fix(deps): update module github.com/onsi/ginkgo to v2

[red-hat-konflux-kflux-prd-rh03]

ℹ️ Note

This PR body was truncated due to platform limits.

This PR contains the following updates:

Package	Change	Age	Confidence
github.com/onsi/ginkgo	v1.16.5 → v2.32.1

---

Release Notes

onsi/ginkgo (github.com/onsi/ginkgo)

v2.32.1

Compare Source

2.32.1

Fixes

• Defer AfterAll until repeated spec completes [e647b3b]

v2.32.0

Compare Source

2.32.0

-fd generate RSpec-style documentation output. Thank @​woodie !
--sleep-on-failure pauses a failed spec before teardown. Thanks @​qinqon !

v2.31.0

Compare Source

2.31.0

Add a bunch of Claude Skills via the marketplace:

/plugin marketplace add onsi/ginkgo
/plugin install ginkgo@ginkgo

v2.30.0

Compare Source

2.30.0

Features

Ginkgo now allows extentions/global.Reset to support running multiple suites from within a single process. This may take some massaging on your part (see 1672) but can dramatically speed up codebases with O(hundreds) of test suites.

Thanks @​lawrencejones !

Fixes

• Fix nested --github-output group for progress report nested inside timeline [4f62d7a]

v2.29.0

Compare Source

2.29.0

GinkgoHelperGo makes it easier to write test helpers that need to run in goroutines. Specifically, it makes managing the failure state and capturing failure panics correctly straightforward.

ginkgo outline now includes entries defined in DescribeTableSubtree

v2.28.3

Compare Source

2.28.3

Maintenance

Bump all dependencies

v2.28.2

Compare Source

2.28.2

• Add ArtifactDir() to support Go 1.26 testing.TB interface [f3a36b6]
• Implement shell completion [94151c8]
• Add asan CLI option mirroring msan implementation [4d21dbb]
• Bump uri from 1.0.3 to 1.0.4 in /docs (#​1630) [c102161]
• fix aspect ratio [9619647]
• update logos [5779304]

v2.28.1

Compare Source

2.28.1

Update all dependencies. This auto-updated the required version of Go to 1.24, consistent with the fact that Go 1.23 has been out of support for almost six months.

v2.28.0

Compare Source

2.28.0

Ginkgo's SemVer filter now supports filtering multiple components by SemVer version:

It("should work in a specific version range (1.0.0, 2.0.0) and third-party dependency redis in [8.0.0, ~)", SemVerConstraint(">= 3.2.0"), ComponentSemVerConstraint("redis", ">= 8.0.0") func() {
    // This test will only run when version is between 1.0.0 (exclusive) and 2.0.0 (exclusive) and redis version is >= 8.0.0
})

can be filtered in or out with an invocation like:

ginkgo --sem-ver-filter="2.1.1, redis=8.2.0"

Huge thanks to @​Icarus9913 for working on this!

v2.27.5

Compare Source

2.27.5

Fixes

Don't make a new formatter for each GinkgoT(); that's just silly and uses precious memory

v2.27.4

Compare Source

2.27.4

Fixes

• CurrentTreeConstructionNodeReport: fix for nested container nodes [59bc751]

v2.27.3

Compare Source

2.27.3

Fixes

report exit result in case of failure [1c9f356]
fix data race [ece19c8]

v2.27.2

Compare Source

2.27.2

Fixes

• inline automaxprocs to simplify dependencies; this will be removed when Go 1.26 comes out [a69113a]

Maintenance

• Fix syntax errors and typo [a99c6e0]
• Fix paragraph position error [f993df5]

v2.27.1

Compare Source

2.27.1

Fixes

• Fix Ginkgo Reporter slice-bounds panic [606c1cb]
• Bug Fix: Add GinkoTBWrapper.Attr() and GinkoTBWrapper.Output() [a6463b3]

v2.27.0

Compare Source

2.27.0

Features

Transforming Nodes during Tree Construction

This release adds support for NodeArgsTransformers that can be registered with AddTreeConstructionNodeArgsTransformer.

These are called during the tree construction phase as nodes are constructed and can modify the node strings and decorators. This enables frameworks built on top of Ginkgo to modify Ginkgo nodes and enforce conventions.

Learn more here.

Spec Prioritization

A new SpecPriority(int) decorator has been added. Ginkgo will honor priority when ordering specs, ensuring that higher priority specs start running before lower priority specs

Learn more here.

Maintenance

• Bump rexml from 3.4.0 to 3.4.2 in /docs (#​1595) [1333dae]
• Bump github.com/gkampitakis/go-snaps from 0.5.14 to 0.5.15 (#​1600) [17ae63e]

v2.26.0

Compare Source

2.26.0

Features

Ginkgo can now generate json-formatted reports that are compatible with the go test json format. Use ginkgo --gojson-report=report.go.json. This is not intended to be a replacement for Ginkgo's native json format which is more information rich and better models Ginkgo's test structure semantics.

v2.25.3

Compare Source

2.25.3

Fixes

• emit --github-output group only for progress report itself [f01aed1]

v2.25.2

Compare Source

2.25.2

Fixes

Add github output group for progress report content

Maintenance

Bump Gomega

v2.25.1

Compare Source

2.25.1

Fixes

• fix(types): ignore nameless nodes on FullText() [10866d3]
• chore: fix some CodeQL warnings [2e42cff]

v2.25.0

Compare Source

2.25.0

AroundNode

This release introduces a new decorator to support more complex spec setup usecases.

AroundNode registers a function that runs before each individual node. This is considered a more advanced decorator.

Please read the docs for more information and some examples.

Allowed signatures:

• AroundNode(func()) - func will be called before the node is run.
• AroundNode(func(ctx context.Context) context.Context) - func can wrap the passed in context and return a new one which will be passed on to the node.
• AroundNode(func(ctx context.Context, body func(ctx context.Context))) - ctx is the context for the node and body is a function that must be called to run the node. This gives you complete control over what runs before and after the node.

Multiple AroundNode decorators can be applied to a single node and they will run in the order they are applied.

Unlike setup nodes like BeforeEach and DeferCleanup, AroundNode is guaranteed to run in the same goroutine as the decorated node. This is necessary when working with lower-level libraries that must run on a single thread (you can call runtime.LockOSThread() in the AroundNode to ensure that the node runs on a single thread).

Since AroundNode allows you to modify the context you can also use AroundNode to implement shared setup that attaches values to the context.

If applied to a container, AroundNode will run before every node in the container. Including setup nodes like BeforeEach and DeferCleanup.

