From b1f9e9c25935b3fbc02c23331315b75d8e43cc93 Mon Sep 17 00:00:00 2001
From: Luke Granger-Brown <git@lukegb.com>
Date: Tue, 1 Dec 2020 09:49:47 +0000
Subject: [PATCH] cacert: fix blacklist

It's broken under Python 3, ironically due to the patch we're carrying.
Fix it, and add a test to check it works.

Fixes #93230.
---
 pkgs/data/misc/cacert/default.nix             | 53 ++++++++++++++++++-
 .../misc/cacert/fix-unicode-ca-names.patch    | 20 -------
 2 files changed, 52 insertions(+), 21 deletions(-)
 delete mode 100644 pkgs/data/misc/cacert/fix-unicode-ca-names.patch

diff --git a/pkgs/data/misc/cacert/default.nix b/pkgs/data/misc/cacert/default.nix
index 34e0ec132a9f..7ac6f96f66ec 100644
--- a/pkgs/data/misc/cacert/default.nix
+++ b/pkgs/data/misc/cacert/default.nix
@@ -1,6 +1,11 @@
 { stdenv, fetchurl, nss, python3
 , blacklist ? []
 , includeEmail ? false
+
+# Used for tests only
+, runCommand
+, cacert
+, openssl
 }:
 
 with stdenv.lib;
@@ -37,7 +42,6 @@ stdenv.mkDerivation {
     EOF
 
     cat ${certdata2pem} > certdata2pem.py
-    patch -p1 < ${./fix-unicode-ca-names.patch}
     ${optionalString includeEmail ''
       # Disable CAs used for mail signing
       substituteInPlace certdata2pem.py --replace \[\'CKA_TRUST_EMAIL_PROTECTION\'\] '''
@@ -66,6 +70,53 @@ stdenv.mkDerivation {
   setupHook = ./setup-hook.sh;
 
   passthru.updateScript = ./update.sh;
+  passthru.tests = {
+    # Test that building this derivation with a blacklist works, and that UTF-8 is supported.
+    blacklist-utf8 = let
+      blacklistCAToFingerprint = {
+        # "blacklist" uses the CA name from the NSS bundle, but we check for presence using the SHA256 fingerprint.
+        "CFCA EV ROOT" = "5C:C3:D7:8E:4E:1D:5E:45:54:7A:04:E6:87:3E:64:F9:0C:F9:53:6D:1C:CC:2E:F8:00:F3:55:C4:C5:FD:70:FD";
+        "NetLock Arany (Class Gold) Főtanúsítvány" = "6C:61:DA:C3:A2:DE:F0:31:50:6B:E0:36:D2:A6:FE:40:19:94:FB:D1:3D:F9:C8:D4:66:59:92:74:C4:46:EC:98";
+      };
+      mapBlacklist = f: concatStringsSep "\n" (mapAttrsToList f blacklistCAToFingerprint);
+    in runCommand "verify-the-cacert-filter-output" {
+      cacert = cacert.unbundled;
+      cacertWithExcludes = (cacert.override {
+        blacklist = builtins.attrNames blacklistCAToFingerprint;
+      }).unbundled;
+
+      nativeBuildInputs = [ openssl ];
+    } ''
+      isPresent() {
+        # isPresent <unbundled-dir> <ca name> <ca sha256 fingerprint>
+        for f in $1/etc/ssl/certs/*.crt; do
+          fingerprint="$(openssl x509 -in "$f" -noout -fingerprint -sha256 | cut -f2 -d=)"
+          if [[ "x$fingerprint" == "x$3" ]]; then
+            return 0
+          fi
+        done
+        return 1
+      }
+
+      # Ensure that each certificate is in the main "cacert".
+      ${mapBlacklist (caName: caFingerprint: ''
+        isPresent "$cacert" "${caName}" "${caFingerprint}" || ({
+          echo "CA fingerprint ${caFingerprint} (${caName}) is missing from the CA bundle. Consider picking a different CA for the blacklist test." >&2
+          exit 1
+        })
+      '')}
+
+      # Ensure that each certificate is NOT in the "cacertWithExcludes".
+      ${mapBlacklist (caName: caFingerprint: ''
+        isPresent "$cacertWithExcludes" "${caName}" "${caFingerprint}" && ({
+          echo "CA fingerprint ${caFingerprint} (${caName}) is present in the cacertWithExcludes bundle." >&2
+          exit 1
+        })
+      '')}
+
+      touch $out
+    '';
+  };
 
   meta = {
     homepage = "https://curl.haxx.se/docs/caextract.html";
diff --git a/pkgs/data/misc/cacert/fix-unicode-ca-names.patch b/pkgs/data/misc/cacert/fix-unicode-ca-names.patch
deleted file mode 100644
index 07d3629196a7..000000000000
--- a/pkgs/data/misc/cacert/fix-unicode-ca-names.patch
+++ /dev/null
@@ -1,20 +0,0 @@
---- a/certdata2pem.py	2017-08-01 23:10:00.000000000 +0300
-+++ b/certdata2pem.py	2017-08-01 23:08:21.131297636 +0300
-@@ -88,7 +88,7 @@
- \# Read blacklist.
- blacklist = []
- if os.path.exists('blacklist.txt'):
--    for line in open('blacklist.txt', 'r'):
-+    for line in io.open('blacklist.txt', 'r', encoding='utf-8'):
-         line = line.strip()
-         if line.startswith('#') or len(line) == 0:
-             continue
-@@ -101,7 +101,7 @@
-     if obj['CKA_CLASS'] != 'CKO_NSS_TRUST':
-         continue
-     if obj['CKA_LABEL'] in blacklist:
--        print("Certificate %s blacklisted, ignoring." % obj['CKA_LABEL'])
-+        print("Certificate %s blacklisted, ignoring." % unicode(obj['CKA_LABEL']).encode('utf-8'))
-     elif obj['CKA_TRUST_SERVER_AUTH'] == 'CKT_NSS_TRUSTED_DELEGATOR':
-         trust[obj['CKA_LABEL']] = True
-     elif obj['CKA_TRUST_EMAIL_PROTECTION'] == 'CKT_NSS_TRUSTED_DELEGATOR':