Make sqlite3 statement cache optional

[palaviv]

BPO	30271
Nosy	@rhettinger, @berkerpeksag, @palaviv
PRs	bpo-30271: Make sqlite3 statement cache optional  #1463

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2019-05-09.09:53:10.677>
created_at = <Date 2017-05-04.19:03:50.059>
labels = ['extension-modules', 'type-feature', '3.8']
title = 'Make sqlite3 statement cache optional'
updated_at = <Date 2019-05-09.09:55:24.913>
user = 'https://github.com/palaviv'

bugs.python.org fields:

activity = <Date 2019-05-09.09:55:24.913>
actor = 'berker.peksag'
assignee = 'none'
closed = True
closed_date = <Date 2019-05-09.09:53:10.677>
closer = 'palaviv'
components = ['Extension Modules']
creation = <Date 2017-05-04.19:03:50.059>
creator = 'palaviv'
dependencies = []
files = []
hgrepos = []
issue_num = 30271
keywords = []
message_count = 6.0
messages = ['293006', '322399', '323412', '341927', '341964', '341976']
nosy_count = 4.0
nosy_names = ['rhettinger', 'ghaering', 'berker.peksag', 'palaviv']
pr_nums = ['1463']
priority = 'normal'
resolution = 'rejected'
stage = 'resolved'
status = 'closed'
superseder = None
type = 'enhancement'
url = 'https://bugs.python.org/issue30271'
versions = ['Python 3.8']

[palaviv]

Currently the minimum size of the statement cache is 10. I suggest that it will be any value above 1 or no cache at all if the size is set to 0.

[berkerpeksag]

Thanks for the report. What is your use case for making the statement caching optional? Did you notice any performance improvement when you use it with newer SQLite versions?

[palaviv]

I don't have any specific use case for making the statement cache optional. I expected that by changing the cache size to 0 there will be no statement cache. I think that this is a common assumption as can be seen in ghaering/pysqlite#126 (comment).

rogerbinns did give a use case where we would like to disable the statement cache in ghaering/pysqlite#126 (comment). I think that statement cache should be disable implicitly in that case as you suggest. The code change will be very similar and I do believe we should allow the user to disable the cache. I will be happy to open a new PR once this is merged that will disable the statement cache implicitly when calling set_authorizer()

[berkerpeksag]

I'd prefer disabling statement cache implicitly when set_authorizer() is called. There's no need to expose this to end users.

[rhettinger]

I do believe we should allow the user to disable the cache.

Why?

[palaviv]

I think we can close this issue and open a different one that will disable the cache implicitly on set_authorizer().