AroundNode can also be applied to RunSpecs to run before every node in the suite. This opens up new mechanisms for instrumenting individual nodes across an entire suite.

v2.24.0

Compare Source

2.24.0

Features

Specs can now be decorated with (e.g.) SemVerConstraint("2.1.0") and ginkgo --sem-ver-filter="2.1.1" will only run constrained specs that match the requested version. Learn more in the docs here! Thanks to @​Icarus9913 for the PR.

Fixes

• remove -o from run command [3f5d379]. fixes #​1582

Maintenance

Numerous dependency bumps and documentation fixes

v2.23.4

Compare Source

2.23.4

Prior to this release Ginkgo would compute the incorrect number of available CPUs when running with -p in a linux container. Thanks to @​emirot for the fix!

Features

• Add automaxprocs for using CPUQuota [2b9c428]

Fixes

• clarify gotchas about -vet flag [1f59d07]

Maintenance

• bump dependencies [2d134d5]

v2.23.3

Compare Source

2.23.3

Fixes

• allow - as a standalone argument [cfcc1a5]
• Bug Fix: Add GinkoTBWrapper.Chdir() and GinkoTBWrapper.Context() [feaf292]
• ignore exit code for symbol test on linux [88e2282]

v2.23.2

Compare Source

2.23.2

🎉🎉🎉

At long last, some long-standing performance gaps between ginkgo and go test have been resolved!

Ginkgo operates by running go test -c to generate test binaries, and then running those binaries. It turns out that the compilation step of go test -c is slower than go test's compilation step because go test strips out debug symbols (ldflags=-w) whereas go test -c does not.

Ginkgo now passes the appropriate ldflags to go test -c when running specs to strip out symbols. This is only done when it is safe to do so and symbols are preferred when profiling is enabled and when ginkgo build is called explicitly.

This, coupled, with the instructions for disabling XProtect on MacOS yields a much better performance experience with Ginkgo.

v2.23.1

Compare Source

2.23.1

🚨 For users on MacOS 🚨

A long-standing Ginkgo performance issue on MacOS seems to be due to mac's antimalware XProtect. You can follow the instructions here to disable it in your terminal. Doing so sped up Ginkgo's own test suite from 1m8s to 47s.

Fixes

Ginkgo's CLI is now a bit clearer if you pass flags in incorrectly:

• make it clearer that you need to pass a filename to the various profile flags, not an absolute directory [a0e52ff]
• emit an error and exit if the ginkgo invocation includes flags after positional arguments [b799d8d]

This might cause existing CI builds to fail. If so then it's likely that your CI build was misconfigured and should be corrected. Open an issue if you need help.

v2.23.0

Compare Source

2.23.0

Ginkgo 2.23.0 adds a handful of methods to GinkgoT() to make it compatible with the testing.TB interface in Go 1.24. GinkgoT().Context(), in particular, is a useful shorthand for generating a new context that will clean itself up in a DeferCleanup(). This has subtle behavior differences from the golang implementation but should make sense in a Ginkgo... um... context.

Features

• bump to go 1.24.0 - support new testing.TB methods and add a test to cover testing.TB regressions [37a511b]

Fixes

• fix edge case where build -o is pointing at an explicit file, not a directory [7556a86]
• Fix binary paths when precompiling multiple suites. [4df06c6]

Maintenance

• Fix: Correct Markdown list rendering in MIGRATING_TO_V2.md [cbcf39a]
• docs: fix test workflow badge (#​1512) [9b261ff]
• Bump golang.org/x/net in /integration/_fixtures/version_mismatch_fixture (#​1516) [00f19c8]
• Bump golang.org/x/tools from 0.28.0 to 0.30.0 (#​1515) [e98a4df]
• Bump activesupport from 6.0.6.1 to 6.1.7.5 in /docs (#​1504) [60cc4e2]
• Bump github-pages from 231 to 232 in /docs (#​1447) [fea6f2d]
• Bump rexml from 3.2.8 to 3.3.9 in /docs (#​1497) [31d7813]
• Bump webrick from 1.8.1 to 1.9.1 in /docs (#​1501) [fc3bbd6]
• Code linting (#​1500) [aee0d56]
• change interface{} to any (#​1502) [809a710]

v2.22.2

Compare Source

What's Changed

• Bump golang.org/x/net from 0.32.0 to 0.33.0 by @​dependabot in #​1496
• Bump golang.org/x/crypto from 0.17.0 to 0.31.0 in /ginkgo/performance/_fixtures/performance_fixture by @​dependabot in #​1498
• Bump github.com/onsi/gomega from 1.36.1 to 1.36.2 by @​dependabot in #​1499

Full Changelog: onsi/ginkgo@v2.22.1...v2.22.2

v2.22.1

Compare Source

2.22.1

Fixes

Fix CSV encoding

• Update tests [aab3da6]
• Properly encode CSV rows [c09df39]
• Add test case for proper csv escaping [96a80fc]
• Add meta-test [43dad69]

Maintenance

• ensure *.test files are gitignored so we don't accidentally commit compiled tests again [c88c634]
• remove golang.org/x/net/context in favour of stdlib context [4df44bf]

v2.22.0

Compare Source

2.22.0

Features

• Add label to serial nodes [0fcaa08]

This allows serial tests to be filtered using the label-filter

Maintenance

Various doc fixes

v2.21.0

Compare Source

2.21.0

Features

• add support for GINKGO_TIME_FORMAT [a69eb39]
• add GINKGO_NO_COLOR to disable colors via environment variables [bcab9c8]

Fixes

• increase threshold in timeline matcher [e548367]
• Fix the document by replacing SpecsThatWillBeRun with SpecsThatWillRun
  [c2c4d3c]

Maintenance

• bump various dependencies [7e65a00]

v2.20.2

Compare Source

2.20.2

Require Go 1.22+

Maintenance

• bump go to v1.22 [a671816]

v2.20.1

Compare Source

2.20.1

Fixes

• make BeSpecEvent duration matcher more forgiving [d6f9640]

v2.20.0

Compare Source

2.20.0

Features

• Add buildvcs flag [be5ab95]

Maintenance

• Add update-deps to makefile [d303d14]
• bump all dependencies [7a50221]

v2.19.1

Compare Source

2.19.1

Fixes

• update supported platforms for race conditions [63c8c30]
• [build] Allow custom name for binaries. [ff41e27]

Maintenance

• bump gomega [76f4e0c]
• Bump rexml from 3.2.6 to 3.2.8 in /docs (#​1417) [b69c00d]
• Bump golang.org/x/sys from 0.20.0 to 0.21.0 (#​1425) [f097741]

v2.19.0

Compare Source

2.19.0

Features

Label Sets allow for more expressive and flexible label filtering.

