A quick fix#37317
---
The current behavior for forks when an organization or repository is
changed to private differs from GitHub.
On GitHub, when a parent repository becomes private, the fork
relationship is removed, which keeps the behavior simple and avoids
visibility conflicts.
I think we need a similar solution to handle cases where the parent
repository becomes private while a fork remains public and the fork
relationship is still preserved.
---------
Signed-off-by: Lunny Xiao <xiaolunwen@gmail.com>
Co-authored-by: silverwind <me@silverwind.io>
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
Co-authored-by: Giteabot <teabot@gitea.io>
Fail the vite build on any rolldown warnings when `NODE_ENV=test` is
set. This gate is set on the CI `make frontend` steps (compliance and
e2e workflows) and on the local `make test-e2e` target, so warnings fail
the build both in CI and when running e2e tests locally. Regular `make
frontend` / production builds are unaffected.
Example output:
```
[plugin test-warning-injector] first synthetic warning
[plugin test-warning-injector] second synthetic warning
transforming...✗ Build failed in 14ms
error during build:
Build failed with 1 error:
[plugin fail-on-warnings]
Error: 2 warnings present
at PluginContextImpl.buildEnd (vite.config.ts:50:13)
...
```
---
This PR was written with the help of Claude Opus 4.7
---------
Signed-off-by: silverwind <me@silverwind.io>
Co-authored-by: Claude (Opus 4.7) <noreply@anthropic.com>
Co-authored-by: Nicolas <bircni@icloud.com>
Follow up #37327. See the comments.
* Root problem: the design of OAuth2 providers is a mess, the display
name is used as provider's name and used in the URL directly
* The regressions:
* When trying to fix https://github.com/go-gitea/gitea/issues/36409 , it
introduced inconsistent URL escaping for the "path" part.
* This fix: always use "path escaping" for the path part, add more tests
to cover all escaping cases.
Now, frontend "pathEscape" and "pathEscapeSegments" generate exactly the
same result as backend.
The link to authentication sources is now escaped with the QueryEscape.
This commit fixes that by unescaping the provider name in the URL.
---------
Signed-off-by: prettysunflower <me@prettysunflower.moe>
Signed-off-by: wxiaoguang <wxiaoguang@gmail.com>
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
## Problem
Workflow-level concurrency groups were evaluated — and jobs were parsed
— before the run was persisted, so `run.ID` was `0` and `github.run_id`
in the expression context resolved to an empty string. Expressions like:
```yaml
concurrency:
group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
cancel-in-progress: true
```
collapsed to `<workflow>-` on every push event (`head_ref` is empty on
push), so `cancel-in-progress` cancelled in-progress runs across
**unrelated branches**, not just the current one.
Reproduced on a 1.26 instance:
- push to `master` → `ci` run starts
- push to `feature-branch` → the `master` run gets cancelled
GitHub Actions' documented semantic: on push events `github.run_id` is
unique per run, so the group is unique → no cancellation; on PR events
`github.head_ref` is the source branch → cancellation is per-PR.
## Fix
Insert the run **before** parsing jobs or evaluating workflow-level
concurrency, so `run.ID` is populated in time for every expression that
reads `github.run_id` — not just the concurrency group, but also
`run-name`, job names, and `runs-on`.
`jobparser.Parse` now runs inside the `InsertRun` transaction, after
`db.Insert(ctx, run)`. Workflow-level concurrency evaluation runs next
and only mutates `run` in memory. All concurrency-derived fields
(`raw_concurrency`, `concurrency_group`, `concurrency_cancel`) plus
`status` and `title` are persisted in a single final `UpdateRun` at
end-of-transaction — one `INSERT` + one `UPDATE` per run in both the
concurrency and non-concurrency paths (matches pre-branch parity, one
fewer `UpdateRepoRunsNumbers` `COUNT` than the interim state).
`GenerateGiteaContext` now sets `run_id` from `run.ID` unconditionally;
every caller passes a persisted run.
