ANY23-445 Review spotbugs issues

[lewismc]

First pass at addressing https://issues.apache.org/jira/browse/ANY23-445.
Still loads to do.

[lewismc]

Current branch results in following errors

[INFO] --- spotbugs-maven-plugin:3.1.12.2:check (default-cli) @ apache-any23-api ---
[INFO] BugInstance size is 2
[INFO] Error size is 0
[INFO] Total bugs: 2
[ERROR] org.apache.any23.mime.MIMEType defines compareTo(MIMEType) and uses Object.equals() [org.apache.any23.mime.MIMEType] At MIMEType.java:[line 134] EQ_COMPARETO_USE_OBJECT_EQUALS
[ERROR] Possible null pointer dereference in org.apache.any23.plugin.Any23PluginManager.loadJARDir(File) due to return value of called method [org.apache.any23.plugin.Any23PluginManager, org.apache.any23.plugin.Any23PluginManager] Dereferenced at Any23PluginManager.java:[line 201]Known null at Any23PluginManager.java:[line 201] NP_NULL_ON_SOME_PATH_FROM_RETURN_VALUE

@HansBrende if you get a chance can you review. You can run spotbugs analysis as follows

mvn spotbugs:check

You can also get detailed information by running

mvn clean install -DskipTests && mvn spotbugs:gui

You can then load the file in ./api/target/spotbugsXml.xml

[HansBrende]

There's actually a bug here, so adding @SuppressWarnings kinda defeats the purpose of the spotbugs plugin. getSupportedMIMETypes() returns a list of MIMEType, so checking if it contains a String will always return false!

[lewismc]

I've removed both the SupressWarnings javac annotation and thought about this one again.

I've decided to introduce <omitVisitors>FindUnrelatedTypesInGenericContainer</omitVisitors> which will omit spotbugs from identifying this issue as a bug. My justification is that, in this case, because the ExtractorFactory#getSupportedMIMETypes() supports wildcards, and as far as I can see there is currently no way to determine the full Collection (e.g. */*) of supported MimeType's then we cannot enforce another check whether an ExtractorGroup supports all MimeType's.

What do you think about this assessment?

[HansBrende]

Sorry, I wasn't clear. The code inside this method has a bug, as correctly identified by the spotbugs plugin. The code is testing if a Collection<MIMEType> contains a String. This will always return false, as a String cannot be an instance of MIMEType.

Therefore, the method needs to be fixed, and the spotbugs plugin is doing its job correctly: no suppression of warnings is needed here.

[HansBrende]

typo

[lewismc]

@HansBrende do you have any issue with us pushing this off to the 2.5 development drive? This has blocked the 2.4 release for months and I don't really have the time to invest in addressing it right now. Thanks

[HansBrende]

@lewismc I can potentially do a quick fix-up on this PR this weekend when I get a chance, and then we should be good to go.