On Thu, Dec 09, 2021 at 03:38:00PM -0500, Antoine Beaupré wrote: > Control: forwarded -1 https://gitlab.com/anarcat/feed2exec/-/issues/22 > > I'm not sure what to do with this bug report. I had the same upstream, > and it seems like the issue is more with the cachecontrol package than > with feed2exec itself: > > https://github.com/ionrock/cachecontrol/issues/262 > > Could you take a look at that package? See also the feed2exec bug > above...
Ah, I missed looking into the closed bugs.
The fact is that cachecontrol changed the interface it expects from the
cache store objects, by adding a expires argument to the `cache.set()` call.
Any use cases that provides a custom cache store implementation needs to
either adapt to that new interface, or cachecontrol needs to be made
smart enough to still support cache stores with the old implementation.
For example this patch makes feed2exec at least not crash:
----------------8<----------------8<----------------8<-----------------
diff --git a/feed2exec/model.py b/feed2exec/model.py
index 7d9f979..24c863e 100644
--- a/feed2exec/model.py
+++ b/feed2exec/model.py
@@ -344,7 +344,7 @@ class SqliteStorage(object):
% (self.value_name, self.table_name,
self.key_name), (key, )).fetchone()
return val[0] if val else None
- def set(self, key, value):
+ def set(self, key, value, expires=None):
with self.connection() as con:
con.execute("INSERT OR REPLACE INTO `%s` (`%s`, `%s`) VALUES (?,
?)"
% (self.table_name, self.key_name, self.value_name),
----------------8<----------------8<----------------8<-----------------
Of course, ideally you would also want to handle the expires by saving
it to the database, and checking for it in the implementation of get(),
(and maybe __iter__() as well).
signature.asc
Description: PGP signature
