bin/comics/xkcd: get comic for specific date via the archive page

It is possible to get a xkcd comic for a specific date via the archive
page. This eleminates the need of the state directory and xkcd was the
only script that used it.

2 files changed, 37 insertions, 41 deletions

diff --git a/bin/comics/xkcd b/bin/comics/xkcd
index 7079294..78c81b3 100755
--- a/bin/comics/xkcd
+++ b/bin/comics/xkcd
@@ -15,38 +15,37 @@ cd "${tmp}"
 browser="Mozilla/4.76 [de] (X11; U; Linux 2.2.18 i586)"
 newn="${day}.png"
 
-wget ${wget_args} -U "$browser" http://www.xkcd.com -O index.html
-
-url=$(grep -A 1 "Image URL" index.html | sed 'N;s/.*\(http.*png\)<\?.*/\1/')
-imgname=$(echo "$url" | tr '/' '\n' | tail -1)
-
-# get additional information
-img_tag=$(grep -A2 "src=\"${url}" index.html | sed 'N;N;s/.*\(<img [^>]\+>\).*/\1/')
-title=$(sed 's/.*title="\([^"]*\)".*/\1/'<<< $img_tag)
-alt=$(sed 's/.*alt="\([^"]*\)".*/\1/'<<< $img_tag)
-
-if [ ! -s "${image_dir}/$newn" ]; then
-    # no image availalable for current date
-
-    if ! grep -q "${url}" "${state_file}" >/dev/null; then
-        # new image
-        echo $url > "${state_file}"
-
-        wget ${wget_args} -U "$browser" -O "${tmp}/${newn}" \
-            --header="Referer: http://www.xkcd.com/" "$url"
-
-        if [ -s "${tmp}/${newn}" ]; then
-            # save
-            mv "${tmp}/${newn}" "${image_dir}/$newn"
-            echo "$title" > "${image_dir}/${day}.title"
-            echo "$alt" > "${image_dir}/${day}.alt"
-
-            # update symlinks
-            rm -f "${comic_dir}/latest.png" "${comic_dir}/latest.title" "${comic_dir}/latest.alt"
-            ln -s "${image_offset}/$newn" "${comic_dir}/latest.png"
-            ln -s "${image_offset}/${day}.title" "${comic_dir}/latest.title"
-            ln -s "${image_offset}/${day}.alt" "${comic_dir}/latest.alt"
-        fi
+# get url
+echo $(date -d"$daysago days ago" +%Y-%-m-%-d)
+url="$(wget ${wget_args} -U "$browser" http://xkcd.com/archive/ -O - | \
+    grep -o "href=\".*\" .*\"$(date -d"$daysago days ago" +%Y-%-m-%-d)\"" | \
+    sed 's/^href="\(.*\)" title.*$/\1/')"
+
+if [ -n "$url" ]; then
+    # fetch comic
+    wget ${wget_args} -U "$browser" "http://www.xkcd.com${url}" -O index.html
+
+    url=$(grep -A 1 "Image URL" index.html | sed 'N;s/.*\(http.*png\)<\?.*/\1/')
+    imgname=$(echo "$url" | tr '/' '\n' | tail -1)
+
+    # get additional information
+    img_tag=$(grep -A2 "src=\"${url}" index.html | sed 'N;N;s/.*\(<img [^>]\+>\).*/\1/')
+    alt=$(sed 's/.*title="\([^"]*\)".*/\1/'<<< $img_tag)
+    title=$(sed 's/.*alt="\([^"]*\)".*/\1/'<<< $img_tag)
+
+    wget ${wget_args} -U "$browser" -O "${tmp}/${newn}" --header="Referer: http://www.xkcd.com/" "$url"
+
+    if [ -s "${tmp}/${newn}" ]; then
+        # save
+        mv "${tmp}/${newn}" "${image_dir}/$newn"
+        echo "$title" > "${image_dir}/${day}.title"
+        echo "$alt" > "${image_dir}/${day}.alt"
+
+        # update symlinks
+        rm -f "${comic_dir}/latest.png" "${comic_dir}/latest.title" "${comic_dir}/latest.alt"
+        ln -s "${image_offset}/$newn" "${comic_dir}/latest.png"
+        ln -s "${image_offset}/${day}.title" "${comic_dir}/latest.title"
+        ln -s "${image_offset}/${day}.alt" "${comic_dir}/latest.alt"
     fi
 fi
 
diff --git a/etc/settings.sh b/etc/settings.sh
index d07293f..00947ac 100644
--- a/etc/settings.sh
+++ b/etc/settings.sh
@@ -1,7 +1,11 @@
 # this file is sourced by the get scripts
 
-if [ "$1" = "" ]
-then
+if [ -z "$top" ]; then
+    echo "\$top needs to be set before sourcing settings.sh" 1>&2
+    exit 1
+fi
+
+if [ -z "$1" ]; then
     daysago=0
 else
     daysago=$1
@@ -12,9 +16,6 @@ year=$(date -d"$daysago days ago" +%Y)
 month=$(date -d"$daysago days ago" +%Y-%m)
 day=$(date -d"$daysago days ago" +%d)
 
-top="$(pwd)/$(dirname $0)/../../"
-state_dir=${top}/var/state/
-state_file=${state_dir}/${comic}.state
 comic_dir=${top}/htdocs/imgs/${comic}
 image_offset=${month}/
 image_dir=${comic_dir}/${image_offset}
@@ -23,8 +24,4 @@ if [ ! -d "${image_dir}" ]; then
     mkdir -p "${image_dir}"
 fi
 
-if [ ! -d "${state_dir}" ]; then
-    mkdir -p "${state_dir}"
-fi
-
 wget_args="--timeout 100 --wait=1000 --random-wait --no-cache"