NullPointerException in LanguageDetectTask

[wladimirleite]

This was reported to another user who was processing a folder with a lot of 7Z, ZIP, TAR and ISO files, using 4.1.5. Processing only the 7z that caused the error I was able to reproduce it (using master branch).

2023-10-25 16:35:25     [ERROR] [app.processing.Main]                   Processing Error:
java.lang.Exception: Worker-11 Error while processing f2653796544.7z>>XXX/YYY.jpg (147456bytes)
        at iped.engine.core.Worker.process(Worker.java:186) ~[iped-engine-4.2-snapshot.jar:?]
        at iped.engine.core.Worker.run(Worker.java:265) ~[iped-engine-4.2-snapshot.jar:?]
Caused by: java.lang.NullPointerException
        at iped.engine.task.LanguageDetectTask.process(LanguageDetectTask.java:77) ~[iped-engine-4.2-snapshot.jar:?]

Line 77 of LanguageDetectTask is if (evidence.getMediaType().equals(MediaType.OCTET_STREAM)), and evidence.getMediaType() is null, which caused the exception.

@lfcnassif, this seems similar to the issue I reported last night. Obviously we could check if evidence.getMediaType() is not null at this point of the code, or invert the equals order, but I see that there are many other places in the code with a similar structure evidence.getMediaType().someMethod(). So, was anything changed recently (4.1.5, maybe before) that may cause an item with null mediaType? Or having null mediaType is expected?

Would it be a good idea to change Item.getMediaType() to return some default mediaType when mediaType is null? Or could it cause bad side effects?

[lfcnassif]

I think it is not expected. After SignatureTask, all items are expected to have a mediaType.

[lfcnassif]

So, was anything changed recently (4.1.5, maybe before) that may cause an item with null mediaType?

AFAIK nothing related to this was changed recently.

Would it be a good idea to change Item.getMediaType() to return some default mediaType when mediaType is null? Or could it cause bad side effects?

I think it shouldn't return a default value, explicitly aborting bugs like this can help us to fix them in the right place.

[lfcnassif]

At SignatureTask.java:122 the mediaType should be set for all items...

[wladimirleite]

Thanks @lfcnassif! I will take a closer look and see if I can identify the root cause.

[wladimirleite]

At SignatureTask.java:122 the mediaType should be set for all items...

The 7z was partially overwritten, so although the file extension is ".jpg", signature detected is "application/xhtml+xml", and correctly set in the line you mentioned.

Later, in the following part of the code of ParsingTask, parsedMediaType is just "application/", making MediaType.parse(parsedMediaType) to return null.

        String prevMediaType = evidence.getMediaType().toString();
        String parsedMediaType = metadata.get(StandardParser.INDEXER_CONTENT_TYPE);
        if (!prevMediaType.equals(parsedMediaType)) {
            evidence.setMediaType(MediaType.parse(parsedMediaType));
        }

I am going to submit a PR with a simple check to avoid such situation.