Rewrite

1 files changed, 134 insertions, 124 deletions

@@ -1,97 +1,105 @@
 #!/bin/sh
-# generate a Atom feed from a directory, recursively
-# using POSIX sh
-# AUTHOR: Case Duckworth <acdw@acdw.net>
-# LICENSE: MIT
+# SHATOM: generate a feed from a directory of files
+# Copyright (C) 2020--2022 Case Duckworth <acdw@acdw.net>
+# Licensed under the Good Choices License.  See COPYING for details.
 
-usage() {
-    cat <<END
-$0: generate an Atom feed from directories of files
-INVOCATION:
-        $0 [-h] [-c CONFIG] DIRECTORY...
-
-OPTIONS:
-        -h                      show this help
-        -c CONFIG       change the CONFIG file.
-                                Default: $PWD/$0.conf.sh
-END
-}
-
-# FEED VARIABLES
-FEED_TITLE=example
-FEED_SUBTITLE="an example"
-FEED_URL="https://example.com/atom.xml"
-SITE_URL="https://example.com"
-FEED_ID="${SITE_URL#*//}"
-FEED_AUTHOR="nobody"
-FEED_COPYRIGHT="(c) 2020 CC-BY-SA $FEED_AUTHOR"
-FEED_UPDATED=$(date -u +'%FT%TZ')
-
-# FUNCTIONS
-
-recent_files() { # recent_files DIRECTORY FIND_ARG...
-    # list files, in order of recency
-    # BOO: REQUIRES GNU FIND :( :( :( :(
-    # possibly look into https://unix.stackexchange.com/questions/9247/
-    # for more possibilities if you don't have GNU find.
-    dir="$1"
-    shift
-
-    find "$dir" -maxdepth 1 \
-        "$@" \
-        -not -name '.*' \
-        -printf '%T@\t%p\n' |
-        sort -nr |
-        cut -f2
-}
+### Entry point
 
-# ENTRY FUNCTIONS
+usage() {
+        cat <<EOF
+SHATOM: generate a web feed from directories, recursively
+USAGE:  shatom -h
+        shatom [-c CONFIG] [-C DIR] [VAR=VALUE...] DIRECTORY...
 
-skip_entry() {
-    return 1
-}
+FLAGS:
+ -h     Show this help and exit.
 
-entry_url() {
-    basename="${1#$DIRECTORY}"
-    echo "$FEED_URL/$basename"
+OPTIONS:
+ -c CONFIG      Specify shatom's configuration path, a shell
+                script overriding functions in shatom itself.
+                Default: \$PWD/shatom.conf.sh
+ -C DIR Change the current working directory to DIR
+                before executing the rest of the script.
+
+PARAMETERS:
+ VAR=VALUE...   Override configuration variables for this
+                invocation of shatom.  These are eval'd, so
+                be careful!
+ DIRECTORY...   Directory(s) containing feed content.  These
+                can be files as well, but really directories
+                makes more sense, don't you think?
+EOF
+        exit "${1:-0}"
 }
 
-entry_title() {
-    awk '/^#+[  ]\S/{
-        for(i=2;i<=NF;i++) {
-                printf $i;
-                if (i!=NF) printf " ";
-        }
-        printf "\n";exit}' "$1"
+main() {
+        # Command-line flags
+        while getopts hc: opt; do
+                case "$opt" in
+                h) usage ;;
+                c) CONFIG="$OPTARG" ;;
+                *) usage 1 ;;
+                esac
+        done
+        shift $((OPTIND - 1))
+
+        # Configuration file
+        if [ -f "${CONFIG:=$PWD/shatom.conf.sh}" ]; then
+                . "$CONFIG"
+        fi
+
+        # Command-line variable assignment
+        for arg; do
+                case "$arg" in
+                *=*) eval "$arg" ;;
+                *) break ;;
+                esac
+        done
+
+        # Finally, default values for variables
+        FEED_TITLE="${FEED_TITLE:-example}"
+        FEED_SUBTITLE="${FEED_SUBTITLE:-an example}"
+        FEED_AUTHOR="${FEED_AUTHOR:-nobody}"
+        FEED_COPYRIGHT="${FEED_COPYRIGHT:-Copyright (C) $(date +%Y) $FEED_AUTHOR}"
+        FEED_URL="${FEED_URL:-https://example.com/feed.xml}"
+        FEED_ID="${FEED_ID:-${FEED_URL#*//}}"
+        FEED_UPDATED="${FEED_UPDATED:-$(date -u +'%F %TZ')}"
+        SITE_URL="${SITE_URL:-https://example.com/}"
+
+        # Do the damn thing already
+        feed_header
+        for DIR; do
+                recent_files "$DIR" -type f |
+                        while read -r item; do
+                                skip_item "$item" && continue
+                                feed_item "$item"
+                        done
+        done
+        feed_footer
 }
 
