You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
* Unrestrict runtime deps
* Remove rubocop-capybara as its no longer needed or suggested and slightly update rubocop gems
* Slight tidy to interceptor spec
* Update CCK testing version
* Downgrade CCK to v17 whilst investigating issues
* WIP for adding id to test run id
* Fix duplicate testing block
* Run release pipeline tests on 3.4 also
* Update docs
* Minor simplification to main runtime
* Fix bug with deprecator message
* Fix deprecator spec
* Minor tidy to backtrace filter
* Change ID generation to be much simpler - Add this to output envelope during the test run started event
* Remove redundant input event when finding / generating test run id
* Add Test run ID to test case ready event
Change order of broadcast such that test case ready comes after test run started
* Add test run started id to testrunfinished message
* Hook refactors
Rename phase to type for type of hook
Add hook type to envelope creation
Ensure enum conversion is adherent to message enum validation (Around hook is the outlier here - InstallPlugin can be converted adherently)
* Downgrade CCK back to 16.2 and hardcode cck tests to just retry
* Recommit v17 of CCK
* Experiment move TestCaseReadyEvent lower down
* Re-run all CCK tests again;
* Change to CCK steps for v18 conformance
* Update to v18 of cck
* Triage one failing test
* Bugfix wrong media type
* Add in deprecated steps for CCK conformance (currently)
* Add in truffleruby guard for some tests
* Hard code the truffleruby test to use v24
* Add truffleruby 24 into release pipeline
* Fix rubocop
* Compress a few lines down
* Re-test the whole of the CCK again
* Clarify deprecation notice for CCK attachment steps
* Remove redundant require for testrunstarted query
* Fix missing arg in step
* Tidy up of lines / style
* Fix up bad copy and paste
* Add CCK strawman for validating the testRunId consistency
* Clean up test for id checking
* Tidy up backtrace filter test
* Optimise iterator for testRunStartId checks
* Update CCK conformance in readme;
* Switch back to truffleruby head
* Add date time stamp for removal of pragma skip for truffleruby error;
* Refactor filter into non-procedural code
* Refactor specs partially to use a agnostic anonymous class
* Remove iVar references and fix up remaining cop offenses
* Initial mapped backtrace must always exist as some array so no need to check if it exists
* Update CCK conformance note
* Add in final changelogs for backtrace filtering
* Bump minimum ruby to 3.1 over 3.0 owing to time taken for v10 conformance
* Flag truffleruby support as v24+
- Added dependencies that will no longer be part of the ruby stdlib ([jeromeag](https://github.com/jerome))
14
-
- Updated `cucumber-compatibility-kit` to v16 ([luke-hill](https://github.com/luke-hill))
14
+
- Updated `cucumber-compatibility-kit` to v18
15
+
- NB: This change has also facilitated a tweak to some output messages to include new fields such as `testRunStartedId`
16
+
for full changes and expectations of what will be new check the [messages](https://github.com/cucumber/messages/blob/main/CHANGELOG.md#2700---2024-10-26)
17
+
([luke-hill](https://github.com/luke-hill))
15
18
- Changed compatibility testing to fully lean on external assets instead of duplicating them ([luke-hill](https://github.com/luke-hill))
16
19
- Permit usage of latest versions of all cucumber internal gems
17
20
- The auto-generation of all placeholder `pending` steps has been refactored (_There should be no visible changes
@@ -20,6 +23,8 @@ but the code is now refactored to work using newer ruby standards_)
20
23
should lean on an appropriate testing library for this)
21
24
-`Cucumber.use_legacy_autoloader` that was intended as a stop-gap for v4/v5 is now deprecated (People
22
25
that need to rely on procedural loading / reloading of files should use method invocations) ([luke-hill](https://github.com/luke-hill))
26
+
- The backtrace filtering logic has been changed (_There should be no visible changes but the code is now
27
+
refactored to use newer, less procedural ruby_)
23
28
24
29
### Fixed
25
30
- Fixed an issue where a change to one example in compatibility testing wasn't fully adhered to ([luke-hill](https://github.com/luke-hill))
@@ -32,7 +37,7 @@ that need to rely on procedural loading / reloading of files should use method i
32
37
- Associated legacy scripts for updating dependencies that are no longer used
33
38
- Tag Expressions using legacy syntax that were handled / sanitized are no longer done so
34
39
(This applies to both regular usage and internal testing)
35
-
- Removed support for Ruby 2.7 ([luke-hill](https://github.com/luke-hill))
40
+
- Removed support for Ruby 2.7 and 3.0 (Minimum ruby is now 3.1) ([luke-hill](https://github.com/luke-hill))
36
41
- Unindentation support for snippet generator / tests (Heredocs are much better now) ([luke-hill](https://github.com/luke-hill))
37
42
38
43
## [9.2.1] - 2025-01-12
@@ -70,7 +75,7 @@ All logic contained in [compatibility](./compatibility) ([luke-hill](https://git
70
75
71
76
## [9.1.0] - 2023-11-14
72
77
### Changed
73
-
- First couple of passes of tidying up approximately 40% of the manual fix cops
78
+
- First iteration to tidy up approximately 40% of the manual fix cops
See [.github/RELEASING](https://github.com/cucumber/.github/blob/main/RELEASING.md).
2
-
3
-
## When done ##
4
-
5
-
Update the cucumber-ruby version in the documentation project:
6
-
7
-
*https://github.com/cucumber/docs.cucumber.io
8
-
9
-
The cucumber-ruby version for the docs is specified in the docs [versions.yaml](https://github.com/cucumber/docs.cucumber.io/blob/master/data/versions.yaml)
0 commit comments