v2.18.0

Compare Source

2.18.0

Features

• Add --slience-skips and --force-newlines [f010b65]
• fail when no tests were run and --fail-on-empty was set [d80eebe]

Fixes

• Fix table entry context edge case [42013d6]

Maintenance

• Bump golang.org/x/tools from 0.20.0 to 0.21.0 (#​1406) [fcf1fd7]
• Bump github.com/onsi/gomega from 1.33.0 to 1.33.1 (#​1399) [8bb14fd]
• Bump golang.org/x/net from 0.24.0 to 0.25.0 (#​1407) [04bfad7]

v2.17.3

Compare Source

2.17.3

Fixes

ginkgo watch now ignores hidden files [bde6e00]

v2.17.2

Compare Source

2.17.2

Fixes

• fix: close files [32259c8]
• fix github output log level for skipped specs [780e7a3]

Maintenance

• Bump github.com/google/pprof [d91fe4e]
• Bump github.com/go-task/slim-sprig to v3 [8cb662e]
• Bump golang.org/x/net in /integration/_fixtures/version_mismatch_fixture (#​1391) [3134422]
• Bump github-pages from 230 to 231 in /docs (#​1384) [eca81b4]
• Bump golang.org/x/tools from 0.19.0 to 0.20.0 (#​1383) [760def8]
• Bump golang.org/x/net from 0.23.0 to 0.24.0 (#​1381) [4ce33f4]
• Fix test for gomega version bump [f2fcd97]
• Bump github.com/onsi/gomega from 1.30.0 to 1.33.0 (#​1390) [fd622d2]
• Bump golang.org/x/tools from 0.17.0 to 0.19.0 (#​1368) [5474a26]
• Bump github-pages from 229 to 230 in /docs (#​1359) [e6d1170]
• Bump google.golang.org/protobuf from 1.28.0 to 1.33.0 (#​1374) [7f447b2]
• Bump golang.org/x/net from 0.20.0 to 0.23.0 (#​1380) [f15239a]

v2.17.1

Compare Source

2.17.1

Fixes

• If the user sets --seed=0, make sure all parallel nodes get the same seed [af0330d]

v2.17.0

Compare Source

2.17.0

Features

• add --github-output for nicer output in github actions [e8a2056]

Maintenance

• fix typo in core_dsl.go [977bc6f]
• Fix typo in docs [e297e7b]

v2.16.0

Compare Source

2.16.0

Features

• add SpecContext to reporting nodes

Fixes

• merge coverages instead of combining them (#​1329) (#​1340) [23f0cc5]
• core_dsl: disable Getwd() with environment variable (#​1357) [cd418b7]

Maintenance

• docs/index.md: Typo [2cebe8d]
• fix docs [06de431]
• chore: test with Go 1.22 (#​1352) [898cba9]
• Bump golang.org/x/tools from 0.16.1 to 0.17.0 (#​1336) [17ae120]
• Bump golang.org/x/sys from 0.15.0 to 0.16.0 (#​1327) [5a179ed]
• Bump github.com/go-logr/logr from 1.3.0 to 1.4.1 (#​1321) [a1e6b69]
• Bump github-pages and jekyll-feed in /docs (#​1351) [d52951d]
• Fix docs for handling failures in goroutines (#​1339) [4471b2e]

v2.15.0

Compare Source

2.15.0

Features

• JUnit reports now interpret Label(owner:X) and set owner to X. [8f3bd70]
• include cancellation reason when cancelling spec context [96e915c]

Fixes

• emit output of failed go tool cover invocation so users can try to debug things for themselves [c245d09]
• fix outline when using nodot in ginkgo v2 [dca77c8]
• Document areas where GinkgoT() behaves differently from testing.T [dbaf18f]
• bugfix(docs): use Unsetenv instead of Clearenv (#​1337) [6f67a14]

Maintenance

• Bump to go 1.20 [4fcd0b3]

v2.14.0

Compare Source

2.14.0

Features

You can now use GinkgoTB() when you need an instance of testing.TB to pass to a library.

Prior to this release table testing only supported generating individual Its for each test entry. DescribeTableSubtree extends table testing support to entire testing subtrees - under the hood DescrieTableSubtree generates a new container for each entry and invokes your function to fill our the container. See the docs to learn more.

• Introduce DescribeTableSubtree [65ec56d]
• add GinkgoTB() to docs [4a2c832]
• Add GinkgoTB() function (#​1333) [92b6744]

Fixes

• Fix typo in internal/suite.go (#​1332) [beb9507]
• Fix typo in docs/index.md (#​1319) [4ac3a13]
• allow wasm to compile with ginkgo present (#​1311) [b2e5bc5]

Maintenance

• Bump golang.org/x/tools from 0.16.0 to 0.16.1 (#​1316) [465a8ec]
• Bump actions/setup-go from 4 to 5 (#​1313) [eab0e40]
• Bump github/codeql-action from 2 to 3 (#​1317) [fbf9724]
• Bump golang.org/x/crypto (#​1318) [3ee80ee]
• Bump golang.org/x/tools from 0.14.0 to 0.16.0 (#​1306) [123e1d5]
• Bump github.com/onsi/gomega from 1.29.0 to 1.30.0 (#​1297) [558f6e0]
• Bump golang.org/x/net from 0.17.0 to 0.19.0 (#​1307) [84ff7f3]

v2.13.2

Compare Source

2.13.2

Fixes

• Fix file handler leak (#​1309) [e2e81c8]
• Avoid allocations with (*regexp.Regexp).MatchString (#​1302) [3b2a2a7]

v2.13.1

Compare Source

2.13.1

Fixes

• 1296 fix(precompiled test guite): exec bit check omitted on Windows (#​1301) [26eea01]

Maintenance

• Bump github.com/go-logr/logr from 1.2.4 to 1.3.0 (#​1291) [7161a9d]
• Bump golang.org/x/sys from 0.13.0 to 0.14.0 (#​1295) [7fc7b10]
• Bump golang.org/x/tools from 0.12.0 to 0.14.0 (#​1282) [74bbd65]
• Bump github.com/onsi/gomega from 1.27.10 to 1.29.0 (#​1290) [9373633]
• Bump golang.org/x/net in /integration/_fixtures/version_mismatch_fixture (#​1286) [6e3cf65]

v2.13.0

Compare Source

2.13.0

Features

Add PreviewSpect() to enable programmatic preview access to the suite report (fixes #​1225)

v2.12.1

Compare Source

2.12.1

Fixes

• Print logr prefix if it exists (#​1275) [90d4846]

