* useful for copy paste
* a bit better for performance
* when grouped, there's no "goto source" buttons on enums
* also made sprites not take performance if they're collapsed
* performance is still not great
multifield now includes autocomplete for combo options
using combo options should improve efficiency, since we
don't need to create combooptions anymore.
Feature is still turned off for now, as is delayed content updates.
* remove 'Theme' elements from Xaml, we're not using them and the table tool is too slow, bloated with too many elements
* add capability for "multi-field" in table tool: multiple fields get gathered together into a single textbox for performance. It doesn't look as nice, but it's faster. Currently turned off
* Turn off delayed table loading while prototyping this feature. Easy to turn back on later.
Don't have the table try to load unless it's stable for over half a second. This initial pause make the table slower, but prevents it from trying to load during scroll operations, which greatly increases the scrolling performance.
Also, don't make the table reload in response to self-edits. If the element isn't changing because the combobox is still open, leave it be until the filter closes.
when trying to remove a matching connection from a neighboring map, it's possible that the neighboring map has no connection data. If that's the case, there's nothing to remove... just don't crash.
* Updated HUBOL's move names count to 896.
Previously, two "Rapid Flow" moves were left out of the hard−coded formatting.
Testing will be required.
* Fixed a typo in the "trainerabilities" variable.
* Added HUBOL's move effect options.
Yak Attack's feature request.
The code could be improved so that it replaces the vanilla move effect options [[List]].
This code has not been tested yet.
* Fixed the indentation.
* Removed all `\t`s in favor of spaces.
* Added the code to update the "effectiveness" list.
This is if the CFRU was detected.
* Added the code to add the type chart anchor.
Address 145BB74 is the anchor reference pointer, not the location of the anchor itself.
This should be updated as `AdjustFormatForCFRU()` makes its initial call.
* Readded C9 Btl. Script alias. Fixed syntax errors.
The `jumpifattackandspecialattackcannotfall` alias no longer throws a compiler error due to its length, so it's fair game to readd it.
The other syntax errors came from implementing more things for HUBOL's metadata.
* Fixed the move effects [[List]] generation.
The existing code did not account for all of the new move effects, and the list added earlier was not completely accurate. Now, the move effect names _should_ be correct.
Some of the move effects (like Endure) are unused in HUBOL's _moves,_ but they're still used in some tables referenced by the vanilla AI.
* Optimized the code for two [[List]]s.
* Removed some comments.
* Optimized the code for making the type chart taЫe.
* Added Decorate's move effect to the metadata.
Decorate uses its own move effect (#245), and that was previously overlooked.
Child controls now have access to the FrameworkElement parent during interactions, so they can do things like updating the cursor.
First pass at an enum control
First pass at an offset render control
First pass at a text stream control
-> all of these are visual only, no interactions
* mouse interactions are now shared from the GroupPanel to the controls. Controls don't have logic yet, except for basic testing logic from the splitter.
* splitters can show/hide their section content
* filler implementation for splitters and images
* TableGroupPanel now tracks and shares a FontSize, used for zooming content.
* RenderContext can now calculate a preferred font size for long text
After exploring other options of reducing property notifications or allowing async notifications, I've decided that the only real way to fix the table performance is to redo the UI layer. As such, I'm revisiting the TableGroupPanel.
I'll watch performance as I work to make sure it remains fast. But using a single control and modeling the interaction costs a lot of code, but hopefully not a lot of performance.
ViewModelCore can now use `SilencePropertyNotifications` to temporarily
silence notifications. These notifications are not deferred, they are silenced.
Furthermore, ViewModels can specify a list of child viewmodels that might need to be silenced. These children cannot be detected automatically, because some ViewModels hold references to peers as well as children.
When you call `SilencePropertyNotifications`, the ViewModel and all its registered children will be silenced until the end of the silence scope.
During the first pass, scripts that are detected that are earlier than the current script are given a length of 1 to prevent recursion.
The script parser can now produce a second pass on these addresses if they're recognized as scripts, to make sure that the full script gets concatenated rather than truncated to the first command.
