* Fixed cards being blurry. The problem isn't completely fixed, but the fix will suffice in most cases. Card Images are only loaded in once, and the cards were loaded in initially at a small size instead of their eventual size after layout. Really they should reload images when resized, but I will tackle that later, and it's only noticable when significantly resizing cards within one usage of the editor anyways.
* Removed some of the unneeded UI on the sideboard half of the DeckView so that componently are less likely to overlap at a small size. Not sure how to ideally handle this, as Swing doesn't really provide an elegant way to hide components at small sizes, and I'm sure the user would rather the components overlap than be unable to resize the window as small as they want due to a minimum size requirement.
* Shift-Click / Shift-Drag now work as expected as far as multi-selection
* Deck editor saves split pane split positions
* Card layout and sort settings are now saved along side the a deck when saving to the .dck format, so that you have back the exact same deck layout when you re-load the deck.
* Fixed the symbol image downloader to work around some of the large-size symbol images being missing on gatherer. Falls back to the medium sized images currently for those symbols.
* On specific platforms (it seems Windows 7-10, JRT 7+), under our specific configuration, in the case where the user selects the exact same file as they chose during a previous selection, JFileChooser reports that the user selected a file, but returns chosen file = null. This commit works around that by assuming they chose the same file as last time in that case.
* Working standalone proof-of-concept application in DragCardGrid.
* Still needs listener functionality in order to integrate it into the existing DeckEditorPanel class.
* Rather than respecting the spacing setting from Preferences, the spacing is now calculated from the card size if characteristic based rendering is enabled, since from the renderer we know exactly what spacing is actually needed.
* Simply invalidating the image render is not enough, a repaint call is also needed in a few cases where a repaint hadn't already been scheduled by the same change that trigged the selectable/chosen status change.
* Turns out that drawing black borders on non-rendered cards, while it fixes what was probably a bug, has unintentional and undesirable implications. Needs further work in the future, but revert it for now.
