Refactor conditional_get.

bin/cget.sh

@@ -25,36 +25,39 @@ display_help () {
 	This script retrieves and caches HTTP resources on disk. 
 	A previously cached resource is retrieved via HTTP Conditional 
 	GET [RFC 7232]. If the web server responds with HTTP 200 OK,
-	the resource is cached and written to stdout. OTOH, if the web 
-	server responds with 304 Not Modified, the cache is refreshed
-	but no output.
+	the resource is cached and written to stdout. If the web 
+	server responds with 304 Not Modified, the cached resource 
+	is output instead.
 	
 	Usage: ${0##*/} [-hv] [-F | -C] URL
 	
 	This script takes a single command-line argument. The URL 
-	argument is the absolute URL of an HTTP resource. By default, 
-	the script requests the resource at the given URL using the 
-	curl command-line tool.
+	argument is the absolute URL of an HTTP resource. The script 
+	requests the resource at the given URL using the curl 
+	command-line tool.
 	
 	This script requires two environment variables. CACHE_DIR is
 	the absolute path to the cache directory (which may or may not
 	exist) whereas LIB_DIR specifies a directory containing various 
 	helper scripts.
 
 	Options:
-	   -h      Display this message
+	   -h      Display this help message
 	   -v      Write verbose messages to stdout
 	   -F      Enables "Force Output Mode"
 	   -C      Enables "Cache Only Mode"
 
 	Option -h is mutually exclusive of all other options.
 	
 	The default behavior of the script may be modified by using 
-	option -F or -C, which are mutually exclusive. Force Output Mode 
-	outputs the response body even if the server response is 304. 
-	Cache Only Mode bypasses the GET request altogether and goes 
-	directly to cache. If the resource resides in cache, it is 
-	output on stdout, otherwise an error is thrown.
+	option -F or -C, which are mutually exclusive. Force Output 
+	Mode (option -F) forces the return of a fresh resource. The 
+	resource is output on stdout if and only if the server 
+	responds with 200. If the response is 304, an error is thrown.
+	 
+	Cache Only Mode (option -C) bypasses the GET request altogether 
+	and goes directly to cache. If the resource resides in cache, 
+	it is output on stdout, otherwise an error is thrown.
 	
 	LIBRARY
 	
@@ -66,9 +69,10 @@ display_help () {
 
 	EXAMPLES
 	
-	  ${0##*/} URL      # Retrieve the resource using HTTP conditional GET
-	  ${0##*/} -F URL   # Enable Force Output Mode
-	  ${0##*/} -C URL   # Enable Cache Only Mode
+	  url=http://md.incommon.org/InCommon/InCommon-metadata.xml
+	  ${0##*/} \$url      # Retrieve the resource using HTTP conditional GET
+	  ${0##*/} -F \$url   # Enable Force Output Mode
+	  ${0##*/} -C \$url   # Enable Cache Only Mode
 HELP_MSG
 }
 
@@ -81,8 +85,7 @@ script_name=${0##*/}  # equivalent to basename $0
 # library filenames (always list command_paths first)
 LIB_FILENAMES="command_paths.sh
 compatible_mktemp.sh
-http_tools.sh
-conditional_get.sh"
+http_tools.sh"
 
 #######################################################################
 # Process command-line options and arguments
@@ -182,34 +185,46 @@ for lib_filename in $LIB_FILENAMES; do
 	fi
 done
 
-# create a temporary directory
-tmp_dir=$( make_temp_file -d )
-if [ ! -d "$tmp_dir" ] ; then
-	printf "ERROR: $script_name unable to create temporary dir\n" >&2
-	exit 2
+# determine temporary directory
+if [ -n "$TMPDIR" ] && [ -d "$TMPDIR" ]; then
+	# use system temporary directory (remove trailing slash)
+	TMP_DIR="${TMPDIR%%/}/${script_name%%.*}"
+	$verbose_mode && printf "$script_name using temp dir: %s\n" "$TMP_DIR"
+else
+	# create temporary directory
+	tmp_dir="$( make_temp_file -d )"
+	if [ ! -d "$tmp_dir" ] ; then
+		printf "ERROR: $script_name unable to create temporary dir\n" >&2
+		exit 2
+	fi
+	# use temporary directory (remove trailing slash)
+	TMP_DIR="${tmp_dir%%/}/${script_name%%.*}"
+	$verbose_mode && printf "$script_name creating temp dir: %s\n" "$TMP_DIR"
+fi
+if [ ! -d "$TMP_DIR" ]; then
+	/bin/mkdir "$TMP_DIR"
+	exit_status=$?
+	if [ $exit_status -ne 0 ]; then
+		echo "ERROR: $script_name failed to create dir: $TMP_DIR" >&2
+		exit $exit_status
+	fi
 fi
-$verbose_mode && printf "$script_name creating temp dir: %s\n" "$tmp_dir"
+
+# temporary file
+tmp_file="${TMP_DIR}/http_resource_$$.xml"
 
