Github user kevdoran commented on a diff in the pull request:
https://github.com/apache/nifi/pull/2648#discussion_r182941468
--- Diff: nifi-toolkit/nifi-toolkit-encrypt-config/pom.xml ---
@@ -167,10 +167,12 @@
<groupId>org.apache.rat</groupId>
<artifactId>apache-rat-plugin</artifactId>
<configuration>
+ <consoleOutput>true</consoleOutput>
<excludes combine.children="append">
<exclude>src/test/resources/scrypt.py</exclude>
-
<exclude>src/test/resources/secure_hash.key</exclude>
-
<exclude>src/test/resources/secure_hash_128.key</exclude>
+ <!-- use wildcard for below files as tests
generate additional files during the build -->
+ <exclude>**/secure_hash.key</exclude>
+ <exclude>**/secure_hash_128.key</exclude>
--- End diff --
Ok, here's what I've got so far for the secure_hash.key file issue. It does
not look trivial to find and disable the tests that could be creating this
file, as I think it can come from both the existing (pre-NIFI-4942) and the
recently introduced test cases.
The path is hardcoded to ./secure_hash.key, but the test class does appear
to try to account for that here:
https://github.com/kevdoran/nifi/blob/master/nifi-toolkit/nifi-toolkit-encrypt-config/src/test/groov...
I'm not sure why that is not working on some Linux platforms. I'm wondering
if it has to do with the fact that it is a static field and how the classes get
loaded for these test cases, but in any case it means that any execution of the
tool could create the file that we want to avoid.
So options I see are:
- leave the tests enabled, leave the wildcard directory for the RAT
exclusions, and maybe add the file to a .gitignore listing so as to prevent it
from getting accidentally committed
- try to find every test that is creating this file and temporarily disable
it (non-trivial from what I can tell, but maybe I am missing something obvious)
- disable the entire ConfigEncryptionToolTest suite (for now)
I think all of these approaches only address the immediate issues and still
require a longer-term solution.
---
