Don't rely on xmlstarlet transparently decompressing files

This feature was removed by default in libxml2 2.15.0 and decompression
must now be explicitly done. Fortunately, gzip offers a mode where it
will decompress when possible or just pass the data through if
uncompressed, so all file types that are or might be compressed now call
gzip to read the data before passing it to xmlstarlet.

Author: dfandrich

autodescribe

@@ -733,25 +733,12 @@ comment_stl () {
 }
 
 # File type: svg (Scalable Vector Graphics image)
-# requires: xmlstarlet
-comment_svg () {
-	COMMENT=$(xmlstarlet sel -t -v "/*[local-name()='svg']/*[local-name()='title']" < "$1" 2> /dev/null)
-	if [ -z "$COMMENT" ] ; then
-		# Try for title in metadata
-		COMMENT=$(xmlstarlet sel -t -v "/*[local-name()='svg']/*[local-name()='metadata']/*[local-name()='RDF']/*[local-name()='Work']/*[local-name()='title']" < "$1" 2> /dev/null)
-	fi
-}
-
-# File type: svgz (Compressed Scalable Vector Graphics image)
 # requires: gzip, xmlstarlet
-# TODO: xmlstarlet can optionally transparently decompress, but that feature is
-# likely to be removed some time after 2025 so we shouldn't rely on it here (or
-# for certain other file types).
-comment_svgz () {
-	COMMENT=$(gzip -dc < "$1" | xmlstarlet sel -t -v "/*[local-name()='svg']/*[local-name()='title']" 2> /dev/null)
+comment_svg () {
+	COMMENT=$(gzip -dcf < "$1" | xmlstarlet sel -t -v "/*[local-name()='svg']/*[local-name()='title']" 2> /dev/null)
 	if [ -z "$COMMENT" ] ; then
 		# Try for title in metadata
-		COMMENT=$(gzip -dc < "$1" | xmlstarlet sel -t -v "/*[local-name()='svg']/*[local-name()='metadata']/*[local-name()='RDF']/*[local-name()='Work']/*[local-name()='title']" 2> /dev/null)
+		COMMENT=$(gzip -dcf < "$1" | xmlstarlet sel -t -v "/*[local-name()='svg']/*[local-name()='metadata']/*[local-name()='RDF']/*[local-name()='Work']/*[local-name()='title']" 2> /dev/null)
 	fi
 }
 
@@ -1084,9 +1071,9 @@ comment_gif () {
 }
 
 # File type: gnumeric (Gnumeric spreadsheet)
-# requires: xmlstarlet
+# requires: gzip, xmlstarlet
 comment_gnumeric () {
-	COMMENT=$(xmlstarlet sel -N gnm=http://www.gnumeric.org/v10.dtd -N office=urn:oasis:names:tc:opendocument:xmlns:office:1.0 -N dc=http://purl.org/dc/elements/1.1/ -t -v '/gnm:Workbook/office:document-meta/office:meta/dc:title' < "$1")
+	COMMENT=$(gzip -dcf < "$1" | xmlstarlet sel -N gnm=http://www.gnumeric.org/v10.dtd -N office=urn:oasis:names:tc:opendocument:xmlns:office:1.0 -N dc=http://purl.org/dc/elements/1.1/ -t -v '/gnm:Workbook/office:document-meta/office:meta/dc:title')
 }
 
 # File type: gpx (GPX GPS track)
@@ -1695,9 +1682,7 @@ for f in "$@" ; do
 				;;
 			*.stl)	TYPE=stl
 				;;
-			*.svg)	TYPE=svg
-				;;
-			*.svgz)	TYPE=svgz
+			*.svg | *.svgz)	TYPE=svg
 				;;
 			*.tar) TYPE=tar
 				;;

automtime

@@ -327,9 +327,9 @@ mtime_bgcode () {
 }
 
 # File type: gnumeric (Gnumeric spreadsheet)
-# requires: xmlstarlet
+# requires: gzip, xmlstarlet
 mtime_gnumeric () {
-	RAWTIME=$(xmlstarlet sel -N gnm=http://www.gnumeric.org/v10.dtd -N office=urn:oasis:names:tc:opendocument:xmlns:office:1.0 -N dc=http://purl.org/dc/elements/1.1/ -t -v '/gnm:Workbook/office:document-meta/office:meta/dc:date' < "$1")
+	RAWTIME=$(gzip -dcf < "$1" | xmlstarlet sel -N gnm=http://www.gnumeric.org/v10.dtd -N office=urn:oasis:names:tc:opendocument:xmlns:office:1.0 -N dc=http://purl.org/dc/elements/1.1/ -t -v '/gnm:Workbook/office:document-meta/office:meta/dc:date')
 	# $RAWTIME is like 2022-04-04T17:07:51Z
 	TIME="$(normalize_iso_time "$RAWTIME")"
 }
@@ -525,9 +525,9 @@ mtime_kra () {
 }
 
 # File type: kmy (KMyMoney file)
-# requires: xmlstarlet
+# requires: gzip, xmlstarlet
 mtime_kmy () {
-	RAWTIME=$(xmlstarlet sel -t -v /KMYMONEY-FILE/FILEINFO/LAST_MODIFIED_DATE/@date < "$1" )
+	RAWTIME=$(gzip -dcf < "$1" | xmlstarlet sel -t -v /KMYMONEY-FILE/FILEINFO/LAST_MODIFIED_DATE/@date)
 	# $RAWTIME is like 2022-01-01
 	TIME="$RAWTIME"
 }
@@ -839,9 +839,9 @@ mtime_stardict () {
 }
 
 # File type: svg (Scalable Vector Graphics image)
-# requires: xmlstarlet
+# requires: gzip, xmlstarlet
 mtime_svg () {
-	RAWTIME=$(xmlstarlet sel -t -v "/*[local-name()='svg']/*[local-name()='metadata']/*[local-name()='RDF']/*[local-name()='Work']/*[local-name()='date']" < "$1" 2>/dev/null)
+	RAWTIME=$(gzip -dcf < "$1" | xmlstarlet sel -t -v "/*[local-name()='svg']/*[local-name()='metadata']/*[local-name()='RDF']/*[local-name()='Work']/*[local-name()='date']" 2>/dev/null)
 	# $RAWTIME can be just about anything, but DublinCore recommends ISO 8601-1.
 	# Anything else is too ambiguous, so ignore them.
 	if echo "$RAWTIME" | grep '^[012][0-9][0-9][0-9]-[0-1][0-9]-[0-3][0-9]' >/dev/null; then