This fixes three pre-existing issues in the visual test runner that were
never caught because visual tests aren't run in CI.
## Fixes
**1. Re-entrant read panic in sidebar creation (3 places)**
`Sidebar::new` reads the `MultiWorkspace` entity, so creating it inside
a `multi_workspace.update()` closure causes a `cannot read while being
updated` panic. Fixed by using `update_window` with
`root_view.downcast()` to get the handle without holding a mutable
borrow.
**2. Tooltip entity leak in breakpoint_hover test**
The tooltip system has an Rc cycle (`ActiveTooltip::Visible` holds a
`check_visible_and_update` closure that captures the
`Rc<RefCell<Option<ActiveTooltip>>>`). This cycle is normally broken
when the tooltip is dismissed, but the test closed the window while the
tooltip was still visible, causing a panic from the leak detector. Fixed
by simulating a click before cleanup to dismiss the tooltip.
**3. Visual tests not in CI**
Added a `cargo_run_visual_tests` step to the macOS test job via xtask so
these tests run on every PR and regressions are caught early.
Release Notes:
- N/A
The workflow run at
https://github.com/zed-industries/zed/actions/runs/23557683707 succeeded
but threw some warnings for a rather-soon Node.js 20 deprecation (June
2nd).
Hence, this PR updates in that context mentioned workflows to newer
versions from which on the actions will use Node.js 24.
Namely, this updates
- `actions/checkout`
- `actions/create-github-app-token` and
- `peter-evans/create-pull-request`
to their latest version which includes said updates. As for their most
recent versions, all of these actions just updated their versions to
account for said deprecation.
Release Notes:
- N/A
Closes#51339
This should address issues with too new libstdc++.so on older/more
conservative distros such as RHEL9.
Release Notes:
- Relaxed requirement for libstdc++.so available on Linux distros.
Reverts zed-industries/zed#48607
The problem we've ran into is that CI started reusing already-cached
runs from newer branches (that did not have that change pulled in).
Yet this unveiled another potential issue, which is that relying on
mtime might make it so that an artifact for a different (newer) commit
would be reused by an unrelated run of CI.
The answer is: checksum-based-freshness.
Release Notes:
- N/A
GitHub allows defining a default shell for all jobs on the workflow
level, which we did not use before, yet practically did when it comes
down to our usage of `named::bash`. Since this makes stuff quite
verbose, I decided on using the defaults instead so the workflows become
somewhat easier to audit when reading the generated files.
Powershell steps continue to use Powershell, only the default for bash
scripts was modified.
Release Notes:
- N/A
I think we're not triggering the after-release workflow because of
github's loop detection when you use the default GITHUB_TOKEN
Closes #ISSUE
Release Notes:
- N/A
This reverts #39643, effectively
For the record, @SomeoneToIgnore found it quite cumbersome to scroll
through logs just to see which tests have failed. I kinda see the
argument. At the same time, I wish nextest could do both: it could
aggregate logs of failed tests and then print out the summary.
Release Notes:
- N/A
This starts the work on a workflow that can be invoked in extension CI
to test changes on extension repositories.
Release Notes:
- N/A
---------
Co-authored-by: Agus Zubiaga <agus@zed.dev>
Co-authored-by: Conrad Irwin <conrad.irwin@gmail.com>
This hopefully resolves the lingering test failures on linux,
but also adds some logging just in case this isn't the problem...
Release Notes:
- N/A
---------
Co-authored-by: Ben Kunkle <ben@zed.dev>
One of the reasons we didn't spot that we were missing the telemetry env
vars for the production builds was that nightly (which was working) had
its own set of build steps. This re-uses those and pushes the env vars
down from the workflow to the job.
It also fixes nightly releases to upload all-in-one go so that all
platforms update in sync.
Closes#41655
Release Notes:
- N/A
Finn Evers