warn instead of crashing on invalid dates
We used to completely crash when a feed had invalid or missing dates. After reviewing the standards, it turns out this is not quite valid behavior: RSS 0.90 and 0.91, for example, do not have dates at...
View Articleexpand and cleanup date validation routines
We used to have a long list of fields. Now we just iterate over the item, then the feed, and look for the fields we want. It's cleaner visually and might even make some feeds validate, as we now look...
View Articleuse dateparser module if available
This behaves better than the feedparser date parser in most scenarios. Still needs unit tests and dependency checks. Closes: #6
View Articlehandle broken pipe correctly from plugins
Before this, doing "feed2exec parse foo | head" would yield an error message for *every* feed item. This silences the warnings completely.
View Articleadd JSON output plugin
This plugin is very simple, if not trivial: it simply dumps all the feed items in a JSON stream. This can be parsed by `jq` on the commandline to diagnose feed problems, do scripting or whatever. This...
View Articleswitch to dateparser for PyPI and tests as well
Because the Debian package recommends dateparser, I had different results running tests natively on Debian and within tox. This harmonizes things and makes use of dateparser everywhere, warts and...
View Articlereduce noise level of 'missing time' problems
It seems like previous versions of feedparser would never trigger that problem, and would fill in the date instead. Now I have feeds that have this on *every* item and it generates a lot of noise in...
View Articlesplit large feeds.py into model.py and controller.py
This rearchitecture seems to make sense to me. I would like to keep those files smaller and that naming will force me to follow that model/controller distinction more clearly. Already, the parse/fetch...
View Articlerename FeedCacheStorage to FeedItemCacheStorage
This is, effectively, a per-item cache, not a full feed cache. We want to implement the latter as part of #10 so it makes sense to rename this first. This is an API breaking change.
View Article