Vinay Sajip <vinay_sa...@yahoo.co.uk> added the comment: On Ubuntu, without the locks around the os.system calls, I was not getting any errors other than file-not-found errors on blah.txt - which you would expect if another thread had just deleted the file. You didn't post the errors which your test script was generating - and I assumed you may have been referring to the file-not-founds on blah.txt.
You definitely need the join calls, as without them atexit processing in the main thread will close the handlers when the threads are still running. This will lead to errors like the ones you saw (I/O operation on closed file), but this is not a bug - you definitely need to join on all created threads before exiting the main thread - whether test or production. Please post the exact errors you are getting, after removing the locks around the os.system calls. Delete all logthred.log files before the test run, and please post the console output as Frans did. As I said in my earlier comment - ensure that anti-virus, Windows indexing and any other software which may open files at non-deterministic times is disabled. If you are seeing a WindowsError with an error code of 32 (this might get lost in all the other output, but it's the first error I found - the other messages look like they are a consequence of that initial error). ---------- status: pending -> open _______________________________________ Python tracker <rep...@bugs.python.org> <http://bugs.python.org/issue4749> _______________________________________ _______________________________________________ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
