From f91218360b8bce56dd96fac888c83a6a1c6bdc32 Mon Sep 17 00:00:00 2001
From: sharonenoch <52180175+sharonenoch@users.noreply.github.com>
Date: Mon, 7 Mar 2022 18:21:32 -0500
Subject: Extracting the Azure Origin url from the download link (#1480)

* Extracting the Azure Origin url from the download link

* Response code check to address double quotes

Co-authored-by: Sharon Enoch <sharone@amzetta.com>
---
 utils/azure_ip_addresses_download.sh | 25 ++++++++++++++++++++-----
 1 file changed, 20 insertions(+), 5 deletions(-)

(limited to 'utils/azure_ip_addresses_download.sh')

diff --git a/utils/azure_ip_addresses_download.sh b/utils/azure_ip_addresses_download.sh
index 7f0bd81dd..8105fb87f 100755
--- a/utils/azure_ip_addresses_download.sh
+++ b/utils/azure_ip_addresses_download.sh
@@ -3,26 +3,41 @@
 cd "$(dirname "${0}")"
 
 DEST=../src/lib/ndpi_azure_match.c.inc
+LINK_TMP=/tmp/azure_link.txt
 TMP=/tmp/azure.json
 LIST=/tmp/azure.list
 # https://www.microsoft.com/en-us/download/confirmation.aspx?id=56519
-ORIGIN="https://download.microsoft.com/download/7/1/D/71D86715-5596-4529-9B13-DA13A5DE5B63/ServiceTags_Public_20220124.json"
+# Azure links have the format https://download.microsoft.com/download/7/1/D/71D86715-5596-4529-9B13-DA13A5DE5B63/ServiceTags_Public_<date>.json
+LINK_ORIGIN="https://www.microsoft.com/en-us/download/confirmation.aspx?id=56519"
 
+echo "(1) Downloading file... ${LINK_ORIGIN}"
+http_response=$(curl -s -o ${LINK_TMP} -w "%{http_code}" ${LINK_ORIGIN})
+if [ "${http_response}" != "200" ]; then
+    echo "Error $http_response: you probably need to update the link origin url!"
+    return
+fi
+
+ORIGIN="$(grep -E 'ServiceTags_Public_[[:digit:]]+.json' ${LINK_TMP} | grep -o -E 'href=\"[^"]+' | sed 's/href="//' | uniq)"
+rm -f ${LINK_TMP}
+if [ -z "${ORIGIN}" ]; then
+    echo "Error ${LINK_ORIGIN} does not contain the url format!"
+    return
+fi
 
-echo "(1) Downloading file..."
+echo "(2) Downloading file... ${ORIGIN}"
 http_response=$(curl -s -o $TMP -w "%{http_code}" ${ORIGIN})
-if [ $http_response != "200" ]; then
+if [ "${http_response}" != "200" ]; then
     echo "Error $http_response: you probably need to update the list url!"
     return
 fi
 
-echo "(2) Processing IP addresses..."
+echo "(3) Processing IP addresses..."
 # Note: the last "grep -v :" is used to skip IPv6 addresses
 tr -d '\r' < $TMP | grep / | tr -d '"' | tr -d " " | tr -d "," | grep -v : > $LIST
 ./ipaddr2list.py $LIST NDPI_PROTOCOL_MICROSOFT_AZURE > $DEST
 rm -f $TMP $LIST
 
-echo "(3) Microsoft Azure IPs are available in $DEST"
+echo "(4) Microsoft Azure IPs are available in $DEST"
 
 
 
-- 
cgit v1.2.3