getarkpatch

#!/bin/bash
#
# getarkpatch
#

declare -i optcount=0
declare -i number=1	# assume single patch
declare homelink
declare homebase
declare patchfile="/dev/shm/getarkpatchfile"
declare directory="$PWD"
declare patlist
declare line
declare hrefstr='href="'
declare href
declare -a hrefary
declare -i hrefstrsiz=${#hrefstr}
declare -i chindex
declare -i count
declare -i index

declare usagestr=$(
cat <<EOF

$(basename $0) [options] link


  link - the link to the patch or to the prologue for a patch set.

  Options
  -n number    - number of patches (default is 1)
  -d directory - optional output directory, default is PWD: $PWD
  -h           - this help text
\0
EOF
)

usage() {
	echo -e "$usagestr"
	exit $1
}

# strindex string substring
#
# Return zero-based index of substring in string
# Returns -1 if substring is not found in string
#
function strindex {
	x="${1%%$2*}"
	[[ "$x" = "$1" ]] && echo -1 || echo ${#x}
}

# renapatch() - rename a patchfile taken from a web archive
#
# $1 - the path to the patchfile downloaded from the archive
# $2 - the directory into which the renamed file is to be moved
# $3 - returns the new path to the renamed file
#
renapatch(){
	local patfil="$1"
	local dir="$2"
	local found=false
	local line
	local lastto
	local lastfrom
	local temp
	local newname

	while IFS= read -r line; do

		if ! $found  \
			&& [ $(strindex "$line" "To:") -ge 0 ]; then
			lastto=$line
		fi
		if ! $found  \
			&& [ $(strindex "$line" "From:") -ge 0 ]; then
			lastfrom=$line
		fi
		if ! $found && [ $(strindex "$line" "Subject: ") -ge 0 ]; then

			found=true

			# Remove the word "Subject: " from the temp
			#
			temp=$(echo "$line" | cut -d' ' -f2-)

			# Remove any brackets, parentheses, and quotes
			#
			temp=$(echo "$temp" | sed 's/[][{}<>()|,"!@#$%^&*]//g')
			temp=$(echo "$temp" | sed "s/'//g")

			# Replace spaces, tabs, slashes, colons and semicolons
			# with hyphens
			#
			temp=$(echo "$temp" | sed "s\[/:; ]\-\g")
			temp=$(echo "$temp" | tr -d '[:space:]' )

			# Create the new name
			#
			newname=$(printf "%s/%s.patch" "$dir" "$temp")

			echo "*********************"
			echo $newname
			echo "*********************"
			echo $lastto > "$newname"
			echo $lastfrom >> "$newname"
			echo $line >> "$newname"
			continue
		fi

		$found && echo "$line" >> "$newname"

	done < $patfil

	eval $3="$newname"
}

getapatch() {

	# $directory is global defined at top of this script
	#
	local patlin="$1"
	local patfil="$2"
	local filename
	local newpat

	echo "link: $patlin"
	echo "file: $patfil"

	# Download the patch and strip the HTML comment markers to expose
	# comments for parsing.
	#
	curl -k "$patlin" | sed -r -e 's/(<!--X-)|(-->)//g' \
		-e 's/<[!a-zA-Z\/][^>]*>//g'> "$patfil"

	renapatch "$patfil" "$directory" newpat

	# Strip all the remaining HTML tags from the file
	#
	# sed -i 's/<[!a-zA-Z\/][^>]*>//g' $newpat

	# Convert the HTML entities in the file back to ASCII characters
	#
	sed -i 's/&amp;/\&/g; s/&lt;/\</g; s/&gt;/\>/g; s/&quot;/\"/g; s/#&#39;/\'"'"'/g; s/&ldquo;/\"/g; s/&#45;/-/g; s/&rdquo;/\"/g; s/&nbsp/ /g;' $newpat
}

while getopts n:d:h OPTION; do
    case "$OPTION" in

	n ) number=$OPTARG
	    optcount=$((optcount+1))
	    ;;
	d ) directory=$OPTARG
	    [ -d "$directory" ] || {
		echo "$directory is not a valid directory"
		exit 1
	    }
	    optcount=$((optcount+2))
	    ;;
	h ) optcount=$((optcount+1))
	    ;;
	* ) echo "unrecognized option -$OPTION"
	    echo -e "$usagestr"
	    exit 127
    esac
done

shift $optcount
[ $# -eq 1 ] || usage 1

# Home link must be the prologue of a patch set or a single patch.
#
# Single patch is assumed, unless user submitted -n number.
#
# If the number of patches to be extracted is greater than 1, then that
# number is used to help parse the list of links on the home link page
# for the links to the patches in the set and replies to them.
#
# The list of links must be parsed to remove the html tags, leaving only
# the relative links.
#
# To create the complete URL of each patch, the home link is stripped of
# the last field after the last '/' and the relative link of the
# corresponding patch is appended.
#
# Each patch extracted will have its html tags stripped.
#

homelink="$1"
echo "homelink: $homelink"

# If there is only one patch to process, then do that and exit with
# good status
#
[ $number -eq 1 ] && { getapatch "$homelink" "$patchfile"; exit 0; }

# If we get this far, we are processing a patch set.
# Get the list of patches and replies in the set.
#
patlist=$(curl -k "$homelink" | awk '/\[/ && /'$number'\]/')

# Extract the base link, everything but the html filename at the end
# of the link.
#
homebase=$(echo "$homelink" | rev | cut -d'/' -f2- | rev)/
echo "homebase: $homebase"

# Loop through the list of links discovered on the homepage and
# create the urls of the links to the individual patches in the
# patchset.
#
index=0
while read line; do

	# Find the href in the line. If there is no href, then
	# skip the line.
	#
	chindex=$(strindex "$line" "$hrefstr")
	[ $chindex -eq -1 ] && continue
	chindex=$((chindex + hrefstrsiz))

	# Create the URL of the patch
	#
	href="$homebase"$(echo ${line:$chindex} | cut -d'"' -f1)
	echo "href: $href"

	# Get the remainder of the line after the href
	#
	line="$(echo ${line:$chindex} | cut -d'>' -f2-)"

	# Not gonna process replies as patches
	#
	[ "${line:0:3}" == "Re:" ] && continue

	hrefary[$index]="$href"
	let index++
done <<< "$patlist"

# At this point, the hrefary array only contains urls of patches in
# this patch set.
# Loop through the hrefary array and get the patches.
#
count=$index
for ((index=0; index < count; ++index)); do
	getapatch "${hrefary[$index]}" "$patchfile"
done

patrmhead $directory

exit 0