Landon Curt Noll
aa263427e3
Added `bin/sgit`. This directory is NOT intended to become a huge software project right now. For now it is just a convenient place to store tools. Please avoid overthinking, over documenting tools under bin. Later on there will be a tool that builds links in the web site from JSON files, perhaps forms HTML files from markdown files, etc. However we will have to make major progress on issue #1, issue #2, issue #3 and issue #4 before such code can be written. In some cases tools put under bin might go away later on if they are no longer needed.
73 lines
2.7 KiB
Bash
Executable File
73 lines
2.7 KiB
Bash
Executable File
#!/usr/bin/env bash
|
|
#
|
|
# sgit - sed on all files under git control
|
|
#
|
|
# usage: sgit <sed command...> <glob>
|
|
#
|
|
# sed must be GNU sed and we use -i'' just in case it's not. Even so sed under
|
|
# macOS will fail. This script could be improved to check for GNU sed and it
|
|
# might be an easy way to fix it in other sed implementations but it works for
|
|
# my purposes and since I want to work on the reason for this script (and not
|
|
# the script itself) I'm not going to worry about that.
|
|
#
|
|
# DISCLAIMER AND WARNING: this is hack and most likely error prone!
|
|
#
|
|
# - Cody Boone Ferguson (@xexyl)
|
|
#
|
|
|
|
# firewall
|
|
#
|
|
# first check that this is a git repo!
|
|
|
|
git status 2>/dev/null 1>&2
|
|
status="$?"
|
|
if [[ "$status" -ne 0 ]]; then
|
|
echo "$(basename "$0"): ERROR: ${PWD} not a git repository" 1>&2
|
|
exit 1
|
|
fi
|
|
|
|
# check number of args
|
|
if [[ "$#" -lt 2 ]]; then
|
|
echo "$(basename "$0"): ERROR: usage: $(basename "$0") <sed command...> <glob>" 1>&2
|
|
exit 2
|
|
fi
|
|
|
|
export SED_COMMANDS=""
|
|
export GLOB=""
|
|
# go through the command line, extracting the sed commands. When we've reached
|
|
# the end (one arg left) we know we have the glob to pass to git ls-files.
|
|
# We do it this way because apparently $BASH_ARG[VC] are not set unless in
|
|
# extended debugging mode (shopt -s extdebug).
|
|
i=0;
|
|
while [[ "$i" -le "$#" ]]; do
|
|
SED_COMMANDS="$SED_COMMANDS -e $1";
|
|
((i++))
|
|
shift 1
|
|
done
|
|
GLOB="$1"
|
|
|
|
# This is not the best performance because it does this on all files of the glob
|
|
# rather than only those files matched by the glob with matching text but it's
|
|
# better than it used to be (it used to run git ls-files on the repo for each
|
|
# sed command so the more you had the slow it would become as it had to get the
|
|
# list of files each time!).
|
|
#
|
|
# A better approach would be to somehow use git grep to find files that match
|
|
# but this would complicate the command line (at least in anyway I can think of
|
|
# how to do it) so we do it on all files in the repo.
|
|
#
|
|
# But these would possibly require additional args to the command and
|
|
# unnecessarily complicate matters. Anyway it's a hack so it doesn't have to be
|
|
# perfect (not that there even is such a thing as perfect).
|
|
#
|
|
# NOTE: it is a burden on the user to specify a glob that will only match
|
|
# regular files. Although we could try and figure out if each matching file is a
|
|
# regular file it would mean we would have to go through the list a line at a
|
|
# time and then add it to a new list if it's a regular file and this seems
|
|
# unnecessary. Besides doing '.' will not cause a problem so it might be fine.
|
|
|
|
# shellcheck disable=SC2086
|
|
# SC2086 (info): Double quote to prevent globbing and word splitting. We can't
|
|
# disable this because we need to have word splitting.
|
|
git ls-files "$GLOB" | xargs sed -i'' $SED_COMMANDS
|