Maintenance

• Bump actions/checkout from 3 to 4 (#​1271) [555f543]
• Bump golang.org/x/sys from 0.11.0 to 0.12.0 (#​1270) [d867b7d]

v2.12.0

Compare Source

2.12.0

Features

• feat: allow MustPassRepeatedly decorator to be set at suite level (#​1266) [05de518]

Fixes

• fix-errors-in-readme (#​1244) [27c2f5d]

Maintenance

Various chores/dependency bumps.

v2.11.0

Compare Source

2.11.0

In prior versions of Ginkgo specs the CLI filter flags (e.g. --focus, --label-filter) would override any programmatic focus. This behavior has proved surprising and confusing in at least the following ways:

• users cannot combine programmatic filters and CLI filters to more efficiently select subsets of tests
• CLI filters can override programmatic focus on CI systems resulting in an exit code of 0 despite the presence of (incorrectly!) committed focused specs.

Going forward Ginkgo will AND all programmatic and CLI filters. Moreover, the presence of any programmatic focused tests will always result in a non-zero exit code.

This change is technically a change in Ginkgo's external contract and may require some users to make changes to successfully adopt. Specifically: it's possible some users were intentionally using CLI filters to override programmatic focus. If this is you please open an issue so we can explore solutions to the underlying problem you are trying to solve.

Fixes

• Programmatic focus is no longer overwrriten by CLI filters [d6bba86]

Maintenance

• Bump github.com/onsi/gomega from 1.27.7 to 1.27.8 (#​1218) [4a70a38]
• Bump golang.org/x/sys from 0.8.0 to 0.9.0 (#​1219) [97eda4d]

v2.10.0

Compare Source

2.10.0

Features

• feat(ginkgo/generators): add --tags flag (#​1216) [a782a77]
  adds a new --tags flag to ginkgo generate

Fixes

• Fix broken link of MIGRATING_TO_V2.md (#​1217) [548d78e]

Maintenance

• Bump golang.org/x/tools from 0.9.1 to 0.9.3 (#​1215) [2b76a5e]

v2.9.7

Compare Source

2.9.7

Fixes

• fix race when multiple defercleanups are called in goroutines [07fc3a0]

v2.9.6

Compare Source

2.9.6

Fixes

• fix: create parent directory before report files (#​1212) [0ac65de]

Maintenance

• Bump github.com/onsi/gomega from 1.27.6 to 1.27.7 (#​1202) [3e39231]

v2.9.5

Compare Source

2.9.5

Fixes

• ensure the correct deterministic sort order is produced when ordered specs are generated by a helper function [7fa0b6b]

Maintenance

• fix generators link (#​1200) [9f9d8b9]
• Bump golang.org/x/tools from 0.8.0 to 0.9.1 (#​1196) [150e3f2]
• fix spelling err in docs (#​1199) [0013b1a]
• Bump golang.org/x/sys from 0.7.0 to 0.8.0 (#​1193) [9e9e3e5]

v2.9.4

Compare Source

2.9.4

Fixes

• fix hang with ginkgo -p (#​1192) [15d4bdc] - this addresses a long standing issue related to Ginkgo hanging when a child process spawned by the test does not exit.

• fix: fail fast may cause Serial spec or cleanup Node interrupted (#​1178) [8dea88b] - prior to this there was a small gap in which specs on other processes might start even if one process has tried to abort the suite.

Maintenance

• Document run order when multiple setup nodes are at the same nesting level [903be81]

v2.9.3

Compare Source

2.9.3

Features

• Add RenderTimeline to GinkgoT() [c0c77b6]

Fixes

• update Measure deprecation message. fixes #​1176 [227c662]
• add newlines to GinkgoLogr (#​1170) (#​1171) [0de0e7c]

Maintenance

• Bump commonmarker from 0.23.8 to 0.23.9 in /docs (#​1183) [8b925ab]
• Bump nokogiri from 1.14.1 to 1.14.3 in /docs (#​1184) [e3795a4]
• Bump golang.org/x/tools from 0.7.0 to 0.8.0 (#​1182) [b453793]
• Bump actions/setup-go from 3 to 4 (#​1164) [73ed75b]
• Bump github.com/onsi/gomega from 1.27.4 to 1.27.6 (#​1173) [0a2bc64]
• Bump github.com/go-logr/logr from 1.2.3 to 1.2.4 (#​1174) [f41c557]
• Bump golang.org/x/sys from 0.6.0 to 0.7.0 (#​1179) [8e423e5]

v2.9.2

Compare Source

2.9.2

Maintenance

• Bump github.com/go-task/slim-sprig (#​1167) [3fcc5bf]
• Bump github.com/onsi/gomega from 1.27.3 to 1.27.4 (#​1163) [6143ffe]

v2.9.1

Compare Source

2.9.1

Fixes

This release fixes a longstanding issue where ginkgo -coverpkg=./... would not work. This is now resolved and fixes #​1161 and #​995

• Support -coverpkg=./... [26ca1b5]
• document coverpkg a bit more clearly [fc44c3b]

Maintenance

• bump various dependencies
• Improve Documentation and fix typo (#​1158) [93de676]

v2.9.0

Compare Source

2.9.0

Features

• AttachProgressReporter is an experimental feature that allows users to provide arbitrary information when a ProgressReport is requested [28801fe]

• GinkgoT() has been expanded to include several Ginkgo-specific methods [2bd5a3b]

  The intent is to enable the development of third-party libraries that integrate deeply with Ginkgo using GinkgoT() to access Ginkgo's functionality.

v2.8.4

Compare Source

2.8.4

Features

• Add OmitSuiteSetupNodes to JunitReportConfig (#​1147) [979fbc2]
• Add a reference to ginkgolinter in docs.index.md (#​1143) [8432589]

Fixes

• rename tools hack to see if it fixes things for downstream users [a8bb39a]

Maintenance

• Bump golang.org/x/text (#​1144) [41b2a8a]
• Bump github.com/onsi/gomega from 1.27.0 to 1.27.1 (#​1142) [7c4f583]

v2.8.3

Compare Source

2.8.3

Released to fix security issue in golang.org/x/net dependency

Maintenance

• Bump golang.org/x/net from 0.6.0 to 0.7.0 (#​1141) [fc1a02e]
• remove tools.go hack from documentation [0718693]

v2.8.2

Compare Source

2.8.2

Ginkgo now includes a tools.go file in the root directory of the ginkgo package. This should allow modules that simply go get github.com/onsi/ginkgo/v2 to also pull in the CLI dependencies. This obviates the need for consumers of Ginkgo to have their own tools.go file and makes it simpler to ensure that the version of the ginkgo CLI being used matches the version of the library. You can simply run go run github.com/onsi/ginkgo/v2/ginkgo to run the version of the cli associated with your package go.mod.

