Success vs tempfail vs permfail

[RussNelson]

It can be useful to keep track of whether an upload has failed temporarily (e.g. network error) or permanently (e.g. configuration error). This would be done in Uploader.upload() where the uploader.upload_file() would return a ternary value (success, tempfail, permfail) rather than the current binary value ( complete==True is success, complete==False is tempfail, and a traceback is permfail). Further in that routine, success would result in "%s uploaded successfully", tempfail would result in adding it to the failed_uploads, and permfail wouldn't.

The advantage? A permfail wouldn't retry, wouldn't issue a trackback, _retry_upload() could be more aggressive about attempting retries.

Not suggesting that anybody else write code, just wondering if you want to move in this direction. I hate writing code that maintainers never wanted in the first place.

[RussNelson]

The specific reason I suggest this is that if you configure remote_dir with an unwritable path, you get a traceback. A traceback looks like an error in the code rather than an error in the configuration. Plus, it's a lousy way to signal an error because you don't get a chance to suggest how the user might fix the code, e.g. "Unable to write to the remote_dir you configured in config.pyu.