I have noticed some spurious updates due to the hash checking (#179):
For feeds like https://github.com/$USER.private.atom, entries related to a repo have text like Python (star)123 Updated Mar 17; the date keeps changing.
Feeds like http://www.scpwiki.com/feed/pages/created_by/$USER have links like http://www.wikidot.com/avatar.php?userid=$ID&size=small×tamp=161600557 in their content; the timestamp keeps changing, I assume it's for cache busting.
I think that's acceptable, at least for now. Possible alternatives:
allow users to enable/disable hash checking per-feed (bad, because they have to do stuff)
use a heuristic to detect cases like the one above
infer stuff from the content (bad, hard to do right)
if all the entries in the feed are either new(er) or with a changed hash, ignore the hash (might have false positives / negatives)
limit the amount of (consecutive?) updates an entry can receive due to hash changes (seems most reasonable)
lemon24
commented
3 years ago
I have noticed some spurious updates due to the hash checking (#179):
https://github.com/$USER.private.atom, entries related to a repo have text likePython (star)123 Updated Mar 17; the date keeps changing.http://www.scpwiki.com/feed/pages/created_by/$USERhave links likehttp://www.wikidot.com/avatar.php?userid=$ID&size=small×tamp=161600557in their content; the timestamp keeps changing, I assume it's for cache busting.I think that's acceptable, at least for now. Possible alternatives:
Originally posted by @lemon24 in https://github.com/lemon24/reader/issues/179#issuecomment-801327048