#jira UE-191009
#test Did repro steps from the associated UDN, and from the UDN associated to the CL that introduced the regression. Both are fixed.
[CL 26571478 by yohann dossantos in 5.3 branch]
The code presently creates two nodes as long as one of them exceeds the minimum dimension, which in practice this has resulted in us eventually having to iterate over 94+ useless elements every time we're adding a glyph. We should only create the smaller node if it exceeds the minimum.
[CL 26414944 by evgenii babinets in 5.3 branch]
Note: This doesn't impact any OS specific dialogs
Breaking: FCoreDelegates::ModalErrorMessage is now called FCoreDelegates::ModalMessageDialog and takes an extra EAppMsgCategory argument
#jira
#rb Dave.Belanger
[CL 26322797 by jamie dale in 5.3 branch]
Removed a reference counting workaround that was added in CL# 3718295 for an issue that was fixed in CL# 3718782.
Removed unnecessary reference count churn from TDecl, and thus every SNew call.
#rb partick.boutot
[CL 26088858 by steve robb in 5.3 branch]
- Converted the font size to float, to allow specifying more precise size, and support the sizes that would be wanted at (for instance) 72 DPI, without losing information (ie rounding to the closest 96DPI int).
- Change various font caches behavior (glyph, advance, kerning...) to match behavior of Freetype, and ensure that a given size in Freetype will be stored only once in the font caches.
- Added cvar commands:
- Slate.DumpFontCacheStats: will display the list of all font combinations used, with the number of glyphs for each combination.
- Slate.FlushFontCache: will clear the font caches as well as the font atlases, then will invalidate all the widgets to force the cache to be refilled.
- Fixed the combo box for font typeface selection not disabled when necessary.
#tests:
- changed the Font display settings between various values and customDPI, validatet that everything was save correctly, and that the display was reacting to the value change.
- tested all the tooltips, simple, extended (ctrl + alt), and dynamic (display of the current DPI setting next to the font size).
- validated that the font atlas/cache was containing not more than one time each glyph rendered by Freetype.
- validated that the edition of the font size was still fully functionnal (spinbox, undo/redo, support of disable state,...).
- tested default value for a new project (72DPI).
- compiled Slate/SlateCore/UMG with UnsafeTypeCastWarningLevel = WarningLevel.Warning to fix warnings related to the change.
- tested in a game with latest changes from BP team (CL 26034633), ensuring that there are no regression related to the change of type (int to float) from the Font size.
- running with it in a game and UE branch for months.
#jira UE-152161
#rb patrick.boutot
[CL 26074535 by yohann dossantos in 5.3 branch]
* Moved dll export from type to methods/static variables. This is changelist is produced by a script and contains 900+ files.
#rb none
[CL 25898513 by henrik karlsson in ue5-main branch]
Bonus: added the ability to scale the fonts in the preview from the font editor, for clearer preview.
#jira UE-184155 Expose Ascent/Descent overides
#rb vincent.gauthier
#tests Using the font provided in the jira and the new settable values, ensured that I was able to get result similar to a 'standard' font, on single line text, multi line text, rich text combining the problematic font and another one, as well as using a composite font mixing the fonts. Tested with different size. Validated that modifying the new font settings was refreshing the font displayed. Replaced in a game project the font "burbank" by the new font and got it properly vertically aligned using new settings.
#preflight 647a377ed4596f242c113c75
[CL 25790038 by yohann dossantos in ue5-main branch]
-----------------------------------------------------------------
Slate batching improvements for large element sets
---
Group similar elements at compile time using tuples instead of maps (Also preps for slate payload removal).
Merge similar renderbatches in same-typed element buckets (Currently for box and lines only).
Reserve verticies / elements in a batch before using them
Add some comments to invalidation.
#rb Patrick.Boutot
#preflight 646ff8505b484acfee7cc9fa
#jira UE-173455
[CL 25738553 by daren cheng in ue5-main branch]
[FYI] daren.cheng
Original CL Desc
-----------------------------------------------------------------
Re-submit 25712585 with manual tuple expansion to prevent use before define.
-----------------------------------------------------------------
Slate batching improvements for large element sets
---
Group similar elements at compile time using tuples instead of maps
Merge similar renderbatches in same-typed element buckets (Currently for box and lines only).
Reserve verticies / elements in a batch before using them
Add some comments to invalidation.
#rb Patrick.Boutot
#preflight 646ff8505b484acfee7cc9fa
#jira UE-173455
[CL 25720199 by daren cheng in ue5-main branch]
-----------------------------------------------------------------
Slate batching improvements for large element sets
---
Group similar elements at compile time using tuples instead of maps
Merge similar renderbatches in same-typed element buckets (Currently for box and lines only).
Reserve verticies / elements in a batch before using them
Add some comments to invalidation.
#rb Patrick.Boutot
#preflight 646ff8505b484acfee7cc9fa
#jira UE-173455
[CL 25717402 by daren cheng in ue5-main branch]
[FYI] daren.cheng
Original CL Desc
-----------------------------------------------------------------
Slate batching improvements for large element sets
---
Group similar elements at compile time using tuples instead of maps
Merge similar renderbatches in same-typed element buckets (Currently for box and lines only).
Reserve verticies / elements in a batch before using them
Add some comments to invalidation.
#rb Patrick.Boutot
#preflight 646ff8505b484acfee7cc9fa
#jira UE-173455
[CL 25716613 by daren cheng in ue5-main branch]
---
Group similar elements at compile time using tuples instead of maps
Merge similar renderbatches in same-typed element buckets (Currently for box and lines only).
Reserve verticies / elements in a batch before using them
Add some comments to invalidation.
#rb Patrick.Boutot
#preflight 646ff8505b484acfee7cc9fa
#jira UE-173455
[CL 25712708 by daren cheng in ue5-main branch]
#preflight 647773ab0d55081f5494ab67
#jira none
--------------------------------------------------
Add ability to add a widget inside the global window overlay
#rb Patrick.Boutot
#ushell-cherrypick of 25473883 by kristof.morva1
----------------------------------------------
Fix ensure when hiding mouse capture label after exiting PIE
#rb Vincent.Gauthier
#ushell-cherrypick of 25507685 by kristof.morva1
[CL 25709904 by ben hoffman in ue5-main branch]
[FYI] yohann.dossantos
Original CL Desc
-----------------------------------------------------------------
Fix incorrectly displayed text. In Slate, wrapping text needs 2 draws to be displayed properly, otherwise some incorrect value are used. In the case of this bug, the text was offseted with a value based on the text width without wrapping (as if it was single line) instead of the correct width with wrapping applied. An invalidation of the widget was called, but as the invalidation was not supported in this code path (WidgetComponent + WidgetRendere::DrawWindow), it was ignored.
Now the WidgetRenderer::DrawWindow support Invalidation Root, and the WidgetComponent enabled the Invalidation on the VirtualWindow it uses, and do only the necessary amount of Tick.
#jira UE-FORT-594200
#rb patrick.boutot
#tests Repro steps from the jira, no longer occurs. Tested that hit testing was still working properly by using a widget with buttons in the widgetcomponent. Tested with and without the new setting bIsInvalidationEnabled activated.
#preflight 646664b7a133ca477bc93c69
[CL 25545951 by yohann dossantos in ue5-main branch]
