Catching errors that aren't a proper Exception

[BlakeTerhune]

More specifically, I'm using ftplib and if the server times out, my traditional except Exception as err: is not catching it. I found one solution to use except ftplib.all_errors and it seems to work but I'd like to know if there's a more generic way to catch these "other" exceptions.

[DanPatterson_Retired]

I gave up with try except blocks for returning error messages since I found that when something fails, I just follow the error stream back to its source.  Have you really found a difference in what is meaningfully returned with and without the blocks?  The only time that I use one is if I have to delete something and if it can't I produce a new filename and move on.  So I guess my comment is, what do you want to do if something might error out... quit? or do something else?

[BlakeTerhune]

I need to catch the exception because the script is running as a scheduled task that sends an email and logs a success or failure.

[JoshuaBixby]

What behavior are you seeing that leads you to say the errors aren't being caught?  What are you doing with the exceptions after catching them?  Logging?  Exiting?  Re-trying?

[BlakeTerhune]

There's an FTP site hosted by another organization that we routinely download data from. One of the downloads is very large and takes a while to download. The FTP server has a 120 second timeout so if the download goes too slowly for whatever reason, the script is disconnected. The error message is from ftplib and says 421 Timeout (120 seconds): closing control connection.

Our script has everything in a big try/except/finally. The main logic with a log success message is in the try, there's an except Exception as err: block that logs a failure message, then there is send email in the finally. It only logs the success or failure, then sends an email with the result.

try:
    main logic here
    log success message here
except Exception as err:
    log failure message here
finally:
    send email here

I know it's skipping the except block because the failure message is not being written to the log but I get the email. I also wrote a little test script that just connects to the ftp, checks for a file, then sleeps for 150 seconds. Again, the except Exception block does not catch the ftp timeout error.

EDIT

I just looked at the script again and I have it logging with err.message instead of just err. That might be it...

[JoshuaBixby]

I just tested out ftplib and forced a connection timeout.  The try-except block caught the error.  The err.message was empty, but err.strerror and err.errno were populated.

[BlakeTerhune]

This is helping me narrow it down. I have ftp.quit() in the finally block but it seems that trying to quit an ftp connection that has timed out throws the same error but as a different error object ("error_temp"). Additionally, error_temp has no attribute "strerror". I had ftp.quit() in a try/except but I was only catching AttributeError or NameError.

I think the solution to this is two-fold:

1. Change error logging to use just err insetad of err.message.
2. Catch (and pass) all Exceptions on ftp.quit() instead of just AttributeError or NameError.

try:
    ftp.quit()
except:
    pass