aboutsummaryrefslogtreecommitdiffstats
path: root/.github/workflows
Commit message (Collapse)AuthorAgeFilesLines
* workflows: gemini cli ignores pull_request_review eventsTaras Madan2025-08-201-3/+5
| | | | | It is needed to reduce "skipped" workflow notifications. Eventually it could be reviewed.
* workflow: don't trigger gemini pr review by pull_requestTaras Madan2025-08-191-4/+5
| | | | Workers created by the forked repository triggers don't have access to github secrets.
* workflows: enable pr review with gemini cliTaras Madan2025-08-191-0/+474
|
* ci.yml: enable self-hosted-x86Taras Madan2025-04-041-8/+4
|
* ci.yaml: use github org runnersTaras Madan2025-04-021-4/+4
|
* ci.yml: fix cachingTaras Madan2025-01-301-18/+43
| | | | | | | | Aux requires more dependencies than other jobs. It also doesn't fail as often as others. Because of these two points aux will be the cache updating job. Other jobs use actions/cache/restore and are the cache readers. Old and gvisor are the special cases that are not gcr.io/syzkaller/env based and require additional GOMODCACHE param.
* .github/workflows: use upload-artifacts v4 in oss-fuzz workflowTaras Madan2025-01-091-1/+1
|
* vm/gvisor: add gvisor presubmit testAndrei Vagin2024-07-081-0/+20
| | | | | | | Download the latest gvisor release and run the syz-manager smoke-test suite. Signed-off-by: Andrei Vagin <avagin@google.com>
* ci.yml: move codecov.yml and point it from configTaras Madan2024-05-101-2/+4
|
* ci.yml: verbose codecov reportsTaras Madan2024-05-101-23/+25
|
* ci.yml: add codecov token and switch to v4Taras Madan2024-05-101-2/+4
|
* ci.yml: cancel old workflowsTaras Madan2024-05-081-0/+5
|
* .github/workflows/ci.yml: workaround codecov limitation to detect environmentTaras Madan2024-03-071-0/+2
|
* .github/workflows/ci.yml: fix cache locationTaras Madan2024-03-041-2/+2
|
* .github/workflows/ci.yml: fix cache pathsTaras Madan2024-03-011-2/+2
| | | | Old path works for ./tools/syz-env container and doesn't work for the new configuration.
* ./github/workflows/ci.yml: run race_dashboard in parallelTaras Madan2024-02-281-0/+17
|
* ci.yml: let's dumb the TERMTaras Madan2024-02-261-0/+6
| | | | It is to avoid "tput: No value for $TERM and no -T specified" in logs.
* run.sh, Makefile: ask golangci-lint to preformat outputTaras Madan2024-02-261-1/+0
|
* ci.yml: run on arc-runner-setTaras Madan2024-02-262-50/+83
|
* ci.yml: change runners to org sharedTaras Madan2024-02-221-3/+3
|
* oss-fuzz.yml: fix compilation errorjonathanmetzman2023-09-271-4/+12
| | | | | | | Fix compilation error and make other improvements: * Only upload crash on run fail, not build fail * Enable SARIF notification. Fixes: https://github.com/google/oss-fuzz/issues/10989
* ci.yml: cache build artefacts where it is neededTaras Madan2023-08-231-27/+18
| | | | | | | | | | | | | | If some job is not a build bottleneck - let's remove caching. The main focus points then are "build" and "race" then. To cache test results independently let's use 2 different cache prefixes. See https://markphelps.me/posts/speed-up-your-go-builds-with-actions-cache/. The idea is to always generate cash miss using the second precision key. After the miss we download latest uploaded one using "restore-keys". Potential improvements: 1. https://github.com/golang/go/issues/23565 asks to cache -coverprofile results since 2018. 2. https://github.com/golang/go/issues/61608 to cache -race results.
* all: remove closestale.yamlAleksandr Nogikh2023-08-021-19/+0
| | | | | | | This reverts commit 71f8554dae7330f8cfa1b22775ebafdff2315623. The bot has begun to actually close stale issues and it does not seem to be the behavior we want. Let's disable it for now.
* depsreview.yaml: bump actionsTaras Madan2023-07-261-2/+2
|
* ci.yml: bump actions and covecovTaras Madan2023-07-261-14/+14
|
* closestale.yaml: notify about stale PRsTaras Madan2023-07-261-0/+19
|
* ci.yml: pull syz-env separatelyTaras Madan2023-07-251-0/+6
|
* ci.yml: run any-env tests on k8s-envTaras Madan2023-07-251-3/+3
|
* ci.yml: run on any-env not self-runnerTaras Madan2023-07-241-3/+3
|
* all: restructure testsAleksandr Nogikh2023-07-121-1/+1
| | | | | Let presubmit_dashboard run dashboard tests and presubmit_build run all others.
* all: remove mentions of big-envAleksandr Nogikh2023-07-121-2/+2
| | | | Use just env instead of that.
* .github: add a 15 minute timeout for the `dashboard` actionAleksandr Nogikh2023-01-171-1/+1
| | | | | | | | The dashboard testing infrastructure is quite fragile and sometimes just hangs in case of errors. The action then just hangs for 6 hours with no debugging output. Add a 15 minute timeout, this should be enough for this test.
* .github: decrease the OSS fuzzing time (#3336)Aleksandr Nogikh2022-09-051-1/+1
| | | | 10 minutes looks like a too big number, especially given that other CI actions finish much faster.
