vdm/gui/docs/pkg-qa-requests-m1.md

GUI → PKG/QA requests (M1)

Filed by lane GUI. These touch PKG/QA-owned files (tools/bootstrap.sh, the CI jobs, docs/agents/AGENT-PKG-QA.md) and the root README.md, so GUI is not making the edits — CLAUDE.md §1. Everything below is apply-ready.

---

R1 — libqt6svg6-dev is not a real package on 26.04 (hard configure failure now)

apt-cache policy libqt6svg6-dev on a clean Ubuntu 26.04 box gives Candidate: (none). The Qt 6 SVG dev package in the 26.04 archive is qt6-svg-dev (currently 6.10.2-2). libqt6svg6 (no -dev) exists as the runtime lib but carries no headers or CMake config, so find_package(Qt6 COMPONENTS Svg) fails without qt6-svg-dev.

I checked the other 38 apt names in bootstrap.sh against apt-cache policy on 26.04 — libqt6svg6-dev is the only one with no candidate. Everything else resolves.

This is no longer a latent bug. gui/CMakeLists.txt is now on main, so the root CMakeLists.txt line

find_package(Qt6 6.6 REQUIRED COMPONENTS Widgets Svg Network LinguistTools)

is live for every build. On a machine without the SVG dev package that is a hard configure failure for the whole project — not a skipped guard, not a GUI-only problem.

Why nothing has gone red yet — and why that's R2's problem too:

• This dev box already has qt6-svg-dev (pulled in by a qt6-base-dev recommends chain), so pkg-config --exists Qt6Svg succeeds and --check is green here.
• CI runs on ubuntu-latest, which is 24.04, where libqt6svg6-dev most likely still resolves. The project targets 26.04 (the ci.yml env comment says as much). So the one automated place that runs bootstrap.sh is on the wrong release to see this, and --check wouldn't catch it there anyway (R2). Wrong package name + runner-vs-target release mismatch means the class of bug PKG/QA owns is currently unobservable in CI.

Fix — three files, same one-line change

tools/bootstrap.sh (in APT_GUI, ~line 87):

 APT_GUI=(
     qt6-base-dev
     qt6-tools-dev
     qt6-tools-dev-tools        # lrelease/lupdate for i18n
-    libqt6svg6-dev             # GUI: SVG icon rendering
+    qt6-svg-dev                # GUI: SVG icon rendering (was libqt6svg6-dev — no such
+                               #      package on 26.04; this is the one with headers +
+                               #      the Qt6SvgConfig.cmake find_package needs)
 )

docs/agents/AGENT-PKG-QA.md (~line 17):

-   `libqt6svg6-dev`, `libsecret-1-dev`, `nodejs`/`npm` and (optionally) `clang` are
+   `qt6-svg-dev`, `libsecret-1-dev`, `nodejs`/`npm` and (optionally) `clang` are

README.md (toolchain table, ~line 106):

 sudo apt update && sudo apt install -y \
-  libqt6svg6-dev \    # GUI: SVG icon rendering
+  qt6-svg-dev \       # GUI: SVG icon rendering
   libsecret-1-dev \   # DAEMON: Secret Service for site logins

---

R2 — --check verifies the outcome, not the thing that can break

--check today runs the "verify toolchain" block: command -v for the binaries, pkg-config --exists for the dev libs (Qt6Core, Qt6Widgets, Qt6Svg, libcurl, sqlite3, libssl, libsecret-1, libavformat), plus the cmake/g++/node version floors. Every one of those checks a result on a box that already installed everything. None of them touch the apt package names in APT_*, which is the only part of the script that can be wrong on a clean machine — as R1 just demonstrated. So --check reported green for a script that fails apt-get install on its target OS.

