bin:no-ansi: Don't just match CSI sequences

Try to match all ANSI escape sequences, not just CSI (i.e. `\e[`) - also
the ones that are followed by additional bytes (e.g. CSI, OSC).
This commit is contained in:
2025-09-18 23:54:21 +02:00
parent 6b866f5474
commit 8aa76b3ee6

View File

@@ -3,5 +3,21 @@
# Copyright (c) 2025 Julian Prein
#
# Remove ANSI escape sequences.
#
# Additionally to the general form `\e[ -/]*[0-~]` this also tries to cover
# sequences that are followed by additional bytes (e.g. CSI, OSC). This script
# covers no C1 sequences.
#
# See:
# - https://en.wikipedia.org/wiki/ANSI_escape_code
# - https://www.ecma-international.org/wp-content/uploads/ECMA-35_6th_edition_december_1994.pdf
#
# TODO: which bytes do DCS, OSC, SOS, PM & APC (second pattern) actually allow?
# Find some documentation, since `[ -~]` is guessed.
env LC_ALL=C sed "$(printf 's/\033\\[[0-?]*[ -/]*[@-~]//g')"
env LC_ALL=C sed -E "$(printf "%b" \
's/' \
'\033\\[[0-?]*[ -/]*[@-~]' '|' \
'\033[]PX^_][ -~]*\033\\\\' '|' \
'\033[ -/]*[0-~]' \
'//g')"