p.p.s. I have only wrapped my lock around file-descriptor creations. Should I wrap it around closings too? Or the whole open -> close transaction? It sounds like error-prone work to do the latter. What am I missing?
What should I be reading to get a better clue? I'll do some googling. Thank you. On Sat, 15 Jan 2005, Marilyn Davis wrote: > Dearest Tutors, > > Bah! It's not over yet. I don't know why, but again my file > descriptors are being trampled upon now and then. > > This time I can see in my log that I'm not trampling on them myself, > like I used to do, unless I'm making calls to the system that I'm not > aware of. > > And, first I get this worrisome output: > > close failed: [Errno 9] Bad file descriptor > close failed: [Errno 9] Bad file descriptor > close failed: [Errno 9] Bad file descriptor > > coming from nowhere. No traceback, nothing. > > I traced this down once to discover that it came out of python *after* > a raise, after a failed call to open a connection to mysql. In that > particular raise clause, I called my facility to pipe mail to exim, > opening 3 descriptors. I guessed then that when closing down, it > closes down the pipe descriptors that I had already closed? > > So, if it does this, if it closes down descriptors that I've already > closed, it's possible that it's closing down descriptors that have > also already been reopened because I'm running this thing like > lightning? > > And the other thought is that there is some other call I make that > uses a file descriptor that I'm not thinking of. os.listdir? > os.anything? > > Any new thoughts? > > Marilyn > > > > > _______________________________________________ > Tutor maillist - Tutor@python.org > http://mail.python.org/mailman/listinfo/tutor > -- _______________________________________________ Tutor maillist - Tutor@python.org http://mail.python.org/mailman/listinfo/tutor