Proposed fix — make --check validate the install list it would actually run

Add this to the verify block (runs in both modes; in --check mode it is the point of the exercise). It needs neither root nor network — apt-cache policy reads the local package lists that apt-get update already populated:

# Every apt name this script would install must be a real, installable package on THIS
# release. This is the check that would have caught libqt6svg6-dev before the VM did.
log "validating apt package names (${#PKGS[@]})"
missing_pkgs=()
for pkg in "${PKGS[@]}"; do
    # `apt-cache policy <pkg>` prints "Candidate: (none)" for a name with no installable
    # version, and exits 0 either way — so grep the candidate line, don't trust $?.
    cand="$(apt-cache policy "$pkg" 2>/dev/null | sed -n 's/^  Candidate: //p')"
    if [[ -z "$cand" || "$cand" == "(none)" ]]; then
        printf '  \033[31m✗\033[0m %-20s no installable candidate on this release\n' "$pkg"
        missing_pkgs+=("$pkg")
        fail=1
    fi
done
if (( ${#missing_pkgs[@]} == 0 )); then
    printf '  \033[32m✓\033[0m %-20s all %d resolve\n' "apt names" "${#PKGS[@]}"
fi

Notes for whoever applies it:

• PKGS is already assembled above the CHECK_ONLY branch, and it already respects --with-clang / --with-packaging, so the loop covers exactly what would be installed.
• Belt-and-braces alternative: apt-get install -s --no-install-recommends "${PKGS[@]}" (-s = simulate, no root) also catches an un-satisfiable dependency, not just a missing name. Downside: it needs the apt lists reasonably fresh and is noisier to parse. apt-cache policy is enough to catch the whole class of bug in R1; pick whichever you want to own.
• The runner mismatch is part of this. Even with the name-validation loop, the bootstrap-script job on a 24.04 runner validates names against the 24.04 archive, not 26.04. So R2 only fully closes the gap if the check also runs where the project ships — a 26.04 container step in that job (container: ubuntu:26.04), or a documented decision that 24.04 is close enough and why. Without that, a name that is valid on 24.04 and gone on 26.04 (exactly libqt6svg6-dev) still slips through.

---

R3 — the GUI DoD gates have nowhere to run in CI — RESOLVED, ready to wire in

Status: done on GUI's side. The harness tests/integration/README.md was waiting on now exists, builds, and has been run end-to-end (all three gates, all three unhappy-path sub-phases) against a real mockd with no changes needed to the pre-drafted job below.

Path: gui/tests/dod/run.sh <gate> [--json <path>], exactly the contract tests/integration/README.md specifies. <gate> is scroll-60fps | rss-flat | unhappy-path. It builds and tears down its own mockd (isolated XDG_RUNTIME_DIR via mktemp -d), needs no network, and leaves nothing running on any exit path (trap cleanup EXIT INT TERM) — verified live by checking for orphaned tsx/mockd processes after both a passing and a forced-failing run of each gate.

One change from the pre-drafted job: the "Configure + build" step needs to also build the harness binary, not just velox-gui:

- cmake --build --preset dev --target velox-gui
+ cmake --build --preset dev --target velox-gui gui-dod-harness

Nothing else in the pre-drafted YAML needs to change — the # TODO(GUI): path comment on the scroll-60fps line can just come off along with the marker on the line below it.

rss-flat's slack: 20 MiB (VELOX_DOD_RSS_SLACK_KIB, default 20480). Chosen by running the gate locally (VELOX_DOD_RSS_DURATION_SEC=8 override, i.e. not the real 10-minute number) a handful of times against mockd --tasks 10000 --seed 1 and looking at actual post-warm-up growth (single-digit MiB per run here) — 20 MiB gives real headroom above that noise floor without being so loose a genuine per-tick leak in the progress-patch path could hide under it. This has not been proven against a full real 10-minute sanitized run (that's the nightly job's own first execution) or tuned against production-length data yet; treat it as GUI's stated starting number, not a load-tested constant, and expect it may need retuning after the first few real gui-dod-nightly runs land actual series data.

