opengrok XMLAnalyzer slow with attributes with huge value

Describe the bug OpenGrok 1.7.20, openjdk version "11.0.11" 2021-04-20

I'm trying to index sources that include XML files. Some XML files have elements that have attributes with a huge value (MBs of Base64 encoded data). XMLAnalyzer takes forever on these files. For example, a 1.7MB XML file took 13 minutes to process.

For now, I'll filter out (--ignore) these files (they have a unique file extension), but would be nice if XMLAnalyzer could be updated to process such files in reasonable time.

Oct 07 '21 14:10 boekhold

Possibly similar to #907.

Oct 08 '21 07:10 vladak

@vladak just to confirm: are you suspecting that the XMLAnalyzer is trying to parse the attribute value as a path, and that due to the huge length/size of my attribute value this is taking forever?

Oct 10 '21 05:10 boekhold

I rather suspect there is something sub-optimal in the lexical parsing rules that gets triggered by huge values.

Oct 11 '21 08:10 vladak