fix: unused snapshot detection for targeting single parameterized test case #394
Conversation
Codecov Report
@@ Coverage Diff @@
## master #394 +/- ##
=========================================
Coverage 100.00% 100.00%
=========================================
Files 19 17 -2
Lines 1002 992 -10
=========================================
- Hits 1002 992 -10 |
| if any( | ||
| TestLocation(node).matches_snapshot_name(snapshot.name) | ||
| for node in self.ran_items | ||
| ) | ||
| if self._selected_items_match_name(snapshot.name) | ||
| or self._provided_nodes_match_name(snapshot.name, provided_nodes) |
There was a problem hiding this comment.
This is the main gist of the fix. Basically _selected_items_match_name use the keyword expressions when they are present or matches against the ran items when no keywords provided or _provided_nodes_match_name checks if the nodes given for the test file matching this fossil location matches the snapshot name.
| return self.all_items == self.ran_items and not self.is_providing_nodes | ||
| return self.all_items == self.ran_items |
There was a problem hiding this comment.
Moved the and not self.is_providing_nodes check to the point of usage at if self.ran_all_collected_tests and not any(provided_nodes):, makes the method name more accurate.
| maybe_opt_arg = None | ||
|
|
||
| for maybe_opt_arg, arg_value in arg_groups: | ||
| if maybe_opt_arg == "-k": # or maybe_opt_arg == "-m": |
There was a problem hiding this comment.
Ignoring the marker expression as they can not be used to match against snapshot names and are already used to select which pytest items in the collection to run.
src/syrupy/report.py
Outdated
| ) | ||
|
|
||
| import attr | ||
| import pytest | ||
| from _pytest.mark.expression import Expression |
There was a problem hiding this comment.
Alternative to importing the protected class is to roll our own parsing of -k expressions.
There was a problem hiding this comment.
Just confirming this works in the min. version of pytest we support?
There was a problem hiding this comment.
Haven't will test that out, probably should have a way to automate the CI range test
There was a problem hiding this comment.
Yeah no this does not work with pytest 5...
There was a problem hiding this comment.
Changed this to include a naive version of the Expression class for compatibility with older pytest versions while still giving us some of the scaffolding for keyword snapshot inclusion
| assert Path(*snapshot_path, "test_updated_1.ambr").exists() | ||
|
|
||
|
|
||
| def test_update_targets_only_selected_parametrized_tests_for_update_dash_k( |
There was a problem hiding this comment.
This test was updated to catch the case raised by this issue
## [0.8.2](v0.8.1...v0.8.2) (2020-10-30) ### Bug Fixes * unused snapshot detection for targeting single parameterized test case ([#394](#394)) ([e008935](e008935))
|
🎉 This PR is included in version 0.8.2 🎉 The release is available on GitHub release Your semantic-release bot 📦🚀 |
Description
Fix unused snapshot detection for targeting single parameterized test case.
Clarify test methods for integration tests and testing in general.
Additionally
__snapshot__/test_file.where extension is an empty string.Related Issues
Checklist
Additional Comments
No additional comments.