 #######################################################################
 # Main processing
 #######################################################################
 
-# invoke the function
-response_body=$( 
-	conditional_get $local_opts -d "$CACHE_DIR" -t "$tmp_dir" "$location"
-)
+# get the resource
+conditional_get $local_opts -d "$CACHE_DIR" -t "$TMP_DIR" "$location" > "$tmp_file"
 exit_code=$?
 if [ $exit_code -ne 0 ]; then
-	echo "ERROR: ${script_name} failed to get resource: $location" >&2
+	echo "ERROR: $script_name failed to get resource: $location" >&2
+	printf "See output log: %s\n" "$TMP_DIR/$conditional_get_log" >&2
 	exit $exit_code
 fi
 
-if $verbose_mode; then
-	/bin/cat "$tmp_dir/conditional_get_log"
-else
-	if [ -n "$response_body" ]; then
-		echo "$response_body"
-	fi
-fi
-
+/bin/cat "$tmp_file"
 exit 0

install.sh

@@ -108,7 +108,6 @@ done <<SOURCES
 $script_bin/lib/command_paths.sh
 $script_bin/lib/compatible_date.sh
 $script_bin/lib/compatible_mktemp.sh
-$script_bin/lib/conditional_get.sh
 $script_bin/lib/http_tools.sh
 $script_bin/lib/md_tools.sh
 $script_bin/lib/extract_entity.xsl

lib/conditional_get.sh

@@ -23,10 +23,11 @@ fi
 
 #######################################################################
 #
-# Request a resource using HTTP Conditional GET [RFC 7232] if possible.
-# If the server responds with 200, cache the resource and return the 
-# response body. If the server responds with 304, cache the resource
-# but return nothing.
+# Given a web resource and a cache, if the resource is cached, request 
+# the resource using HTTP Conditional GET [RFC 7232], otherwise issue
+# an ordinary GET request. In either case, if the server responds with 
+# 200, cache the resource and return the response body. If the server 
+# responds with 304, return the cached resource instead.
 #
 # Usage: conditional_get [-v] [-F | -C] -d CACHE_DIR -t TMP_DIR HTTP_LOCATION
 #
@@ -36,11 +37,20 @@ fi
 #
 # Options:
 #   -v   verbose mode
-#   -F   force output the HTTP response body even if 304 response
-#   -C   check the cache only (do not request the resource)
+#   -F   force the return of a fresh resource
+#   -C   check the cache only
 #   -d   the cache directory (REQUIRED)
 #   -t   a temporary directory (REQUIRED)
 #
+# Use option -F or -C to alter the default behavior of the function.
+#
+# Option -F forces the return of a fresh resource, that is, if the
+# server responds with 304, an error occurs and the function returns
+# with a nonzero return code.
+#
+# Option -C causes the function to go directly to cache. No GET request
+# is issued. This option is useful in offline mode.
+#
 # The output of the curl command-line tool is stored in the following 
 # temporary files:
 #
@@ -58,7 +68,7 @@ fi
 # TODO:
 #   - follow redirects?
 #
-# This script is compatible with Mac OS and GNU/Linux.
+# This script is compatible with both Mac OS and GNU/Linux.
 #######################################################################
 
 conditional_get () {
@@ -68,7 +78,7 @@ conditional_get () {
 		return 2
 	fi
 
-	local script_version="0.5"
+	local script_version="0.6"
 	local user_agent_string="HTTP Conditional GET client $script_version"
 
 	local hash
@@ -367,9 +377,11 @@ conditional_get () {
 	#
 	#######################################################################
 
-	if [ "$response_code" = "200" ] || $force_output_mode; then
-		/bin/cat "${cached_content_file}"
+	if $force_output_mode && [ "$response_code" = "304" ]; then
+		echo "ERROR: $FUNCNAME failed to get fresh resource: $location" >&2
+		return 1
 	fi
 
+	/bin/cat "${cached_content_file}"
 	return 0
 }