Maintenance

• Bump github.com/onsi/gomega from 1.26.0 to 1.27.0 (#​1139) [5767b0a]
• Fix minor typos (#​1138) [e1e9723]
• Fix link in V2 Migration Guide (#​1137) [a588f60]

v2.8.1

Compare Source

2.8.1

Fixes

• lock around default report output to avoid triggering the race detector when calling By from goroutines [2d5075a]
• don't run ReportEntries through sprintf [febbe38]

Maintenance

• Bump golang.org/x/tools from 0.5.0 to 0.6.0 (#​1135) [11a4860]
• test: update matrix for Go 1.20 (#​1130) [4890a62]
• Bump golang.org/x/sys from 0.4.0 to 0.5.0 (#​1133) [a774638]
• Bump github.com/onsi/gomega from 1.25.0 to 1.26.0 (#​1120) [3f233bd]
• Bump github-pages from 227 to 228 in /docs (#​1131) [f9b8649]
• Bump activesupport from 6.0.6 to 6.0.6.1 in /docs (#​1127) [6f8c042]
• Update index.md with instructions on how to upgrade Ginkgo [833a75e]

v2.8.0

Compare Source

2.8.0

Features

• Introduce GinkgoHelper() to track and exclude helper functions from potential CodeLocations [e19f556]

Modeled after testing.T.Helper(). Now, rather than write code like:

func helper(model Model) {
    Expect(model).WithOffset(1).To(BeValid())
    Expect(model.SerialNumber).WithOffset(1).To(MatchRegexp(/[a-f0-9]*/))
}

you can stop tracking offsets (which makes nesting composing helpers nearly impossible) and simply write:

func helper(model Model) {
    GinkgoHelper()
    Expect(model).To(BeValid())
    Expect(model.SerialNumber).To(MatchRegexp(/[a-f0-9]*/))
}

• Introduce GinkgoLabelFilter() and Label().MatchesLabelFilter() to make it possible to programmatically match filters (fixes #​1119) [2f6597c]

You can now write code like this:

BeforeSuite(func() {
	if Label("slow").MatchesLabelFilter(GinkgoLabelFilter()) {
		// do slow setup
	}

	if Label("fast").MatchesLabelFilter(GinkgoLabelFilter()) {
		// do fast setup
	}
})

to programmatically check whether a given set of labels will match the configured --label-filter.

Maintenance

• Bump webrick from 1.7.0 to 1.8.1 in /docs (#​1125) [ea4966e]
• cdeql: add ruby language (#​1124) [9dd275b]
• dependabot: add bundler package-ecosystem for docs (#​1123) [14e7bdd]

v2.7.1

Compare Source

2.7.1

Fixes

• Bring back SuiteConfig.EmitSpecProgress to avoid compilation issue for consumers that set it manually [d2a1cb0]

Maintenance

• Bump github.com/onsi/gomega from 1.24.2 to 1.25.0 (#​1118) [cafece6]
• Bump golang.org/x/tools from 0.4.0 to 0.5.0 (#​1111) [eda66c2]
• Bump golang.org/x/sys from 0.3.0 to 0.4.0 (#​1112) [ac5ccaa]
• Bump github.com/onsi/gomega from 1.24.1 to 1.24.2 (#​1097) [eee6480]

v2.7.0

Compare Source

2.7.0

Features

• Introduce ContinueOnFailure for Ordered containers [e0123ca] - Ordered containers that ar

✂ Note

PR body was truncated to here.

---

Configuration

📅 Schedule: (in timezone UTC)

• Branch creation
  • Between 02:00 AM and 04:59 AM, Monday through Friday (* 2-4 * * 1-5)
• Automerge
  • At any time (no schedule defined)

🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.

♻ Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

🔕 Ignore: Close this PR and you won't be reminded about this update again.

---

• If you want to rebase/retry this PR, check this box

---

To execute skipped test pipelines write comment /ok-to-test.

---

Documentation

Find out how to configure dependency updates in MintMaker documentation or see all available configuration options in Renovate documentation.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 65.60%. Comparing base (c6e16cb) to head (2518704).

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #298   +/-   ##
=======================================
  Coverage   65.60%   65.60%           
=======================================
  Files          23       23           
  Lines        1608     1608           
=======================================
  Hits         1055     1055           
  Misses        473      473           
  Partials       80       80           

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[openshift-ci]

Hi @red-hat-konflux-kflux-prd-rh03[bot]. Thanks for your PR.

I'm waiting for a openshift member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work.

Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

Details

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

[fullsend-ai-retro]

🤖 Finished Retro · ✅ Success · Started 12:03 PM UTC · Completed 12:10 PM UTC

Commit: c11a260 · View workflow run →

[fullsend-ai-retro]

Retro: PR #298 — 4th retro on bot ginkgo v1→v2 dependency PR (no proposals)

This is the 4th retro dispatch on the same unmerged, bot-authored Renovate/MintMaker dependency update PR over 8 days (Aug 12–20). All four retros reached the same conclusion: zero fullsend code/review/triage/fix agent involvement, no actionable improvements beyond existing issues.

Timeline

1. 2026-06-18: PR fix(deps): update module github.com/onsi/ginkgo to v2 #298 opened by red-hat-konflux-kflux-prd-rh03[bot] to upgrade github.com/onsi/ginkgo from v1.16.5 to v2.32.1 across 20 files (+21/−59 lines).
2. Jun 18 – Aug 12: PR cycled through ~9 close/reopen iterations by the bot. CI (ci/prow/validate) failed consistently. No human review submitted. No /ok-to-test granted by an org member.
3. 2026-08-12: 1st retro dispatched → concluded "no agent workflow to analyze."
4. 2026-08-14: 2nd retro dispatched → concluded "duplicate retro run."
5. 2026-08-18: 3rd retro dispatched → concluded "zero agent involvement."
6. 2026-08-20: 4th retro dispatched (this run). Seven fullsend.yaml shim runs total; only retro was ever dispatched (review, code, triage, fix, prioritize all skipped every time).

Token waste

Four retro runs on a bot PR with zero agent involvement is pure waste. Each retro agent explores the PR, discovers no agent activity, and writes a summary confirming there's nothing to analyze. The 2nd, 3rd, and 4th runs add no new information.

