Discussion
Continuing my reading on how Papyrix renders text...
There's a method for rendering truncated text which 1st sizes the text to see if it needs truncation, for Arabic/Thai (due to greater text-sizing nuances) iterates over all prefixes (longest 1st) to find the 1st one which fits with the appendeded ellipsis, or binary searches for that location, & returns the truncated text with the added ellipsis.
Then there's a suite of methods for breaking line of text into multiple.
1/?
A support routine for hyphenating a word checks if we've got a word to split & whether we've got Arabic or Thai text before repeatedly checking whether the remaining text fits, searching (approach depends again on whether its Arabic or Thai) for where best to split the word, & appends those results to the output with the hyphenation.
A wrapper method shortcircuits if there's no text or no available space, before repeatedly seeing if each word fits.
2/?
If the word does fit it remembers the potential breakpoint for the next iteration. If none do it hyphenates the word as just described. If the current word doesn't it accepts the remembered breakpoint. If it reached the end of string it ensures the last line is incorporated into the results.
After the (nested) loop it truncates the final line if appropriate.
There's methods to measure the font. And a method to iterate over the known hardware buttons & render rectangles & text for them.
3/?
There's a method (used during sizing text) to retrieve the width of a glyph within an "external font".
There's method to precache certain glyphs from an external font.
To measure Thai text it asks the Thai shaper to "shape" a list of "clusters" before summing up the advance-X for all the glyphs within. To draw Thai text it retrieves those clusters & a start-Y deferring each cluster to another method, which iterates over the glyphs.
4/5?
For each Thai glyph it retrieves the glyph from the font with a fallback, sums in an offset onto the X & Y positions, retrieves the bitmap & copies it over to the display pixel-by-pixel, & sums in the advance-X.
As for Arabic it calls the Arabic shaper & sums in the glyphs' advance-Xs. Whilst rendering involves rendering each of those glyphs using the shared method.
There's a shared method which twiddles bits to extract a pixel from a font's glyph.
5/5 Fin for today!
Install bonfire.cafe
Get the full app experience
