Age | Commit message (Collapse) | Author |
|
|
|
|
|
This lets me rewrite the implementation during the 4.x cycle.
|
|
|
|
This reverts commit a0ec39ff
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Also adds nicer tests with client-side commands.
/libgui config: opens the config GUI without Mod Menu
/libgui tab: opens a tab test GUI
|
|
|
|
|
|
The new field is used for both fullscreen and normal screens.
Also fixes titles being incorrectly positioned by default.
|
|
* 21w20a. Fixes #109
* Indentation and spacing fixes
|
|
|
|
Co-authored-by: CoolMineman <62723322+coolmineman@users.noreply.github.com>
Closes #101.
|
|
|
|
|
|
|
|
Closes #100.
|
|
|
|
|
|
|
|
This reverts commit 1a5d29ec
|
|
|
|
|
|
|
|
|
|
This means that mouse events are propagated upwards the widget tree
until they are processed or the root panel is reached.
Code duplication in mouse handling was reduced by moving all logic
into a new MouseInputHandler helper class. (I'd obviously have them
directly in the screen classes if there was only one.)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Changed the MatrixStack translation to direct GL translation
via RenderSystem. I've already fixed this bug in the 4.0 branch
by using MatrixStacks for texture rendering, but that won't work here.
|
|
|
|
|
|
Apparently having a client-only-typed field referenced in a lambda body
can be dangerous, so I moved the packet registration to the client
init class.
|