Existing issue coverage

All candidate improvements are already extensively covered by open issues in fullsend-ai/fullsend. Key existing issues providing this evidence for:

• Skip retro for bot PRs with no agent involvement: #3226, #5295, #5943, #4006, and 8+ more.
• Retro dedup / circuit breaker for repeated runs on same PR: #4064, #5226, #2401.
• Bot PR filtering at dispatch level: #3221, #4154, #5360.
• Consolidation meta-issue: #5817 — explicitly calls out 22+ overlapping issues across these topics and proposes merging them into a single design issue.

This PR provides additional evidence that the bot-skip and retro-dedup gaps are actively causing waste. No new proposals filed — all improvements are already tracked.

[fullsend-ai-retro]

🤖 Finished Retro · ✅ Success · Started 4:06 PM UTC · Completed 4:10 PM UTC

Commit: 34ffb54 · View workflow run →

Runtime: claude · Model: opus → claude-opus-4-6 · Effort: high · Cost: $0.96

[fullsend-ai-retro]

Retro: PR #298 — 5th retro dispatch on zero-involvement bot PR

This is the 5th retro agent run (runs 31594750643, 31755990355, 32097845281, 32366821404, and now 33321451641) on the same bot-authored dependency update PR. All 4 prior runs reached the same conclusion: zero fullsend agent involvement, no actionable improvements beyond what is already tracked. This run confirms that finding once more.

What happened: PR #298 is a Renovate/MintMaker bot PR upgrading github.com/onsi/ginkgo from v1.16.5 to v2.32.1 across ~20 test files. It cycled through 9 close/reopen iterations over 73 days (Jun 18 – Aug 30, 2026) without ever receiving human /ok-to-test approval or review. No fullsend code, review, triage, or fix agents were dispatched — only the retro agent, triggered on each close event.

No new proposals filed. Every improvement opportunity identified is already covered by existing open issues:

• Bot PR retro skip / zero-agent-involvement skip: Skip retro agent for bot-authored trivial PRs with no agent interaction fullsend-ai/fullsend#3226, #5295, #5943, #6297, and 15+ more — all consolidated under Consolidate 22+ overlapping "skip retro/agent dispatch" issues into a single design issue fullsend-ai/fullsend#5817.
• Retro dedup / per-PR circuit breaker: Add retro per-PR circuit breaker to cap retro dispatches after N successful runs with no new activity fullsend-ai/fullsend#4064, #5226, #2401, #6153, #6154.
• Pre-retro early exit: Add bot-PR early exit to pre-retro.sh to skip retro on zero-involvement bot PRs fullsend-ai/agents#276, chore: bump fullsend agents harness base to v0.38.0 #349, #649.
• This run is itself evidence for #5817 — the 5th identical retro on the same PR demonstrates the consolidation meta-issue's core claim that dispatch-skip logic is fragmented across 22+ overlapping issues and none have been implemented yet.

The core CI checks (lint, test, coverage) now pass on the latest commit. The migration is mechanically straightforward (import path changes only). The PR remains blocked on human approval (/ok-to-test, lgtm, approved labels).

[fullsend-ai-review]

🤖 Finished Review · ✅ Success · Started 4:13 AM UTC · Completed 4:22 AM UTC

Commit: bdcb7af · View workflow run →

Runtime: claude · Model: opus → claude-opus-4-6 · Effort: high · Cost: $2.17

[fullsend-ai-review]

Looks good to me

Previous run

Looks good to me

Previous run (2)

Looks good to me

Previous run (3)

Looks good to me

Previous run (4)

Looks good to me

[fullsend-ai-retro]

🤖 Finished Retro · ✅ Success · Started 12:04 PM UTC · Completed 12:14 PM UTC

Commit: bdcb7af · View workflow run →

Runtime: claude · Model: opus → claude-opus-4-6 · Effort: high · Cost: $1.50

[fullsend-ai-retro]

Retrospective: PR #298 — ginkgo v1→v2 bot dependency update (autoclosed)

Timeline: Renovate/MintMaker bot opened PR #298 on Jun 18, 2026 to migrate github.com/onsi/ginkgo from v1.16.5 to v2.32.1 across 20 files (18 test files updating import paths, plus go.mod, go.sum, and tools.go). Over 2.5 months, the Konflux bot (red-hat-konflux-kflux-prd-rh03[bot]) cycled the PR through 11 close/reopen events. No human ever commented, reviewed, or approved. On Aug 31, the review agent was auto-dispatched via a synchronize trigger, spent $2.17 (Opus, high effort), and correctly approved the straightforward import path migration. The PR was auto-closed Sep 1 without merge.

Retro dispatch waste (evidence for existing issues):

• Five retro runs (Aug 12, Aug 14, Aug 18, Aug 20, Aug 30) all concluded with identical findings: zero fullsend agent involvement, no proposals. Estimated ~$5 aggregate waste. This is direct evidence for fullsend-ai/fullsend#5943 — the unmerged-PR-with-zero-agent-involvement skip logic would have prevented all five runs.
• fullsend-ai/fullsend#6297 also relevant — it documents a similar pattern of duplicate retro runs within short time windows.

Bot PR lifecycle (evidence for existing issue): The 2.5-month close/reopen cycle with zero human engagement confirms the pattern described in openshift/ocm-agent-operator#342 — Renovate creating major-version Go module PRs that can never be merged without code migration, cycling indefinitely.

Review quality: The review agent correctly identified the changes as a safe import path migration. No missed issues. The repo's CLAUDE.md explicitly deprecates ginkgo v1 in favor of v2, so the migration aligns with documented preferences.

Autonomy readiness: Not assessable — no human review occurred to compare against the agent's approval.

Agents repo: fullsend-ai/agents@v0.37.0 (commit dc7c805632eb).

