CI: revert temporary test-only isolation; MTU fix ruled out

Clean result, isolated from lint's own unrelated github.com flakiness:
post-MTU-fix (Ryuvia/charts#272), `make test` alone hits the exact same
"TLS handshake timeout" fetching envtest-v1.37.0-linux-amd64.tar.gz as
before the fix. Byte-for-byte identical error. The dind sidecar's MTU
mismatch was real (measured 1450 vs 1500 per the other session's report)
but it was not (solely) the cause of this specific failure.

Separately and incidentally: golangci-lint's own custom-gcl build also
does a plain `git clone https://github.com/...` and that is now failing
too (2/2, ~2.5min hang then generic exit 128) where it briefly succeeded
in an earlier pre-fix run -- noted in the comment but not chased further
here; worth someone's attention if it keeps recurring, since it'll block
`lint` regardless of the envtest question.

.gitea/workflows/ci.yaml

@@ -46,11 +46,18 @@ jobs:
   #   - Running this job on the bare runner host instead of in a container (tried and
   #     reverted, same as terdut-server's `chart` job does for get.helm.sh) fails
   #     earlier and differently: "go: command not found" -- the host has no Go.
+  #   - The act-runner dind sidecar's MTU (Ryuvia/charts#272, merged and reconciled
+  #     2026-09-30: explicit `"mtu": 1450` in its daemon.json, matching Flannel's
+  #     VXLAN-reduced pod MTU, since dockerd's bridge networks default to 1500
+  #     unset) -- tested against this exact failure post-fix (isolated `make test`,
+  #     bypassing lint's own unrelated github.com flakiness) and got the identical
+  #     "TLS handshake timeout" fetching envtest-v1.37.0-linux-amd64.tar.gz. Ruled
+  #     out: the MTU mismatch was real but wasn't (solely) the cause of this.
   # So whatever blocks github.com from the dind bridge sits outside anything this
-  # workspace's repos configure -- a firewall/DNS layer neither Ryuvia/charts nor
+  # workspace's repos configure, and outside the MTU theory -- still unidentified as
-  # Ryuvia/k8s expresses in Kubernetes objects. `make test` fails on the envtest fetch
+  # of 2026-09-30. `make test` fails on the envtest fetch until that's found and
-  # until that's found and fixed (or the binaries are vendored -- see ROADMAP.md/the
+  # fixed (or the binaries are vendored -- see ROADMAP.md/the PR discussion for why
-  # PR discussion for why that was declined for now).
+  # that was declined for now).
   test:
     runs-on: ubuntu-latest
     container:
@@ -75,12 +82,7 @@ jobs:
           fi
 
       - name: Format, lint and test
-        # TEMPORARY: `lint` is blocked by an unrelated github.com git-clone failure
+        run: make fmt lint test
-        # (golangci-lint's own custom-gcl build), which never lets this reach `test`'s
-        # setup-envtest step -- the one the act-runner MTU fix (Ryuvia/charts#272) was
-        # actually meant to affect. Narrowed to `test` alone to get a clean read on
-        # that specific question; revert to `make fmt lint test` once answered.
-        run: make test
 
   # No `chart` job yet -- there's no Helm chart until Stage 6 (ROADMAP.md). No
   # `security` job yet either (govulncheck/gitleaks, as terdut-server has); add one