One caveat worth deciding on explicitly: the pre-drafted job builds with cmake --preset dev, i.e. ASan+UBSan (CLAUDE.md: "default for all lanes"). scroll-60fps's frame budget already compensates (gui/tests/dod/dod_harness.cpp multiplies 16.6 ms by 4× when it detects a sanitized build — measured p99 here was ~50 ms against ASan overhead, well under the scaled 66.4 ms budget, so the multiplier is doing real work, not padding for no reason). rss-flat's slack does not get a similar adjustment — ASan's allocator (redzones, quarantine) can look like real growth over a long run in a way this hasn't been validated against yet. Two honest options: run gui-dod-nightly against a release-preset build instead of dev (loses the sanitizers' own bug-catching value for this one job), or accept VELOX_DOD_RSS_SLACK_KIB may need a second, larger number for the ASan build once real 10-minute data exists. GUI's preference is the second (keep sanitizers on everywhere), but this is genuinely PKG/QA's call since it's their job definition.

Verified live, under the exact ASAN_OPTIONS=detect_leaks=1:halt_on_error=1 the sanitizers job already sets (checked against .github/workflows/ci.yml rather than assumed) — no leak-suppression flag needed, on any of the three gates:

• scroll-60fps against mockd --tasks 10000 --seed 1: PASS at p99 ≈ 50 ms (budget 66.4 ms sanitized). Forced red once via VELOX_DOD_FRAME_BUDGET_MS=1 to confirm the fail path and exit code actually work, not just the pass path.
• rss-flat at VELOX_DOD_RSS_DURATION_SEC=6/8: PASS, ~4-6 MiB growth against the 20 MiB slack.
• unhappy-path, all three phases (--slow 900, --flaky 0.3, --drop-connection 5): PASS. One real finding from building this: --drop-connection is currently a no-op over the Unix socket transport in mockd itself (only wired for WebSocket) — filed as gui/docs/proto-requests-m1.md since that's PROTO's file to fix, not GUI's. The gate still passes today on the weaker (but real, and the actually-documented) condition that the client reaches and holds Connected; it just isn't proving a real mid-session drop yet for that one phase.

Two bugs surfaced and fixed by building this harness, both in gui/'s own RPC client (caught by ASan, not assumed): RpcClient::stop() left conn_ dangling after joining its worker thread, so any caller that called stop() and then let the client destruct hit a double-free — the harness's own client.stop() at shutdown found it on the first run. Separately, RpcClient's initial download.list call had a hardcoded limit: 1000 with no paging, silently capping the table at 1000 rows regardless of how many the daemon actually has — scroll-60fps against --tasks 10000 refused to run rather than "passing" against a 1000-row table, which is what caught it. Both fixed on lane/gui before this request was filed.

---

R4 — root CMakeLists.txt's find_package(Qt6 ...) should list DBus explicitly

Not currently broken — flagging a "works, but by accident" for the record. gui/src/ clipboard/GlobalShortcut.cpp (docs/06-risks-and-spikes.md R2's explicit path #2: org.freedesktop.portal.GlobalShortcuts) needs Qt6::DBus. Verified live: the target already exists and links today, even though the root find_package doesn't list DBus in COMPONENTS — this Qt 6 packaging apparently exports every module's CMake target once any component pulls in the shared prefix, DBus included. gui/CMakeLists.txt still guards the clipboard sources on if(TARGET Qt6::DBus) (same pattern the file already uses for veloxproto), so if that turns out to be environment-specific rather than a general Qt 6 CMake guarantee, the build degrades to "feature skipped," not "build broken," on whatever machine finds out otherwise.

Worth making explicit anyway, since relying on undocumented target leakage is fragile:

- find_package(Qt6 6.6 REQUIRED COMPONENTS Widgets Svg Network LinguistTools)
+ find_package(Qt6 6.6 REQUIRED COMPONENTS Widgets Svg Network DBus LinguistTools)

No apt change needed either way — qt6-base-dev (already in APT_GUI) ships QtDBus's headers directly (verified live: dpkg -L qt6-base-dev | grep -i dbus lists the whole QtDBus/ include tree).