From a489bea480eaf47825632d389a447af5673d6f95 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Gabriel=20Fr=C3=B3es=20Franco?= <gffranco@gmail.com>
Date: Tue, 12 May 2026 13:17:10 +0000
Subject: [PATCH] test(keymaps): hard timeouts + serialise the CI job behind
 `test`
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

User reported the editor-keymaps CI job hung for 30 minutes before
"Has been cancelled" — twice in a row. Two changes to keep that from
happening again:

1. Wrap both harness invocations with `timeout` (60s Neovim, 30s
   Vim). A misbehaving LSP attach, a stuck `vim.defer_fn`, or
   ex-mode hung on stdin can no longer hold the runner indefinitely;
   `timeout` kills the process and the wrapper exits non-zero so CI
   shows the real failure shape.

2. `</dev/null` on `vim -e -s` so ex-mode doesn't sit waiting for
   input on CI's non-TTY stdin (the local zsh run inherits the
   user's tty, masking this).

3. CI: the keymaps job now has `needs: test` and
   `timeout-minutes: 10`. `needs: test` serialises behind the cargo
   jobs so we don't fight for the single Gitea runner, and the
   workflow-level timeout is a belt to the script-level `timeout`'s
   braces — if for some reason both fail the same way, the runner
   reclaims itself in 10 minutes instead of 30.

Local timing: `test-keymaps.sh` 10.4s, `test-keymaps-vim.sh` 0.1s.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---
 .gitea/workflows/ci.yaml    | 5 +++++
 scripts/test-keymaps-vim.sh | 6 ++++--
 scripts/test-keymaps.sh     | 4 +++-
 3 files changed, 12 insertions(+), 3 deletions(-)

diff --git a/.gitea/workflows/ci.yaml b/.gitea/workflows/ci.yaml
index 02d3f94..62466b7 100644
--- a/.gitea/workflows/ci.yaml
+++ b/.gitea/workflows/ci.yaml
@@ -59,6 +59,11 @@ jobs:
   keymaps:
     name: editor keymaps
     runs-on: ubuntu-latest
+    # Serialise behind the cargo jobs so we never compete with them
+    # for runners, and inherit their warm cargo cache instead of
+    # forcing a release build of our own.
+    needs: test
+    timeout-minutes: 10
     steps:
       - uses: actions/checkout@v4
       - uses: dtolnay/rust-toolchain@1.83
diff --git a/scripts/test-keymaps-vim.sh b/scripts/test-keymaps-vim.sh
index 9231ecb..77127c8 100755
--- a/scripts/test-keymaps-vim.sh
+++ b/scripts/test-keymaps-vim.sh
@@ -41,9 +41,11 @@ EOF
 RESULTS="$TMP/results.txt"
 
 log "running harness…"
+# `</dev/null` so Vim's ex-mode doesn't hang waiting on stdin;
+# `timeout` so a misbehaving script can't stall CI for 30 minutes.
 NUWIKI_KEYMAP_RESULTS="$RESULTS" \
-  vim -e -s -u "$VIMRC" -c "edit $SEED" -c "source $REPO_ROOT/scripts/test-keymaps-vim.vim" \
-    >"$TMP/vim.log" 2>&1 || true
+  timeout 30 vim -e -s -u "$VIMRC" -c "edit $SEED" -c "source $REPO_ROOT/scripts/test-keymaps-vim.vim" \
+    </dev/null >"$TMP/vim.log" 2>&1 || true
 
 if [[ ! -f "$RESULTS" ]]; then
   echo 'keymap test harness produced no output' >&2
diff --git a/scripts/test-keymaps.sh b/scripts/test-keymaps.sh
index 6a64e3a..c38a79c 100755
--- a/scripts/test-keymaps.sh
+++ b/scripts/test-keymaps.sh
@@ -46,8 +46,10 @@ log "running harness…"
 # `nvim --headless -u init.lua FILE +luafile …` — the harness's
 # `vim.defer_fn` lets the LSP attach before tests fire, then writes
 # the results file and exits. Output path is plumbed through env.
+# Wrapped in `timeout` so a hung LSP / harness fails fast instead of
+# stalling CI runners indefinitely.
 NUWIKI_KEYMAP_RESULTS="$RESULTS" \
-  nvim --clean -u "$INIT" --headless "$TMP/wiki/index.wiki" \
+  timeout 60 nvim --clean -u "$INIT" --headless "$TMP/wiki/index.wiki" \
     -c "luafile $REPO_ROOT/scripts/test-keymaps.lua" \
     >"$TMP/nvim.log" 2>&1 || true