> My Mac build now works with the files that I have thrown at it, unless > I enable openmp. In that case it usually fails in lines2fsts, but at > what amounts to a random place, as I get differing numbers of lines > processed before it fails.
Yes, because lines are processed in parallel, and that's not deterministic. > I know that you did some magic to eliminate some of the "terminate > called after throwing an instance of 'char*'" Is it possible to do it > a bit more, so that exceptions thrown during processing are at least > logged before termination? We can try to add more exception handlers (or you can try to just add one yourself in ocr-commands in the OMP for loop in main_lines2fsts). Also, even with OpenMP, you can compile unoptimized and with debugging symbols and catch the place where the exception gets thrown; that way, you might be able to figure out where it's coming from. > Up to this point I have been diagnosing problems by running the non- > openmp version and seeing where it failed. But now that the non-openmp > version always works, and the openmp version usually fails, this > tactic is useless. If the non-openmp version doesn't throw the exception and the openmp version does, there's something weird going on. We aren't seeing this on Ubuntu, so it might even be a bug in the Mac OpenMP implementation. Tom --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "ocropus" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/ocropus?hl=en -~----------~----~----~----~------~----~------~--~---
