GNU: add a script to list the failing tests

and sort them by size: (the smaller, the easier to fix in general)

util/remaining-gnu-error.py

@@ -0,0 +1,49 @@
+#!/usr/bin/env python3
+# This script lists the GNU failing tests by size
+# Just like with util/run-gnu-test.sh, we expect the gnu sources
+# to be in ../
+import urllib.request
+
+import urllib
+import os
+import glob
+import json
+
+base = "../gnu/tests/"
+urllib.request.urlretrieve(
+    "https://raw.githubusercontent.com/uutils/coreutils-tracking/main/gnu-full-result.json",
+    "result.json",
+)
+
+tests = glob.glob(base + "/*/*.sh")
+tests_pl = glob.glob(base + "/*/*.pl")
+tests_xpl = glob.glob(base + "/*/*.xpl")
+tests = tests + tests_pl + tests_xpl
+
+# sort by size
+list_of_files = sorted(tests, key=lambda x: os.stat(x).st_size)
+
+with open("result.json", "r") as json_file:
+    data = json.load(json_file)
+
+for d in data:
+    for e in data[d]:
+        # Not all the tests are .sh files, rename them if not.
+        script = e.replace(".log", ".sh")
+        a = "%s%s/%s" % (base, d, script)
+        if not os.path.exists(a):
+            a = a.replace(".sh", ".pl")
+            if not os.path.exists(a):
+                a = a.replace(".pl", ".xpl")
+
+        # the tests pass, we don't care anymore
+        if data[d][e] == "PASS":
+            list_of_files.remove(a)
+
+# Remove the factor tests and reverse the list (bigger first)
+tests = list(filter(lambda k: "factor" not in k, list_of_files))
+
+for f in reversed(tests):
+    print("%s: %s" % (f, os.stat(f).st_size))
+print("")
+print("%s tests remaining" % len(tests))