Repository navigation
ci: restore hide-vt's prepared inputs in every job that builds it - #2632
Merged
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #2631
Summary
Problem
On 2026-11-18 codeberg.org stopped resolving, and every pull request's
verifyfailed at its first build: five jobs could not fetchtranslate-c, one of hide-vt's pinned Zig packages (#2629, run 44276157654, two attempts).The prepared inputs are saved on main and restored by
.github/actions/vt-cache, but eight jobs that build hide-vt never restored them and downloaded every package on every run, and Linux arm64 had no saved entry at all.Before → After
hide-vt/build-inputs.jsonchangesNo screenshot: CI configuration and a script test only, no screen changes.
web-e2e(build, e2e),agent-journey(fixture, fake),agent-trust-contract,desktop-e2e,perfand the device helpers' build.test_vt_cache.pynow fails any job whoseverify-cargo.shcall (or the device helpers' build) builds hide-vt without the action, so a new job cannot drop it again.Decide
Not verified
remote mailbox arm64,helper (linux-aarch64)) still download until main saves the arm64 entry, so they fail on this pull request while codeberg.org is down; on the first run the other jobs restored main's entries and passed.remote-mailbox-armis in main's plan.scripts/tests/test_agent_live_protection.py::test_linux_reaping_adopted_children_leaves_the_root_to_its_own_waitfails on macOS on main too; it is not touched here.Evidence
python3 -m pytest scripts/testslocally, 1057 passed; the one other failure is the Linux-only test above, failing on main as well.~/.cache/hide-build(or%LOCALAPPDATA%\hide-build), not the$RUNNER_TEMP/hide-buildthe action restores; the arm64 job loggedCache not found for input keys: vt-inputs-v1-Linux-ARM64-....Tests
scripts/tests/test_vt_cache.py, reading the workflow files).perf.ymlwithout the step, and the test failed namingperf.yml:measure.Risky files
.github/workflows/remote-mailbox.yml- a new saver; its condition saves only on main and only on an ARM64 runner, so it cannot race the vt lane's Linux x64 entry..github/workflows/agent-journey.yml- the fake-model job's cap rose from 181 to 186 minutes, the sum of its step caps with the 5-minute restore.Verification record
7f624a962678.python3 -m pytest -q scripts/tests(1057 passed, 1 unrelated Linux-only failure),test_vt_cache.pyandtest_agent_journey_workflow.py7 passed.