On Wed, Jan 21, 2026 at 05:25:52PM +0100, [email protected] wrote: > commit af043eaf6ca068f0c4f0f011400d1c30d5f72e5a > Author: elbachir-one <[email protected]> > Date: Wed Jan 21 17:24:26 2026 +0100 > > [st][patches][fontmetrics] Fixed index.md
I ran the page through smu before committing. Did something break? > -This may or may not make the font rendering, particularly block and > box-drawing elements, better or worse. > +Depending on the font, this may improve or worsen font rendering, > particularly > +for block and box-drawing characters. This is an intentional bit of phrasing, because in a significant fraction of the fonts I've checked, this method ultimately yields the exact same values (Fira Code, Go Mono, Iosevka). I'll release a write-up on differences in usage between fonts when (if) I get to it, eventually. > -There is no standard for how monotype font files are supposed to communicate > correct metrics to terminal emulators, > -and actual usage is inconsistent between fonts. > +There is no strict standard for how monospaced font files should communicate > correct > +metrics to terminal emulators, and real-world usage is inconsistent across > fonts. Is there a... *non-strict* standard of some kind? I wrote that based on what I could find in the Microsoft and Apple TrueType/OpenType documentation, which is irrelevant to this use case at best, and contradictory at worst. > -`st` normally uses the descent and ascent heights provided by FreeType, > -which seems to derive them from `hhea` table values. > +By default, `st` uses the ascent and descent values provided by FreeType, > +which appear to be derived from the `hhea` table. I've only really dug through libXft to verify that it passes the numbers from FreeType, and then compared FT output to values read directly from font files. Could they actually be calculated somewhere else? > Download > -------- > +* [st-fontmetrics-0.9.3.diff](st-fontmetrics-0.9.3.diff) > > -- [st-fontmetrics-0.9.3.diff](st-fontmetrics-0.9.3.diff) (0.9.3) > - > -Authors > -------- > - > -- Paul Storkman - <[email protected]> > +Author > +------ > +* Paul Storkman - <[email protected]> > > -- Storkman
