Closed vangheem closed 3 years ago
hmmm, some 3.8 tests failing. Something unrelated is wrong in tests it seems to me.
Merging #1121 (b477135) into master (c2eba32) will decrease coverage by
5.2%
. The diff coverage is16.7%
.:exclamation: Current head b477135 differs from pull request most recent head 4e2ec1e. Consider uploading reports for the commit 4e2ec1e to get more accurate results
@@ Coverage Diff @@
## master #1121 +/- ##
========================================
- Coverage 94.4% 89.3% -5.1%
========================================
Files 372 372
Lines 31919 31333 -586
========================================
- Hits 30131 27969 -2162
- Misses 1788 3364 +1576
Impacted Files | Coverage Δ | |
---|---|---|
guillotina/tests/test_catalog.py | 60.6% <0.0%> (-39.4%) |
:arrow_down: |
guillotina/tests/test_postgres.py | 10.5% <20.0%> (-89.5%) |
:arrow_down: |
guillotina/db/storages/pg.py | 34.9% <21.8%> (-53.6%) |
:arrow_down: |
guillotina/tests/test_cockroach.py | 17.7% <0.0%> (-77.9%) |
:arrow_down: |
guillotina/api/aggregation.py | 24.2% <0.0%> (-69.1%) |
:arrow_down: |
guillotina/db/storages/vacuum.py | 47.0% <0.0%> (-53.0%) |
:arrow_down: |
guillotina/db/storages/cockroach.py | 28.5% <0.0%> (-52.0%) |
:arrow_down: |
guillotina/contrib/catalog/pg/utility.py | 24.9% <0.0%> (-49.7%) |
:arrow_down: |
guillotina/db/factory.py | 47.0% <0.0%> (-45.9%) |
:arrow_down: |
... and 152 more |
The purpose of this PR is to get rid of our dedicated read connection and use of prepared statements:
transaction
mode. asyncpg already automatically creates prepared statements for all sql. By removing the use of prepared statements, you can now set the statement cache size to 0 for asyncpg and have guillotina work with pgbouncer.