Proposals skipped (target repo not allowed)

File manually or update create_issues.allow_targets in config.yaml:

• Skip review dispatch for stale bot PRs with no human engagement (fullsend-ai/agents)

[fullsend-ai-review]

🤖 Review · ❌ Terminated · Started 4:10 AM UTC · Ended 4:21 AM UTC

Commit: 5608aab · View workflow run →

[openshift-ci]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: red-hat-konflux-kflux-prd-rh03[bot]
Once this PR has been reviewed and has the lgtm label, please assign typeid for approval. For more information see the Code Review Process.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[openshift-ci]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: red-hat-konflux-kflux-prd-rh03[bot]
Once this PR has been reviewed and has the lgtm label, please assign typeid for approval. For more information see the Code Review Process.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[fullsend-ai-review]

🤖 Finished Review · ✅ Success · Started 4:10 AM UTC · Completed 4:21 AM UTC

Commit: 5608aab · View workflow run →

Runtime: claude · Model: opus → claude-opus-4-6 · Effort: high · Cost: $2.26

[fullsend-ai-retro]

🤖 Finished Retro · ✅ Success · Started 8:05 AM UTC · Completed 8:17 AM UTC

Commit: 5608aab · View workflow run →

Runtime: claude · Model: opus → claude-opus-4-6 · Effort: high · Cost: $4.33

[fullsend-ai-retro]

Retro: PR #298 — Bot dependency update (ginkgo v1→v2)

PR #298 was a MintMaker bot-generated dependency update migrating github.com/onsi/ginkgo from v1.16.5 to v2.32.1. The changes were purely mechanical (import path swaps across 18 test files, go.mod/go.sum cleanup). The PR was open for 2.5 months (Jun 18 – Sep 3, 2026), cycled through ~9 close/reopen events by the bot, and was ultimately auto-closed without merging. No human ever reviewed or commented on it.

Workflow runs: 12 total, ~$10+ estimated cost

Agent	Runs	Outcome	Est. Cost
Retro	7	All concluded "zero agent involvement"	~$5+
Review	2	Both APPROVED (same commit)	~$4.43
No-op (all skipped)	3	Triggered by review events	~$0

Key findings

All findings corroborate existing open issues. No new proposals are warranted.

1. Retro agent ran 7 times on a zero-involvement bot PR. Each run concluded there was nothing to analyze, producing identical "no proposals" output. This PR provides additional evidence for the cluster of existing issues addressing this pattern: fullsend-ai/fullsend#3226, #4006, #5295, #5943 (skip retro on bot/zero-involvement PRs), fullsend-ai/fullsend#4064, #3034 (per-PR circuit breaker), fullsend-ai/agents#276, #349 (pre-retro early exit), and the meta-consolidation issue fullsend-ai/fullsend#5817.

2. Review agent approved without checking CI status. The review agent ran gh pr diff analysis but never ran gh pr checks or queried CI status. It approved and applied ready-for-merge on a PR where CI was not passing cleanly. This corroborates fullsend-ai/fullsend#1424 and fullsend-ai/agents#833.

3. ready-for-merge label applied despite manual-review-required. The post-review script applied ready-for-merge on approval, even though the PR carried a manual-review-required label and lacked OWNERS approval. This corroborates fullsend-ai/fullsend#5138 and fullsend-ai/agents#501.

4. Review sub-agents failed silently due to model unavailability. The style-conventions and risk-assessment sub-agents failed on both review runs because claude-sonnet-4-5@20250929 was unavailable on Vertex. The review agent continued with only the Opus correctness check, silently dropping two review dimensions. This corroborates fullsend-ai/agents#285, fullsend-ai/fullsend#1795, #1771, #6922, and fullsend-ai/agents#379, #1116.

5. Duplicate review on same commit. Two review runs (Aug 31 and Sep 2) analyzed the identical commit SHA and reached the same APPROVED verdict. This corroborates fullsend-ai/fullsend#5139 and fullsend-ai/agents#323.

Assessment

This PR is a textbook case of agent resource waste on bot-generated dependency updates. The ~$10+ spent across 12 workflow runs produced zero actionable value — no human ever engaged with the PR, and it was never merged. The existing issue backlog comprehensively covers all five identified gaps. The meta-consolidation issue fullsend-ai/fullsend#5817 correctly identifies that 22+ overlapping skip-dispatch proposals need consolidation into a single decision matrix before implementation can proceed.

[fullsend-ai-review]

🤖 Finished Review · ✅ Success · Started 4:10 AM UTC · Completed 4:22 AM UTC

Commit: 3d80f28 · View workflow run →

Runtime: claude · Model: opus → claude-opus-4-6 · Effort: high · Cost: $2.33

[fullsend-ai-review]

Risk Assessment: moderate (2/5)

Details

Bot-authored Ginkgo v2 dependency update with mechanical test import changes across 20 files but only 84 lines changed; high blast radius and two dependency files drive Tier 1 to 2.0, but excellent test coverage ratio (0.85), zero security-sensitive or protected paths, and Tier 2 showing extremely stable files with near-zero churn confirm moderate risk appropriate for standard review.

Previous run

Risk Assessment: moderate (2/5)

Details

Bot-authored Ginkgo v2 dependency update with mechanical test import changes across 20 files but only 84 lines changed; high blast radius and dependency file signals are offset by excellent test coverage ratio (0.85), zero security-sensitive paths, stable low-churn files, and no regression history, yielding a moderate risk level appropriate for standard review.

Previous run (2)

Risk Assessment: moderate (2/5)

Details

Bot-authored Ginkgo v2 dependency update with mechanical test import changes across 20 files but only 84 lines changed; high blast radius and dependency file signals are offset by excellent test coverage ratio (0.85), zero security-sensitive paths, stable low-churn files, and no regression history, yielding a moderate risk level appropriate for standard review.

[fullsend-ai-retro]

🤖 Finished Retro · ✅ Success · Started 12:04 PM UTC · Completed 12:11 PM UTC

Commit: 3d80f28 · View workflow run →

Runtime: claude · Model: opus → claude-opus-4-6 · Effort: high · Cost: $3.49

[fullsend-ai-retro]

Retro: PR #298 — Ginkgo v1→v2 dependency update (bot-generated)

PR #298 is a MintMaker/Renovate bot-generated dependency update that migrates github.com/onsi/ginkgo imports from v1 to v2 across ~18 test files plus go.mod/go.sum cleanup. The PR was open for 2.5 months (Jun 18 – Sep 5, 2026), went through 9 close/reopen cycles driven by the bot rebasing, and was ultimately closed without merge. No human ever reviewed, commented on, or approved this PR.

The fullsend agent system dispatched 11 workflow runs on this PR:

• 8 retro runs — all concluded "zero agent involvement, no proposals"
• 3 review runs — all approved with "Looks good to me" (two on the same commit SHA)
• Estimated total cost: ~$15+ across all runs, producing zero actionable value

No new proposals — all candidate improvements already tracked

Every improvement pattern identified in this retro is already covered by existing open issues. Rather than filing duplicates, here is the evidence this retro adds to each:

• Retro dispatched 8× on zero-involvement bot PR — corroborates fullsend#5817 (consolidation meta-issue for 22+ overlapping skip-retro proposals), fullsend#4064 (per-PR circuit breaker), agents#276 and agents#349 (pre-retro early exit). PR fix(deps): update module github.com/onsi/ginkgo to v2 #298 is among the strongest evidence cases — 8 identical retro runs on one PR, each costing $1.50–$4.33, all reaching the same "nothing to analyze" conclusion.

