Implement aesthetic break point selection to prefer natural break locations
(e.g., after operators) rather than arbitrary positions when line wrapping
mathematical expressions.
When rendering large operators (e.g., sum, integral) with scripts in text
mode, the operator glyph was incorrectly positioned after its subscripts
and superscripts instead of before them. This caused expressions like
\sum_{i=1}^{n} i = \frac{n(n+1)}{2} to render with the equals sign
appearing visually misplaced.
Root cause:
The line-breaking refactoring introduced double-positioning of large
operators. makeLargeOp() internally sets the operator position, advances
currentPosition.x, and adds script displays. However, the calling code
then overwrote the position and advanced currentPosition.x again, causing:
- Double-advancement leading to incorrect width calculations
- Scripts positioned before the operator instead of after
Solution:
Save and restore typesetter state before/after line break dimension checks,
then call makeLargeOp() once at the correct position after handling line
breaks and inter-element spacing.
Extends the width-checking pattern from fractions/radicals to ALL remaining
complex atom types, completing Priority 1 of the multiline implementation.
Changes:
- Large operators (∑, ∫, ∏): Now stay inline with height+width checking
(breaks only if height > fontSize * 2.5 OR width exceeds constraint)
- Delimiters (\left...\right): Stay inline with maxWidth propagation to
inner content for proper nested wrapping
- Colors (.color, .textcolor, .colorBox): All 3 types now stay inline with
maxWidth propagation for proper nested wrapping
- Matrices/tables: Small matrices can now stay inline with surrounding content
- Width constraint propagation: All recursive createLineForMathList() calls
now properly pass maxWidth parameter
Impact:
Before: Complex atoms always forced line breaks, even when they fit
After: ALL complex atoms intelligently stay inline when width permits
Examples:
- a + ∑ xᵢ + b → 1 line instead of 3
- (a+b) + \left(\frac{c}{d}\right) + e → stays inline with wrapping
Implement smart width-checking for complex mathematical displays to enable
inline rendering when space permits, dramatically improving multiline layout.
Changes:
- Add shouldBreakBeforeDisplay() helper to check width before line breaks
- Add performLineBreak() helper for clean line transitions
- Modify fraction handling to stay inline when they fit within maxWidth
- Modify radical handling to stay inline when they fit within maxWidth
- Support radicals with degrees (cube roots, nth roots, etc.)
Add display-style (dfrac) and text-style (tfrac) fraction commands
to SwiftMath's LaTeX parser. These commands force fractions to render
in specific styles regardless of context.
Implementation:
- Add dfrac parsing to prepend displaystyle to numerator/denominator
- Add tfrac parsing to prepend textstyle to numerator/denominator
- Implement in both parser functions in MTMathListBuilder.swift
Testing:
- Add testDisplayStyleFraction() for dfrac validation
- Add testTextStyleFraction() for tfrac validation
- Add testDisplayAndTextStyleFractions() for complex expressions
- All 180 tests pass on macOS and iOS simulator
Documentation:
- Update MISSING_FEATURES.md (7/12 features now implemented, 58%)
- Update README.md feature list to include dfrac and tfrac
Fixes issue where equations like y'=-\dfrac{2}{x^{3}} would fail to
parse with "Invalid command dfrac" error. This was blocking the
StepByStep feature preview rendering.
