#!/bin/sh -e

[ -n "$1" ] \
|| { echo "usage: $(basename $0) <package> [timeout]" && exit 0; }

[ "$1" != "-c" ] || { shift; cached="yes"; }
[ "$1" != "-r" ] || { shift; reparse="yes"; }

name="$1"
timeout=${2:-5}

cleanup() { [ -f "$temp" ] && rm "$temp"; }
trap cleanup EXIT

temp="$(mktemp)"

# main
#url="https://git.alpinelinux.org/aports/tree/$name?h=master"
#pre="tree listing"
#pst="<\/table"
#rex="href='.+\/([^\/]+.patch)'"
#plain="https://git.alpinelinux.org/aports/plain"

# github mirror
url="https://github.com/alpinelinux/aports/tree/master/$name"
pre="repository-content"
pst="footer container"
rex='title="([^\/]+.patch)"'
plain="https://raw.githubusercontent.com/alpinelinux/aports/master"

# extract the text between the pre and post regexes on the target webpage.
{ curl --compressed -m $timeout -L -s $url \
|| { echo "url failed" >&2 && continue; }; } \
| awk "sub(/^.*$pre/,\"\"){echo=1} /$pst/{echo=0} echo" \
> $temp

perl -n -e "/$rex/ && print \"\$1\\n\"" $temp | while read path; do
    curl -LO "$plain/$name/$path"
done