1 files changed, 226 insertions, 0 deletions
diff --git a/src/lib/librexgettext b/src/lib/librexgettext
new file mode 100755
index 0000000..db575d6
--- /dev/null
+++ b/src/lib/librexgettext
@@ -0,0 +1,226 @@
+#!/usr/bin/env bash
+# Copyright (C) 2013-2016 Luke Shumaker <lukeshu@sbcglobal.net>
+#
+# License: GNU GPLv2+
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+export TEXTDOMAIN='librelib'
+export TEXTDOMAINDIR='/usr/share/locale'
+
+default_simple=(
+	--keyword={eval_,}{gettext,'ngettext:1,2'}
+	--keyword={_,print,term_title}
+	--keyword={msg,msg2,warning,error,stat_busy,die}
+	--keyword={lock,slock}:3
+)
+default_prose=(--keyword={prose,bullet})
+
+readonly default_simple default_prose
+
+if ! type gettext &>/dev/null; then
+	gettext() { echo "$@"; }
+fi
+
+errusage() {
+	if [[ $# -gt 0 ]]; then
+		fmt="$(gettext "$1")"; shift
+		printf "${0##*/}: $fmt\n" "$@"
+	fi
+	usage >&2
+}
+
+usage() {
+	. libremessages
+	print 'Usage: %s [OPTIONS] FILES...' "${0##*/}"
+	print 'Generates .pot files for programs using libremessages'
+	echo
+	prose 'In librexgettext, there are 2 types of keywords:'
+	bullet 'simple: Simple keywords are just like normal xgettext'
+	bullet 'prose: Prose keywords are similar, but the text is
+	        word-wrapped'
+	prose 'The keyword format is the same as in GNU xgettext.'
+	echo
+	prose 'The libremessages `flag` command is also handled
+	       specially, and is not configurable as a keyword.'
+	echo
+	prose 'The default simple keywords are: %s' "${default_simple[*]#--keyword=}"
+	echo
+	prose 'The default prose keywords are: %s' "${default_prose[*]#--keyword=}"
+	echo
+	print 'Options:'
+	flag \
+	    '--simple=KEYWORD' 'Look for KEYWORD as an additional simple keyword' \
+	    '--prose=KEYWORD' 'Look for KEYWORD as an additional prose keyword' \
+	    '-k' 'Disable using the default keywords' \
+	    '-h, --help' 'Show this text'
+}
+
+xgettext-sh() {
+	xgettext --omit-header --from-code=UTF-8 -L shell -k -o - "$@"
+}
+
+xgettext-flag() {
+	{
+		# Stage 1: Generate
+		#
+		# Get all of the arguments to `flag`.  Because `flag`
+		# takes an arbitrary number of arguments, just iterate
+		# through arg1, arg2, ... argN; until we've come up
+		# empty 3 times.  Why 3?  Because each flag takes 2
+		# arguments, and because we don't keep track of which
+		# one of those we're on, waiting for 3 empties ensures
+		# us that we've had a complete "round" with nothing.
+		#
+		# Why can't I just do i+=2, and not have to keep track
+		# of empties?  Because, we also allow for arguments
+		# ending in a colon to be headings, which changes the
+		# offsets.
+		declare -i empties=0
+		declare -i i
+		for (( i=1; empties < 3; i++ )); do
+			local out
+			out="$(xgettext-sh --keyword="flag:$i,\"$i\"" "$@")"
+			if [[ -n $out ]]; then
+				printf -- '%s\n' "$out"
+			else
+				empties+=1
+			fi
+		done
+	} | whitespace-collapse | sed '/^\#, sh-format/d' | {
+		# Stage 2: Parse
+		#
+		# Read in the lines, and group them into an array of
+		# (multi-line) msgs.  This just makes working with
+		# them easier.
