build: allow icu download to use other hashes besides md5

- ICU uses sha512 instead of md5 in some recent releases
- Use hashlib.algorithms_guaranteed to choose the following algorithms:
    sha1 sha224 sha384 sha256 sha512 md5
- No preference as to the priority of the algorithms
- This commit does not change the hash used for ICU.

Fixes: #27369
Reviewed-By: Refael Ackermann <refack@gmail.com>
Reviewed-by: Ujjwal Sharma <usharma1998@gmail.com>
Reviewed-by: Richard Lau <riclau@uk.ibm.com>
PR-URL: #27370

Author: srl295

configure.py

@@ -1261,7 +1261,8 @@ def glob_to_var(dir_base, dir_sub, patch_dir):
 
 def configure_intl(o):
   def icu_download(path):
-    with open('tools/icu/current_ver.dep') as f:
+    depFile = 'tools/icu/current_ver.dep';
+    with open(depFile) as f:
       icus = json.load(f)
     # download ICU, if needed
     if not os.access(options.download_path, os.W_OK):
@@ -1270,7 +1271,12 @@ def icu_download(path):
     attemptdownload = nodedownload.candownload(auto_downloads, "icu")
     for icu in icus:
       url = icu['url']
-      md5 = icu['md5']
+      (expectHash, hashAlgo, allAlgos) = nodedownload.findHash(icu)
+      if not expectHash:
+        error('''Could not find a hash to verify ICU download.
+          %s may be incorrect.
+          For the entry %s,
+          Expected one of these keys: %s''' % (depFile, url, ' '.join(allAlgos)))
       local = url.split('/')[-1]
       targetfile = os.path.join(options.download_path, local)
       if not os.path.isfile(targetfile):
@@ -1279,13 +1285,13 @@ def icu_download(path):
       else:
         print('Re-using existing %s' % targetfile)
       if os.path.isfile(targetfile):
-        print('Checking file integrity with MD5:\r')
-        gotmd5 = nodedownload.md5sum(targetfile)
-        print('MD5:      %s  %s' % (gotmd5, targetfile))
-        if (md5 == gotmd5):
+        print('Checking file integrity with %s:\r' % hashAlgo)
+        gotHash = nodedownload.checkHash(targetfile, hashAlgo)
+        print('%s:      %s  %s' % (hashAlgo, gotHash, targetfile))
+        if (expectHash == gotHash):
           return targetfile
         else:
-          warn('Expected: %s      *MISMATCH*' % md5)
+          warn('Expected: %s      *MISMATCH*' % expectHash)
           warn('\n ** Corrupted ZIP? Delete %s to retry download.\n' % targetfile)
     return None
   icu_config = {

tools/configure.d/nodedownload.py

@@ -46,9 +46,19 @@ def retrievefile(url, targetfile):
         print(' ** Error occurred while downloading\n <%s>' % url)
         raise
 
-def md5sum(targetfile):
-    """md5sum a file. Return the hex digest."""
-    digest = hashlib.md5()
+def findHash(dict):
+    """Find an available hash type."""
+    # choose from one of these
+    availAlgos = hashlib.algorithms_guaranteed
+    for hashAlgo in availAlgos:
+      if hashAlgo in dict:
+        return (dict[hashAlgo], hashAlgo, availAlgos)
+    # error
+    return (None, None, availAlgos)
+
+def checkHash(targetfile, hashAlgo):
+    """Check a file using hashAlgo. Return the hex digest."""
+    digest = hashlib.new(hashAlgo)
     with open(targetfile, 'rb') as f:
       chunk = f.read(1024)
       while chunk !=  "":