From 84c2699e22d9660ef5b6de54643c0c78f2e8e855 Mon Sep 17 00:00:00 2001 From: Gabriel Schneider Date: Tue, 17 Nov 2026 00:45:31 -0300 Subject: width: back to libvaxis's gwidth, the user's ruling (keeping a dependency beats reimplementing it in the core): vmkopprq reverted: the core measures a grapheme's cells with vaxis.gwidth.gwidth(.unicode) again (File.graphemeDisplayWidth and its neighbours, surface.zig, draw.zig's test), File.utf8Width and tui.zig's corpus test that held vaxis's width to it go, and build.zig gives libvaxis to the core module again, so vaxis stays a dependency of the core; the widths were the same on every Unicode scalar value (vmkopprq's proof), so every golden and snapshot is unchanged; the changelog's shared entry keeps the palette sentence alone --- build.zig | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'build.zig') diff --git a/build.zig b/build.zig index 26eb0c5c..0e67b61e 100644 --- a/build.zig +++ b/build.zig @@ -842,9 +842,7 @@ pub fn build(b: *std.Build) void { .mvzr = mvzr_mod, .c_heap = c_heap_mod, .ghostty_vt = ghostty_core_vt, - // The tui's and the gui's (main.zig's roots); the core measures - // text with uucode alone (File.utf8Width). - .vaxis = if (m == core_module) null else vaxis_mod, + .vaxis = vaxis_mod, .uucode = uucode_mod, .wuffs = wuffs_mod, .xitdb = xitdb_mod, -- cgit v1.3