* Makefile: split presubmit_arch_other targetDmitry Vyukov2022-08-101-1/+1
| | | | | presubmit_arch_other is now the slowest presubmit step taking 7-8 mins. Split it per OS.
* .github/workflows/ci.yml: stop to use github instances for heavy tests (#3274)Taras Madan2022-08-021-54/+0
|
* .github/workflows/ci.yml: run race step on the self-hosted instance (#3271)Taras Madan2022-07-271-0/+14
| | | Race step takes ~8 minutes and is the 3rd longest after "build" and "dashboard".
* .github/workflows/depsreview: new check (#3156)Naveen2022-07-271-0/+19
| | | | | Dependency Review GitHub Action in your repository to enforce dependency Signed-off-by: naveen <172697+naveensrinivasan@users.noreply.github.com>
* .github/workflows/ci.yml: run dashboard step on the self-hosted instance (#3268)Taras Madan2022-07-261-0/+19
| | | +use v3 codecov instead
* .github/workflows/ci.yml: add self-hosted builder (#3265)Taras Madan2022-07-261-0/+21
| | | | Add self-hosted builder. Use covecov-action v3.
* .github/workflows: pined actions by SHAnaveensrinivasan2022-03-241-14/+17
| | | | - Pinned actions by SHA https://github.com/ossf/scorecard/blob/main/docs/checks.md#pinned-dependencies
* .github/workflows: split work across more jobsDmitry Vyukov2021-12-211-4/+21
| | | | | | | | CI become quite slow, especially the 'arch' job. Split it into 3 separate jobs. Also split main build/lint/test from all of the aux checks. Update #2886
* .github/workflows: re-enable oss-fuzzDmitry Vyukov2021-03-251-5/+1
| | | | | It's claimed to be fixed: https://github.com/google/oss-fuzz/issues/3670
* .github/workflows: remove fuzzit.devDmitry Vyukov2020-12-251-24/+0
| | | | | | | | | | | | | | All fuzzit.dev jobs become broken recently, see e.g.: https://github.com/google/syzkaller/pull/2351/checks?check_run_id=1601794199 ... {"status":"Status: Downloaded newer image for gcr.io/fuzzit-public/stretch-llvm8:64bdedf"} Downloading fuzzit cli/agent... Error: Process completed with exit code 1. https://github.com/fuzzitdev/fuzzit is archived now and the README says: "the standalone service will soon be deperecated". So I guess it's time to remove fuzzit.dev integration.
* .github/workflows: switch fuzzit to syz-old-envDmitry Vyukov2020-10-271-2/+5
| | | | | | Use syz-old-env because it contains Go 1.14. syz-env contains Go 1.15 and go-fuzz is broken with Go 1.15: https://github.com/dvyukov/go-fuzz/issues/294
* tools/docker: add old-envDmitry Vyukov2020-08-221-3/+14
| | | | | | | old-env is based on Ubuntu 16.04 and allows to test executor build on older distributions. Fixes #2055
* tools/check-commits.sh: fix commit range detectionDmitry Vyukov2020-08-051-1/+2
| | | | | | | | | | | | We currently check from github.event.pull_request.base.sha to github.event.pull_request.head.sha, but they may be in different branches if the PR commits are branched not from the latest master HEAD (at the time of PR creation). Then GH will create a merge commit, and the range we try to check is not valid. Check github.event.pull_request.commits commits backwards from github.event.pull_request.head.sha commit.
* .github/workflows: check commit description formatDmitry Vyukov2020-07-311-0/+5
| | | | | Check PR commit subjects for the standard format we use "pkg/name: description of the change". Check PR commit bodies for at most 120 characters.
* .github/workflows: support errors from golangci-lintDmitry Vyukov2020-06-051-0/+1
| | | | | | | | | If there are compilation errors in tests, golangci-lint produces then in the following format: level=warning msg="[runner] Can't run linter unused: buildssa: analysis skipped: errors in package: [/syzkaller/gopath/src/github.com/google/syzkaller/syz-fuzzer/mab_reward_test.go:26:35: cannot use rawCov (variable of type []float64) as float64 value in argument to status.Reward.RawAllTasks.Update /syzkaller/gopath/src/github.com/google/syzkaller/syz-fuzzer/mab_reward_test.go:26:43: cannot use rawTime (variable of type []float64) as float64 value in argument to status.Reward.RawAllTasks.Update /syzkaller/gopath/src/github.com/google/syzkaller/syz-fuzzer/mab_reward_test.go:27:38: cannot use totalReward (variable of type []float64) as float64 value in argument to status.Reward.RewardAllTasks.Update /syzkaller/gopath/src/github.com/google/syzkaller/syz-fuzzer/mab_reward_test.go:33:2: reward declared but not used]" Transform this to github format.
* .github/workflows: support errors without :columnDmitry Vyukov2020-06-051-1/+1
| | | | | | | | Also transform: file.go:line: error message to the github format.
* tools: add syz-big-envDmitry Vyukov2020-06-022-6/+43
| | | | | | | | On top of syz-env it provides akaros/fuchsia/netbsd toolchains and gcloud sdk. With this it's possible to run dashboard/app tests on CI and locally and test executor build and pkg/{csource,cover} for these OSes. Update #1765
generated by ggit (джигит) mrf-deployment (git 2.46.0) at 2026-03-11 21:07:41 +0000