or zoom out fully) on a non-retina monitor whereas map was ok on the laptop screen. Anyway, on a recent Vassal version, the problem had gone away.
“uncaught exception” crash when re-playing some C&C Medieval logs - still occurring.
This build is svn9294 with svn9249 removed. That’s the only commit between 3.2.17 and svn9294 which isn’t a change for Java 9 compatibility or updating our dependencies.
This build is svn9294 with svn9249 removed. That’s the only commit
between 3.2.17 and svn9294 which isn’t a change for Java 9
compatibility or updating our dependencies.
C&C Nap - lost card deck :- I couldn’t reproduce this one
Ok, so it seems that this commit is the offending one for the lost
deck:
I don’t have a guess yet as to (a) why this commit would be the cause
(though it does affect Decks) and (b) why this commit would not affect
Linux, as there’s seemingly no platform-specific code in it.
Can anyone reprodce the C&C Napoleonics deck problem on anything other
than a Mac?
Huh, I got something weird to happen just now with the C&C Napoleonics Vimiero scenario which sounds like the reverse of what was reported: If I open the scenario, close it, reopen it, choose a command deck, and then undo that, the deck doesn’t disappear, but the button comes back! That’s goofy.
I suspect that this is the same problem you guys are seeing but manifesting itself a bit differently for me. I can reproduce this, at least, so I’m going to see about fixing it next.
One other thing though, I often don’t seem to be able to quit this module cleanly. I have to Force Quit. This applied to previous 3.3.0 builds, too. It’s something I have noticed in general (across other modules) under v3.2.17 but it seems to happen much more with 3.3.0 and the C&C Medieval module in particular.
I’m late to this thread but definitely interested in trying my modules out on 3.3.0 and reporting any errors.
One question – are there any special techniques for dealing with “multiple versions of Java at once” (I’m on Windows 10). Because I want to make sure don’t Scrootch Up my ability to run “regular Vassal” (3.2.17) and my Eclipse/java stuff associated with that.
I’m late to this thread but definitely interested in trying my modules
out on 3.3.0 and reporting any errors.
Please do. I’d like to kill as many bugs as we can before releasing a
beta.
One question – are there any special techniques for dealing with
“multiple versions of Java at once” (I’m on Windows 10). Because I want
to make sure don’t Scrootch Up my ability to run “regular Vassal”
(3.2.17) and my Eclipse/java stuff associated with that.
The Windows test builds have their own Java bundled with them. It won’t
interact at all with your existing setup.
(The MP3 that’s in the C&C Medieval module appears to be about 0.6 seconds of silence, which (a) I can’t the the purpose of, and (b) is maybe not even entirely well-formed.)
I am happy to report that both Paths of Glory 9.6 and For The People 3.0 both SEEM to be working fine, after some time putting each of them “through their paces” for a bit. Shockingly, even the “custom java classes” don’t seem to be throwing it (PoG9.6 has a couple tiny aesthetic things; but FTP3.0 has a significant one). That was a preliminary test and I will “stay at it” and see what I find.
I do see a bug reproducing itself, though I think it was already in 3.2.17. Basically a particular situation where moving a counter and then pressing “Undo” does not actually undo the move properly (to be clear, even with no custom classes anywhere in the module). I have a “minimum viable module” for that (UndoBug.vmod now in the upload moderation queue) if you’re interested, but completely understand if you’re only focusing on java-version-upgrade things. The situation the minimum-viable-module shows could of course be “worked around” by turning on “Snap To” for regions, but it’s merely a simple demonstration of an issue that creates other problems that can’t be worked around that way.
Meanwhile, will put some more time into pounding on my two modules with 3.3.0 and see if I can find anything else.