One of the recent patches re-introduced the usage of dmemcache_key() in wildcards(), this resulted in different wildcard keys for getting/setting cache entries.

The result is that $cache->flushes is always 0 and is marked as invalid on the next get, resulting in yet another cache set...

CommentFileSizeAuthor
wildcards_key.patch962 bytesberdir

Comments

catch’s picture

catch
he/him
Primary language English
commented
Status: Needs review » Fixed

Committed, thanks!

catch’s picture

catch
he/him
Primary language English
commented
Title: Wildcard clears are broken again... » Tests for wildcard clears are broken again...
Category: bug » task
Priority: Normal » Major
Status: Fixed » Active

Berdir rightly pointed out in irc we should add tests for this.

berdir’s picture

berdir
Primary language German
Location Switzerland
commented

There are actually test fails related to cache clearing right now, even with this bug fixed. I think something with the wildcard optimization is not correct, but I'm still investigating..

berdir’s picture

berdir
Primary language German
Location Switzerland
commented
Status: Active » Fixed

The tests fail with #1131370: Segment wildcard counts by $string[0] applied and without my fix here. (not the new tests, the existing ones)

Status: Fixed » Closed (fixed)

Automatically closed -- issue fixed for 2 weeks with no activity.