Jakub Wilk <[email protected]> writes: > rawdog.py:36 - catch ImportError; > rawdog.py:41 - ditto;
You'll get ImportError if the module's missing. If the module's present but raises an exception during initialisation then rawdog should still fall back to the alternative, rather than crashing. > rawdog.py:297 - catch UnicodeError; Nope; it's got to deal with the wacky behaviour of any subclass of str, and a mis-decoded string is preferable to a crash (as it is elsewhere in rawdog and feedparser). That bit of code existed originally because BeautifulSoup did highly entertaining things with string subclassing, and rawdog needs to be robust in flattening whatever feedparser returns back to standard Python types so it can reliably pickle them. Strings are less of a concern now that feedparser's use of BeautifulSoup is explicitly disabled, although the dict flattening is still important for feedparser's results. > rawdog.py:1217 - catch AttributeError. Hmm. The behaviour's the same either way, but it's not consistent: I'm catching AttributeError in the other places where I'm looking for attributes added during the 2.x series, so I'll change that in the next upstream release. Thanks, -- Adam Sampson <[email protected]> <http://offog.org/> -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected] Archive: http://lists.debian.org/[email protected]
