Catch NuValidator's StackOverflowError

[tofi86]

For pagina EPUB-Checker I updated the Windows wrapper to add the JVM argument -Xss1024k as suggested in https://github.com/w3c/epubcheck/wiki/Running#javalangstackoverflowerror but it seems that on certain systems this does not help and the SOE is still thrown. This leads to an unwanted situation: EPUBCheck is finished "successfully without errors or warnings" – but only because no checks were performed at all because of the SOE.

This can be reproduced on Mac and Windows by setting -Xss144k for example and checking a medium or large sized book.

Is there any chance to throw a FATAL exception if the NuValidator crashes?

And as a workaround I'm looking for a way to catch the previously uncaught SO-Errors when running EpubCheck.doValidate()? Does anyone know of a solution for that? Hint: Wrapping EpubCheck.doValidate() in a try/catch does not work ;-)

Thanks, Tobias

[rdeltour]

By essence a StackOverflowError indicates a serious issue that is likely unrecoverable, and I don't think it would make much sense (nor would be much helpful) to catch that.

I'm closing the issue as wontfix, but feel free to keep commenting here if anyone can suggest a better handling!