[TIKA-2530] OutlookExtractor "buffer underrun" when parsing .msg with embedded .msg

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 1.16, 1.17
• Component/s: parser

Environment

Reproduced with both Tika 1.16 and Tika 1.17 on Windows but the problem is likely on all platform.

• External issue URL: https://github.com/Norconex/importer/issues/72
• External issue ID: 72
• Assignee: Tim Allison
• Reporter: Pascal Essiembre
• Created: 16/Dec/17 20:44
• Updated: 09/Mar/18 12:57

Description

When parsing certain .msg files containing certain attachments (e.g. other .msg files), I get this error:

...
Caused by: org.apache.poi.util.LittleEndian$BufferUnderrunException: buffer underrun
        at org.apache.poi.util.LittleEndian.readInt(LittleEndian.java:662)
        at org.apache.poi.hmef.CompressedRTF.decompress(CompressedRTF.java:73)
        at org.apache.poi.util.LZWDecompresser.decompress(LZWDecompresser.java:81)
        at org.apache.poi.hmef.attribute.MAPIRtfAttribute.<init>(MAPIRtfAttribute.java:42)
        at org.apache.tika.parser.microsoft.OutlookExtractor.parse(OutlookExtractor.java:270)
...

I think the issue is with MAPIRtfAttribute not liking it when receiving an empty byte array from OutlookExtractor. I was able to eliminate the error at around line 269 of OutlookExtractor with Tika 1.16 code (or around line 322 with Tika 1.17) with the following:

            //--- START FIX ---
            ByteChunk chunk = (ByteChunk) rtfChunk;
            if (chunk != null && chunk.getValue() != null 
                    && chunk.getValue().length > 0 && !doneBody) {
                //ByteChunk chunk = (ByteChunk) rtfChunk;
            //--- END FIX ---

I am not sure if that is a real fix or more should be done than just getting rid of the error to make sure all is extracted properly from all files.

I cannot share the sample file I have to test since it was given to me as sensitive content and I could not recreate a faulty msg file.

Thanks