Tags: gdamore/tcell
Tags
revert: "refactor(simscreen): reuse drawCell encode buffers" reverts 1ccc145. the lift moved 12-byte scratch buffers from per-call make to simscreen fields. measured impact varies by escape decision: in escape-forced callers (gitgum) it saves ~14% allocs, but in the standalone benchmark the original was already stack-allocated, and the indirection through the struct field measurably regresses Show time by 2-8%. dropping in favour of keeping the PR scoped to wins that are unambiguous on the standalone bench.
fix(tscreen): Last column snafu with workaround for wide characters. The hack to fix #976 caused another problem where the wide characters didn't render properly on the last column as the backspace did not go backwards with pending wrap state. (This might actually be considered a terminal bug as well.) We don't use the work around at the right edge of the screen. This may mean that buggy terminals will still see screwy behavior in the far right column, but at least it will be fixed everywhere else.
fix(tscreen): Last column snafu with workaround for wide characters. The hack to fix #976 caused another problem where the wide characters didn't render properly on the last column as the backspace did not go backwards with pending wrap state. (This might actually be considered a terminal bug as well.) We don't use the work around at the right edge of the screen. This may mean that buggy terminals will still see screwy behavior in the far right column, but at least it will be fixed everywhere else. (cherry picked from commit 49eaa4d)
fix: fix for wide not overwriting adjacent character (fixes #976) This is a workaround - and really if it this is happening it means that your terminal emulator is a bug. I'm not sure I want to merge this. I have not observed the behavior that this change is meant to fix.
PreviousNext