**Verification**: tested end-to-end on a 1.26 deployment. Before the
patch, two successive `ci` pushes (one to master, one to a feature
branch) cross-cancelled each other. After the patch, the same pushes —
in both orders (master→branch, branch→master) — run to completion
simultaneously across 15+ runs with zero cancellations.
**Regression tests** in `services/actions/context_test.go`:
- `TestEvaluateRunConcurrency_RunIDFallback` — unit check that
`EvaluateRunConcurrencyFillModel` resolves `github.run_id` from
`run.ID`.
- `TestPrepareRunAndInsert_ExpressionsSeeRunID` — full-flow check: calls
`PrepareRunAndInsert` with `${{ github.run_id }}` in both `run-name` and
the concurrency group, then asserts the persisted `Title`,
`ConcurrencyGroup`, and `RawConcurrency` contain / survive the run's ID.
Re-ordering `db.Insert` relative to either parse or concurrency eval
fails this test.
## Relation to #37119
[#37119](https://github.com/go-gitea/gitea/pull/37119) also moves
concurrency evaluation into `InsertRun` but keeps it **before**
`db.Insert`, then tries to populate `run_id` only when `run.ID > 0` —
which is still `0` at that call site, so the cross-branch leak would
survive that PR as written. This PR fixes the ordering so that `run.ID`
is actually populated at eval time, and broadens it to cover parse-time
expression interpolation too.
Co-authored-by: Claude (Opus 4.7) <noreply@anthropic.com>
Replaces `@silverwind/vue3-calendar-heatmap` with an inlined SVG
implementation. Renders pixel-identically to `main`, drops the
`onMounted` legend viewBox workaround, and uses tippy's
`createSingleton` for the hover tooltip. Adds an e2e test for tooltip
display.
This is a prereq for migrating tippy.js to
[floating-ui](https://github.com/floating-ui/floating-ui) to avoid
having two tooltip libs active.
<img width="861" height="168" alt="image"
src="https://github.com/user-attachments/assets/99343cf6-6e09-42c7-a80d-63dbf33cf56a"
/>
---
This PR was written with the help of Claude Opus 4.7
---------
Co-authored-by: Claude (Opus 4.7) <noreply@anthropic.com>
Co-authored-by: Nicolas <bircni@icloud.com>
Enable full TypeScript `strict` mode and fix issues discovered during
this refactor. Introduced a `errorMessage` helper function to cleanly
extract a error messages from the `unknown` type.
Signed-off-by: silverwind <me@silverwind.io>
Co-authored-by: Claude (claude-opus-4-7) <noreply@anthropic.com>
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
Moves the manifest patching from `closeBundle` to `writeBundle`. Thrown
errors in `writeBundle` work correctly and exit the build.
Signed-off-by: silverwind <me@silverwind.io>
Signed-off-by: wxiaoguang <wxiaoguang@gmail.com>
Co-authored-by: Claude (Opus 4.7) <noreply@anthropic.com>
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
Verified locally with 50 runs, averaging 9 seconds per local test suite
run. Total suite took 15s.
`--with-deps` is needed because webkit's dependencies are not
pre-installed on GHA runners (as opposed to firefox/chrome which are
preinstalled).
---
This PR was written with the help of Claude Opus 4.7
---------
Co-authored-by: Claude (Opus 4.7) <noreply@anthropic.com>
Co-authored-by: Nicolas <bircni@icloud.com>
When generating release notes for v1.26, many pull requests haven't been
given correct labels so that I have to do many manual work. I think this
could be avoid to remove these useless modify labels.
While editing frontend, I found some inconsistencies while testing
transferring repositories:
- No button for accepting/rejecting/cancelling the transfer of an empty
repository.
- The `redirect_to` in `templates/repo/header.tmpl` is useless.
- There's no redirection when there's an error from `handleActionError`
in `routers/web/repo/repo.go`. Therefore, instead of flash message, a
blank page will be displayed.
This pr adds some commits to resolve all these issues.