+		local msgs=()
+		declare -i i=-1
+		local re='^#\. ([0-9]+)$'
+		IFS=''
+		local line
+		while read -r line; do
+			if [[ $line =~ $re ]]; then
+				i+=1
+			fi
+			msgs[$i]+="$line"$'\n'
+		done
+		# Stage 3: Sort
+		#
+		# Now, we have the `msgs` array, and it is
+		# sorted such that it is all of the arg1's to `flag`,
+		# then all of the arg2's, then all of the arg3's, and
+		# so on.  We want to re-order them such that it's all
+		# of the args for the first invocation then all of the
+		# args for the second; and so on.
+		#
+		# We do this by simply sorting them by the location
+		# that they appear in the file.  Then, when we see the
+		# argument number go back down to 1, we know that a
+		# new invocation has started!
+		IFS=$'\n'
+		local locations=($(
+			local i
+			for i in "${!msgs[@]}"; do
+				declare -i arg row
+				local lines=(${msgs[$i]})
+				arg=${lines[0]#'#. '}
+				row=${lines[1]##*:}
+				printf '%d.%d %d\n' "$row" "$arg" "$i"
+			done | sort -n
+		))
+		# Stage 4: Output
+		#
+		# Now, we prune out the arguments that aren't
+		# localizable.  Also, remove the "#." comment lines.
+		# As explained above (in stage 3), when we see $arg go
+		# to 1, that's the beginning of a new invocation.
+		local expectflag=true
+		local location
+		for location in "${locations[@]}"; do
+			IFS=' .'
+			local row arg i
+			read -r row arg i <<<"$location"
+			local msg="${msgs[$i]#*$'\n'}"
+			# Now we operate based on $row, $arg, and $msg
+			if [[ $arg == 1 ]]; then
+				expectflag=true
+			fi
+			if $expectflag; then
+				IFS=$'\n'
+				local lines=(${msg})
+				if [[ ${lines[1]} == *':"' ]]; then
+					# We expected a flag, but got
+					# a heading
+					printf -- '%s\n' "$msg"
+				else
+					# We expected a flag, and got
+					# one!
+					expectflag=false
+				fi
+			else
+				printf -- '%s\n' "$msg"
+				expectflag=true
+			fi
+		done
+	}
+}
+
+whitespace-collapse() {
+	tr '\n' '\r' | sed 's/"\r\s*"//g' | tr '\r' '\n' | # This removes the awkward word-wrapping done by xgettext
+	    sed -r -e 's/(\\n|\\t|\t)/ /g' -e 's/(^|[^.!? ]) +/\1 /g' -e 's/([.!?])  +/\1  /g' # This collapses whitespace
+}
+
+main() {
+	local simple=()
+	local prose=()
+	local files=()
+	local use_defaults=true
+	local error=false
+
+	declare -i i
+	for (( i=1; i <= $#; i++ )); do
+		case "${!i}" in
+			--simple) i+=1; simple+=(--keyword="${!i}");;
+			--simple=*)     simple+=(--keyword="${!i#*=}");;
+			--prose) i+=1; prose+=(--keyword="${!i}");;
+			--prose=*)     prose+=(--keyword="${!i#*=}");;
+			-k) use_defaults=false;;
+			--help|-h) usage; return 0;;
+			--) i+=1; break;;
+			-*) errusage "unrecognized option: %s" "${!i}"; error=true;;
+			*) files+=("${!i}");;
+		esac
+	done
+	files+=("${@:$i}")
+	if [[ ${#files[@]} -lt 1 ]]; then
+		errusage "no input file given"
+		error=true
+	fi
+	if "$error"; then
+		return 1
+	fi
+	if "$use_defaults"; then
+		simple+=("${default_simple[@]}")
+		prose+=("${default_prose[@]}")
+	fi
+
+	# Main code
+	{
+		xgettext-sh "${simple[@]}" -- "${files[@]}"
+		xgettext-sh "${prose[@]}" -- "${files[@]}" | whitespace-collapse
+		xgettext-flag -- "${files[@]}"
+	} | sed '/^\#, sh-format/d' | msguniq -Fi --to-code=UTF-8
+}
+
+main "$@"