Closed thalesmg closed 2 years ago
Anything I can do to help with this?
Pinging @sigrlami as well since you're the maintainer listed on Hackage.
@sigrlami We should probably merge this. Are you still interested in being the maintainer? I can try to find someone else if not, but I've kinda drifted away from the Haskell community...
@IreneKnapp @chreekat sorry, I will update it this week, too much work on my job.
no problem, and totally understood. paying work does have to come first.
@sigrlami Sorry to bother you, but it would be great if this was merged 😅
Since it has been a while, new versions of Sqlite were released.
I made another PR with an update for 3.37.2, if anyone is interested: #91
I've separately updated to 3.38.5. Regarding the failing test, the approach I chose was to remove it as in #93.
Since the current version of the embedded sqlite (3.28.0 from #80 ), some new features have been released. One of them is the experimental trigram tokenizer for fts5.
I had to change the test
testGetAutoCommit
for some reason that is not clear to me. I could not find in the changelog the reason for this change in behavior. At the time of writing, theget_autocommit
function documentation seems to mention thatSQLITE_FULL
is one of the errors that should induce an automatic ROLLBACK.Yet, a manual test of the broken test case seems to show that this is not the case anymore:
That behavior was indeed present in older versions, as the following test in the
haskell:8.10.4
docker image reveals:With that, I changed the test case to reflect this new behavior.