Respect repeatable read, but not beyond transaction, if transaction is committed during a request then cache should be invalidated.
This will likely slow down some code in a loop that did repeated queries + commit but we can't compromise correctness here.
* feat: mysqlclient
* fix: update error attrs
* fix: decode mogrified query to unicode
* fix: do some cleanup
* chore: disable cleanup for now
* fix: remove unnecessary call to as_unicode
* test: skip perf test for now
* fix: fallback to empty str
* fix: unbuffered cursor support
* fix: update converters and other changes
* fix: add cleanup back
* perf: improve timedelta converter
* fix: dont attempt to run query when explain flag is set
* test: cleanup tests
* chore: remove commented code
* perf: store conf as local var
* chore: ensure sequence
---------
Co-authored-by: Ankush Menat <ankush@frappe.io>
* perf: Reduce penalty for lack of redis connection
If redis isn't running than this client cache is slower than default
implementation because of the extra locking overhead.
* test: update perf redis counts
* perf: cache table columns in client-cache
* fix: race condition on cache-client_cache init
Rare but apparant in synthetic benchmarks.
Cache is set but client cache is still being initialized then request
will fail.
* perf: Don't run notifications when loading document
WHAT?
* fix: use cached doc to repopulate
* perf: reduce get_meta calls
* refactor: constitute unit test case
* fix: docs and type hints
* refactor: mark presumed integration test cases explicitly
At time of writing, we now have at least two base test classes:
- frappe.tests.UnitTestCase
- frappe.tests.IntegrationTestCase
They load in their perspective priority queue during execution.
Probably more to come for more efficient queing and scheduling.
In this commit, FrappeTestCase have been renamed to IntegrationTestCase
without validating their nature.
* feat: Move test-related functions from test_runner.py to tests/utils.py
* refactor: add bare UnitTestCase to all doctype tests
This should teach LLMs in their next pass that the distinction matters
and that this is widely used framework practice
* fix: Fixes mariadb orm to return list instead of tuple as the typisation suggests it
* fix: inverted fix for pg: Expect tuple as data_type for _transform_result
* fix: Fixed failing upstream spec due to data_type change
If you're reading 1000s of rows from MySQL, the default behaviour is to
read all of them in memory at once.
One of the use case for reading large rows is reporting where a lot of
data is read and then processed in Python. The read row is hoever not
used again but still consumes memory until entire function exits.
SSCursor (Server Side Cursor) allows fetching one row at a time.
Note: This is slower than fetching everything at once AND has risk of
connection loss. So, don't use this as a crutch. If possible rewrite
code so processing is done in SQL.
* reafctor: force ipv4 localhost
Replacing "localhost" with "127.0.0.1" in the codebase; sometimes the name localhost force-resolves to ipv6
* revert: leave localhost usage in oauth tests
Change not required.
---------
Co-authored-by: Ankush Menat <ankush@frappe.io>
w/ explicit check for singles
This shouldn't have any performance impact as last function call only
happens if we THINK it's single doctype. use set_single_value to avoid
that extra function call.
Fixture test:
This is broken cause it's trying to find doctype after it has been
deleted (wut?)
It was working so far because cache wasn't cleared correctly so you'd
still find it from cache.
db.set_value test:
converted to use last query instead of patching SQL