• Review agent approved 3× including duplicate on same commit — corroborates fullsend#1452 (dedup review dispatches for same HEAD SHA). The Aug 31 and Sep 2 review runs analyzed the identical commit 3d80f28 and both approved.

• Review agent approved without checking CI status — corroborates fullsend#1424 and agents#833. The review agent applied approval while CI was failing.

• ready-for-merge label applied despite manual-review-required — corroborates fullsend#5138 and agents#501. The post-review script applied the merge label without checking for conflicting labels.

• Review sub-agents failed silently due to model unavailability on Vertex — corroborates agents#285 and agents#1116. The style-conventions and risk-assessment sub-agents were silently dropped, leaving only the Opus correctness check.

• Renovate close/reopen cycle generating repeated dispatches — corroborates fullsend#5226 and fullsend#4980. The 9 close/reopen cycles over 2.5 months each triggered new workflow dispatches.

• Unmergeable Renovate major-version bump pattern — corroborates ocm-agent-operator#342, which specifically mentions PR fix(deps): update module github.com/onsi/ginkgo to v2 #298.

Workflow quality assessment

This PR represents a worst-case scenario for agent dispatch efficiency: a mechanically simple bot PR with no human engagement generated $15+ in agent compute across 11 workflow runs that produced no actionable outcomes. The root causes — missing bot-PR skip logic, missing retro circuit breaker, missing review dedup — are all well-documented in existing issues. Resolution of fullsend#5817 (the consolidation meta-issue) would address the majority of waste seen here.

[fullsend-ai-review]

🤖 Finished Review · ✅ Success · Started 4:10 AM UTC · Completed 4:20 AM UTC

Commit: d0103ad · View workflow run →

Runtime: claude · Model: opus → claude-opus-4-6 · Effort: high · Cost: $1.95

[fullsend-ai-retro]

🤖 Finished Retro · ✅ Success · Started 4:03 PM UTC · Completed 4:15 PM UTC

Commit: d0103ad · View workflow run →

Runtime: claude · Model: opus → claude-opus-4-6 · Effort: high · Cost: $4.01

[fullsend-ai-retro]

Retro: PR #298 — ginkgo v1→v2 dependency migration

PR: openshift/ocm-agent-operator#298 — fix(deps): update module github.com/onsi/ginkgo to v2

What happened

A MintMaker/Renovate bot opened this PR on Jun 18 to migrate github.com/onsi/ginkgo from v1 to v2 across 20 files (18 test file import path changes + go.mod/go.sum/tools.go). The change is purely mechanical — every modified file replaces "github.com/onsi/ginkgo" with "github.com/onsi/ginkgo/v2". Core CI checks (lint, test, validate, images, build) all pass.

The PR was open for 81 days (Jun 18 – Sep 7) without any human engagement. It was blocked on /ok-to-test authorization from an openshift org member and required lgtm/approved labels from human approvers (clcollins, jharrington22, rafael-azevedo). The bot rebased/force-pushed at least 9 times, generating close/reopen cycles.

Agent activity:

• 8 completed retro runs (Aug 12 – Sep 5, estimated ~$16+ total). Only retros 1, 6, and 7 produced genuinely new insights; retros 2–5 and 8 repeated the same findings.
• 4 review runs (Aug 31 – Sep 7, ~$8.71 total). All approved with "Looks good to me" — technically correct for a mechanical import change, but the approvals carry no weight in Prow's merge process.
• Estimated total agent cost: ~$22+ on a PR that was never merged and never received human attention.

Autonomy readiness

Inconclusive. No human reviewed this PR, so the agent-vs-human delta cannot be computed. The review agent correctly identified this as a safe mechanical change (risk: moderate 2/5, all modified lines covered by tests), but without human review to compare against, this provides no evidence for or against expanding agent autonomy.

Evidence for existing open issues

This PR is a strong evidence case for several open issues that, if resolved, would prevent the observed waste:

• Consolidate 22+ overlapping "skip retro/agent dispatch" issues into a single design issue fullsend-ai/fullsend#5817 (meta-issue consolidating 22+ skip-retro/dispatch proposals) — PR fix(deps): update module github.com/onsi/ginkgo to v2 #298 exemplifies the core problem: 8 redundant retro runs and 4 redundant review runs on a bot PR with zero human engagement.
• Add retro per-PR circuit breaker to cap retro dispatches after N successful runs with no new activity fullsend-ai/fullsend#4064 (per-PR circuit breaker for retro) — 5 of 8 retro runs produced no new insights beyond what retro 1 already found.
• Consider skipping or minimizing review agent runs on bot-authored dependency update PRs fullsend-ai/fullsend#4796 (minimize review on bot dependency PRs) — 4 review runs on identical mechanical content, each costing ~$2.
• Add bot-PR early exit to pre-retro.sh to skip retro on zero-involvement bot PRs fullsend-ai/agents#276 / chore: bump fullsend agents harness base to v0.38.0 #349 (pre-retro early exit for zero-involvement / bot PRs) — the first 5 retros ran on a PR with zero prior fullsend agent involvement.
• Prioritize same-commit review dedup — PR #944 shows 10x waste on a single SHA fullsend-ai/fullsend#5139 / agents#323 (same-commit review dedup) — review ran on semantically-identical content after bot rebases.
• Review agent should check PR CI status before finalizing verdict fullsend-ai/agents#833 (review should check CI status) — review approved while 3 CI checks were failing (coverage, osd-gcp-e2e, rosa-sts-e2e).
• Suppress Renovate major-version bumps for indirect Go module dependencies #342 (suppress Renovate major-version bumps) — this PR is an example of a Renovate-generated PR that cycles indefinitely without human attention.

Proposals skipped (target repo not allowed)

File manually or update create_issues.allow_targets in config.yaml:

• Retro post-script should use sticky/updateable PR comments instead of appending new ones (fullsend-ai/agents)

[fullsend-ai-review]

🤖 Finished Review · ✅ Success · Started 4:15 AM UTC · Completed 4:26 AM UTC

Commit: 2518704 · View workflow run →

Runtime: claude · Model: opus → claude-opus-4-6 · Effort: high · Cost: $2.55

[openshift-ci]

@red-hat-konflux-kflux-prd-rh03[bot]: all tests passed!

Full PR test history. Your PR dashboard.

Details

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. I understand the commands that are listed here.