After #416 was merged into main, the milestone PR #380 (next into main) no longer merges: script/test conflicts. #416 carried a main-only version of fixes that next already has in its own form (#344, #414, #315): no -v, a memory cap, and cheaper password hashing in tests. Besides script/test it changed internal/database/password.go, internal/database/password_test.go, internal/database/export_test.go and internal/resetpw/resetpw_test.go, which merge without a textual conflict but may now carry a second mechanism next to next's own.
Definition of done:
main is merged into next through a PR into next whose branch holds a real merge commit of origin/main (not a squash, which would leave the conflict in place), merged by the manager with Gitea's merge style.
Every conflict and every auto-merged file is resolved to what next already does where next covers the same fix, so the tree has one mechanism for each (one test-hashing cost, one script/test), and nothing main fixed is lost.
After it lands, #380 reports mergeable, and make check passes on next.
Model: opus-5-5
After https://git.eeqj.de/sneak/webhooker/pulls/416 was merged into `main`, the milestone PR https://git.eeqj.de/sneak/webhooker/pulls/380 (`next` into `main`) no longer merges: `script/test` conflicts. https://git.eeqj.de/sneak/webhooker/pulls/416 carried a `main`-only version of fixes that `next` already has in its own form (https://git.eeqj.de/sneak/webhooker/issues/344, https://git.eeqj.de/sneak/webhooker/issues/414, https://git.eeqj.de/sneak/webhooker/issues/315): no `-v`, a memory cap, and cheaper password hashing in tests. Besides `script/test` it changed `internal/database/password.go`, `internal/database/password_test.go`, `internal/database/export_test.go` and `internal/resetpw/resetpw_test.go`, which merge without a textual conflict but may now carry a second mechanism next to `next`'s own.
Definition of done:
- `main` is merged into `next` through a PR into `next` whose branch holds a real merge commit of `origin/main` (not a squash, which would leave the conflict in place), merged by the manager with Gitea's merge style.
- Every conflict and every auto-merged file is resolved to what `next` already does where `next` covers the same fix, so the tree has one mechanism for each (one test-hashing cost, one `script/test`), and nothing `main` fixed is lost.
- After it lands, https://git.eeqj.de/sneak/webhooker/pulls/380 reports mergeable, and `make check` passes on `next`.
Model: opus-5-5
clawbot
self-assigned this 2026-10-03 14:16:37 +02:00
Plan: a worker branches off next, merges origin/main with a real merge commit, resolves script/test and the four auto-merged files to next's own version wherever next covers the same fix (checking nothing main fixed is lost), runs make check, and opens a PR into next. An independent reviewer checks the resolution file by file and that next merged with it into main gives no conflict. I merge it with Gitea's merge style, not a squash, so main becomes an ancestor of next.
Model: opus-5-5
Plan: a worker branches off `next`, merges `origin/main` with a real merge commit, resolves `script/test` and the four auto-merged files to `next`'s own version wherever `next` covers the same fix (checking nothing `main` fixed is lost), runs `make check`, and opens a PR into `next`. An independent reviewer checks the resolution file by file and that `next` merged with it into `main` gives no conflict. I merge it with Gitea's merge style, not a squash, so `main` becomes an ancestor of `next`.
Model: opus-5-5
#498 merges origin/main into next with a real merge commit, with main as its second parent. The merged tree is identical to next's.
script/test: conflicted; kept next's, which already has main's no -v and -p 4 -parallel 8 memory cap, plus coverage and a -v rerun of failed tests.
internal/database/password.go: merged on its own; main's test-hashing cost is line for line next's, so one mechanism remains.
internal/database/password_test.go: merged on its own; TestHashPassword_ShippedParameters is the same on both sides.
internal/database/export_test.go: merged on its own; main's HashAtShippedCostForTest is already on next, which also keeps ExportReapBatchSize.
internal/resetpw/resetpw_test.go: merged on its own; main's one comment change is already on next.
Nothing main changed is missing from next. git merge-tree of the branch into origin/main reports no conflict.
Judgement call: the PR must be merged with Gitea's merge-commit style. A squash or rebase would drop main as a parent and bring the conflict back.
Model: opus-5-5
https://git.eeqj.de/sneak/webhooker/pulls/498 merges `origin/main` into `next` with a real merge commit, with `main` as its second parent. The merged tree is identical to `next`'s.
- `script/test`: conflicted; kept `next`'s, which already has `main`'s no `-v` and `-p 4 -parallel 8` memory cap, plus coverage and a `-v` rerun of failed tests.
- `internal/database/password.go`: merged on its own; `main`'s test-hashing cost is line for line `next`'s, so one mechanism remains.
- `internal/database/password_test.go`: merged on its own; `TestHashPassword_ShippedParameters` is the same on both sides.
- `internal/database/export_test.go`: merged on its own; `main`'s `HashAtShippedCostForTest` is already on `next`, which also keeps `ExportReapBatchSize`.
- `internal/resetpw/resetpw_test.go`: merged on its own; `main`'s one comment change is already on `next`.
Nothing `main` changed is missing from `next`. `git merge-tree` of the branch into `origin/main` reports no conflict.
Judgement call: the PR must be merged with Gitea's merge-commit style. A squash or rebase would drop `main` as a parent and bring the conflict back.
Model: opus-5-5
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
After #416 was merged into
main, the milestone PR #380 (nextintomain) no longer merges:script/testconflicts. #416 carried amain-only version of fixes thatnextalready has in its own form (#344, #414, #315): no-v, a memory cap, and cheaper password hashing in tests. Besidesscript/testit changedinternal/database/password.go,internal/database/password_test.go,internal/database/export_test.goandinternal/resetpw/resetpw_test.go, which merge without a textual conflict but may now carry a second mechanism next tonext's own.Definition of done:
mainis merged intonextthrough a PR intonextwhose branch holds a real merge commit oforigin/main(not a squash, which would leave the conflict in place), merged by the manager with Gitea's merge style.nextalready does wherenextcovers the same fix, so the tree has one mechanism for each (one test-hashing cost, onescript/test), and nothingmainfixed is lost.make checkpasses onnext.Model: opus-5-5
Plan: a worker branches off
next, mergesorigin/mainwith a real merge commit, resolvesscript/testand the four auto-merged files tonext's own version wherevernextcovers the same fix (checking nothingmainfixed is lost), runsmake check, and opens a PR intonext. An independent reviewer checks the resolution file by file and thatnextmerged with it intomaingives no conflict. I merge it with Gitea's merge style, not a squash, somainbecomes an ancestor ofnext.Model: opus-5-5
#498 merges
origin/mainintonextwith a real merge commit, withmainas its second parent. The merged tree is identical tonext's.script/test: conflicted; keptnext's, which already hasmain's no-vand-p 4 -parallel 8memory cap, plus coverage and a-vrerun of failed tests.internal/database/password.go: merged on its own;main's test-hashing cost is line for linenext's, so one mechanism remains.internal/database/password_test.go: merged on its own;TestHashPassword_ShippedParametersis the same on both sides.internal/database/export_test.go: merged on its own;main'sHashAtShippedCostForTestis already onnext, which also keepsExportReapBatchSize.internal/resetpw/resetpw_test.go: merged on its own;main's one comment change is already onnext.Nothing
mainchanged is missing fromnext.git merge-treeof the branch intoorigin/mainreports no conflict.Judgement call: the PR must be merged with Gitea's merge-commit style. A squash or rebase would drop
mainas a parent and bring the conflict back.Model: opus-5-5