Update: see the new changes
https://github.com/go-gitea/gitea/pull/37277#issuecomment-4276150232
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
Co-authored-by: Giteabot <teabot@gitea.io>
1. stabilize flaky e2e test from
2f5b5a9e9c
2. increase ci timeout factor to 4 as 3 was not enough
3. add a `e2e` category to files-changed so e2e-test-only changes
trigger ci
---
This PR was written with the help of Claude Opus 4.7
---------
Co-authored-by: Claude (Opus 4.7) <noreply@anthropic.com>
Automated changes by the
[update-flake-lock](https://github.com/DeterminateSystems/update-flake-lock)
GitHub Action.
```
Flake lock file updates:
• Updated input 'nixpkgs':
'github:nixos/nixpkgs/4c1018d' (2026-04-09)
→ 'github:nixos/nixpkgs/4bd9165' (2026-04-14)
```
### Running GitHub Actions on this PR
GitHub Actions will not run workflows on pull requests which are opened
by a GitHub Action.
**To run GitHub Actions workflows on this PR, close and re-open this
pull request.**
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Co-authored-by: Nicolas <bircni@icloud.com>
If a workflow is not in default branch the hooks could not be detected
Fixes#37169
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: Giteabot <teabot@gitea.io>
Fixes GHSA-3xc5-wrhm-f963 (credential exposure on HTTP redirects).
---
This PR was written with the help of Claude Opus 4.6
Co-authored-by: Claude (Opus 4.6) <noreply@anthropic.com>
Introduces a frontend external-render framework that runs renderer
plugins inside an `iframe` (loaded via `srcdoc` to keep the CSP
`sandbox` directive working without origin-related console noise), and
migrates the 3D viewer and OpenAPI/Swagger renderers onto it. PDF and
asciicast paths are refactored to share the same `data-render-name`
mechanism.
Adds e2e coverage for 3D, PDF, asciicast and OpenAPI render paths, plus
a regression for the `RefTypeNameSubURL` double-escape on non-ASCII
branch names.
Signed-off-by: silverwind <me@silverwind.io>
Co-authored-by: Claude (Opus 4.6) <noreply@anthropic.com>
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
Patterns starting with "/" (e.g. /docs/.*\.md) never matched because git
returns relative paths without a leading slash. Strip the leading "/"
before compiling the regex since the ^...$ anchoring already provides
root-relative semantics.
Fixes: #28107
Co-authored-by: silverwind <me@silverwind.io>
Co-authored-by: Giteabot <teabot@gitea.io>
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
Use `golangci-lint fmt` to format code, replacing the previous custom
formatter tool. https://github.com/daixiang0/gci is used to order the
imports.
`make fmt` performs ~13% faster while consuming ~57% less cpu while
formatting for me.
`GOFUMPT_PACKAGE` is gone because it's using the builtin package from
golangci-lint.
Co-authored-by: Claude (claude-opus-4-6) <noreply@anthropic.com>
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
This PR adds an External ID Claim Name configuration field to the OIDC
auth source. When set, Gitea uses the specified JWT claim as the user's
`ExternalID` instead of the default `sub` claim.
This PR fixes the bug when migrating from Azure AD V2 to OIDC. When an
admin migrates the same auth source to OIDC, goth's `openidConnect`
provider defaults to using the `sub` claim as `UserID`. However, Azure
AD's `sub` is a pairwise identifier:
> `sub`: The subject is a pairwise identifier and is unique to an
application ID. If a single user signs into two different apps using two
different client IDs, those apps receive two different values for the
subject claim.
https://learn.microsoft.com/en-us/entra/identity-platform/id-token-claims-reference#payload-claims
As a result, every existing user appears as a new account after
migration.
To fix this issue, Gitea should use `oid` claim for `UserID`.
> `oid`: This ID uniquely identifies the user across applications - two
different applications signing in the same user receives the same value
in the oid claim.
Note: The `oid` claim is not included in Azure AD tokens by default. The
`profile` scope must be added to the Scopes field of the auth source.