-entry_summary() {
-    return 1
-}
+### Library functions
 
-entry_author() {
-    echo "$FEED_AUTHOR"
-}
+## File-listing functions
 
-entry_content() {
-    cat "$1"
-}
+recent_files() {
+        # requires stat(1)
+        dir="$1"
+        shift
 
-entry_updated() {
-    # requires stat(1).
-    # probably another way.
-    # possibly using ls(1).
-    stat -c '%y' "$1" |
-        awk '{
-                sub(/\..*/,"",$2);
-                sub(/[0-9][0-9]/,"&:",$3);
-                print $1"T"$2$3;}'
+        find "$dir"/* -prune "$@" \
+                -exec stat -c '%Y %N' '{}' + |
+                sort -k1,1 -nr |
+                cut -d' ' -f2-
 }
 
-# ATOM FUNCTIONS
+## Feed-level functions
+# by default, generate Atom feeds
+# (https://datatracker.ietf.org/doc/html/rfc4287), but an RSS configuration file
+# is included in rss.conf.sh.
 
-atom_header() {
-    cat <<END
+feed_header() {
+        cat <<END
 <?xml version="1.0" encoding="utf-8"?>
 <feed xmlns="http://www.w3.org/2005/Atom">
 <title>$FEED_TITLE</title>
@@ -105,60 +113,62 @@ atom_header() {
 END
 }
 
-atom_footer() {
-    cat <<END
+feed_footer() {
+        cat <<END
 </feed>
 END
 }
 
-atom_entry() { # atom_entry FILE
-    ENTRY_URL="$(entry_url "$1")"
-    ENTRY_TITLE="$(entry_title "$1")"
-    ENTRY_SUMMARY="$(entry_summary "$1")"
-    ENTRY_AUTHOR="$(entry_author "$1")"
-    ENTRY_CONTENT="$(entry_content "$1")"
-    ENTRY_UPDATED="$(entry_updated "$1")"
-    cat <<END
+feed_item() {
+        ITEM_URL="$(item_url "$1")"
+        cat <<END
 <entry>
-<id>$ENTRY_URL</id>
-<link rel="alternate" href="$ENTRY_URL" />
-<title>$ENTRY_TITLE</title>
-<summary>$ENTRY_SUMMARY</summary>
-<updated>$ENTRY_UPDATED</updated>
-<author><name>$ENTRY_AUTHOR</name></author>
-<content type="text"><![CDATA[$ENTRY_CONTENT]]></content>
+<id>$ITEM_URL</id>
+<link rel="alternate" href="$ITEM_URL" />
+<title>$(item_title "$1")</title>
+<summary>$(item_summary "$1")</summary>
+<updated>$(item_updated "$1")</updated>
+<author>$(item_author)</author>
+<content type="text"><![CDATA[$(item_content "$1")]]></content>
 </entry>
 END
 }
 
-main() {
-    CONFIGFILE="$PWD/$0.conf.sh"
-    case "$1" in
-        -h)
-            usage
-            exit 0
-            ;;
-        -c)
-            CONFIGFILE="$2"
-            shift 2
-            ;;
-    esac
-
-    if [ -f "CONFIGFILE" ]; then
-        . "$CONFIGFILE"
-    fi
-
-    atom_header
-    for DIR; do
-        for entry in $(recent_files "$DIR" -type f); do
-            if skip_entry "$entry"; then continue; fi
-
-            atom_entry "$entry"
-        done
-    done
-    atom_footer
+## Item-level functions
+# These can all be, and in fact should be, changed in the configuration file.
+
+skip_item() {
+        return 1
+}
+
+item_url() {
+        basename="${1#$DIRECTORY}"
+        echo "$FEED_URL/$basename"
+}
+
+item_title() {
+        basename="${1#$DIRECTORY}"
+        echo "$basename"
+}
+
+item_summary() {
+        return 1
+}
+
+item_author() {
+        echo "$FEED_AUTHOR"
+}
+
+item_content() {
+        cat "$1"
+}
+
+item_updated() {
+        # requires stat(1)
+        stat -c %y "$1"
 }
 
-if [ $DEBUG ]; then set -x; fi
+### Execution section
 
-main "$@"
+test -n $DEBUG && set -x
+test -n $SOURCE || main "$@"