ImGuiTextBuffer: Added append() function (unformatted).

Examples: Win32: Added support for mouse buttons 4 and 5 via WM_XBUTTON* messages. (#2264)

ImFontAtlas: Added 0x2000-0x206F general punctuation range to default ChineseFull/ChineseSimplifiedCommon ranges. (#2093)

Examples: Win32: Using IsChild() to be compatible with windows created within a parent. (#1951, #2087, #2156, #2232)

Examples: DirectX9: Explicitly disable fog (D3DRS_FOGENABLE) before drawing in case user state has it set. (#2288, #2290)

EditorConfig: Further tweaks (#2038)

ImDrawList: Fixed AddCircle(), AddCircleFilled() angle step being off, which was visible when drawing a "circle" with a small number of segments (e.g. an hexagon). (#2287) [@baktery] ...

+ Demo tweaks

FreeType: Minor tweaks previous commit (#2285)

Examples: Win32: Added support for XInput games (if ImGuiConfigFlags_NavEnableGamepad is enabled).

Examples: Win32: Using GetForegroundWindow() instead of GetActiveWindow() to be compatible with windows created in a different thread. (#1951, #2087, #2156, #2232) [many people]

Version 1.68 WIP

Version 1.67

DragFloat: Fixed broken mouse direction change with power!=1.0. (#2174, #2206) [@Joshhua5]

ImFontAtlas: Rewrote FreeType based builder. ...

- Fixed abnormally high atlas height. (#618)
- Fixed support for any values of TexGlyphPadding (not just only 1). (#618)
- Atlas width is now properly based on total surface rather than glyph count (unless overridden with TexDesiredWidth). (#618)
- Fixed atlas builder so missing glyphs won't influence the atlas texture width. (#2233, #618)
- Fixed atlas builder so duplicate glyphs (when merging fonts) won't be included in the rasterized atlas. (#618)

ImFontAtlas: Rewrote stb_truetype based builder. ...

- Atlas width is now properly based on total surface rather than glyph count (unless overridden with TexDesiredWidth).
- Fixed atlas builder so missing glyphs won't influence the atlas texture width. (#2233)
- Fixed atlas builder so duplicate glyphs (when merging fonts) won't be included in the rasterized atlas.

ImDrawList: Optimized some of the functions for performance of debug builds where non-inline function call cost are non-negligible.

Fixed animated window titles from being updated when displayed in the CTRL+Tab list. + Adding overkill helpers for reusing buffers. (#787)

Tweak changelog + tweak internal render helper functions.

Renamed ImFont::GlyphRangesBuilder to ImFontGlyphRangesBuilder. Keep redirection typedef (will obsolete).

IO: Renamed InputCharacters[], marked internal as was always intended. AddInputCharacter() goes into a queue which can receive as many characters as needed during the frame. This is useful for automation to not have an upper limit on typing speed. Will later transition key/mouse to use the event queue later.

Nav: Fixed an keyboard issue where holding Activate/Space for longer than two frames on a button would unnecessary keep the focus on the parent window, which could steal it from newly appearing windows. (#787)

Examples: Allegro 5: Properly destroy all globals on shutdown. (#2262)

InputFloat: When using ImGuiInputTextFlags_ReadOnly the step buttons are disabled. (#2257)

Error recovery: Extraneous/undesired calls to End() are now being caught by an assert in the End() function itself at the call site (instead of being reported in EndFrame). Past the assert, they don't lead to crashes any more. Missing calls to End(), pass the assert, should not lead to crashes any more, nor to the fallback/debug window appearing on screen. (#1651).

Made it illegal to call Begin("") with an empty string. This somehow accidentally worked before but had various undesirable side-effect as the window would have ID zero. In particular it is causing problems in viewport/docking branches.

IO: Added BackendPlatformUserData, BackendRendererUserData, BackendLanguageUserData void* for storage use by back-ends. (#2004 + for cimgui)

Fixed using SetNextWindowPos() on a child window (which wasn't really documented) position the cursor as expected in the parent window, so there is no mismatch between the layout in parent and the position of the child window. Demo tweak and adding some child window stuff

Window, Focus, Popup: Fixed an issue where closing a popup by clicking another window with the _NoMove flag would refocus the parent window of the popup instead of the newly clicked window.

Internals: Windows hidden with HiddenFramesRegular (but NOT HiddenFramesForResize) preserve their SizeContents, so restoring a auto-resize window after it's been hidden by tabs won't reset its size for a frame. Arguable. Let's see how it goes. (Followup to b48e295)

Contents size is preserved while a window collapsed. Fix auto-resizing window losing their size for one frame when uncollapsed.