Danny Yoo wrote: > Similarly, the try/except for IndexError seems too pervasive: rather than > wrap it around the whole program, we may want to limit its extent to just > around the sys.argv access. Otherwise, any other IndexError has the > potential of being misattributed. Much of the program does indexing of > some sort, so that's why this concerns me. > > Alternatively, doing the explicit length check: > > if not sys.argv[1:]: > print some kind of usage message > raise SystemExit > > might be sufficient. > > > Unless we really want to hide errors from the user, I'd avoid the > exception handlers altogether: if bad things happen, the default exception > handler gives a fairly good stack trace that aids debugging. > > But if we do want to handle the exceptions manually, we should try to make > sure that useful information is preserved in the error messages: it's a > terrible thing to get an error message that says "Something bad happened." > when we can do much better. *grin*
I agree with Danny that in this case there is no need to catch the exceptions - just let the default exception handler do its thing. If you *do* want to handle the exceptions yourself, a good principle is to put the try / except around the least amount of code possible - just the lines that may generate the expected exception. This prevents the except clause from hiding unexpected exceptions. An easy way to do this is to use try / except / else. The else clause is only executed if no exception was caught by the except clause. In your case you could write try: fname = sys.argv[1] except IndexError: # report the error as you like else: # normal processing goes here Kent _______________________________________________ Tutor maillist - Tutor@python.org http://mail.python.org/mailman/listinfo/tutor
