Scope and Limitations

Pretext is intentionally focused. It handles text preparation and line fitting; it does not try to replace a full text rendering stack.

What Pretext does

analyzes text into reusable segments
classifies whitespace and break opportunities
measures text through SkiaSharp
computes wrapped line counts, heights, and materialized lines
supports streamed line walking for custom layouts
carries locale-aware segmentation and bidi metadata where available

What Pretext does not do

draw text for you
choose line height automatically
own baseline alignment
implement selection, caret movement, or hit testing
expose a ready-made document model

Important limitations

Font strings are CSS-like, not full CSS

The font parser only understands a practical subset of CSS-style font strings. Unsupported syntax falls back to 16px Arial.

Locale-aware segmentation is desktop-oriented

ICU-backed locale-aware word segmentation is available on desktop targets when the needed ICU libraries are available. If ICU is unavailable, the engine falls back to its internal tokenization path.

`LayoutWithLines` allocates

If you are on a hot path and only need geometry, prefer WalkLineRanges or LayoutNextLine.

Prepared text is input-specific

Prepared objects are valid only for the text, font, white-space mode, and locale they were created with.

The engine is deterministic, but your host still matters

Rendering quality, baseline handling, and final visuals still depend on your host UI and paint configuration.