feat(release): enable start-at-login by default in research Qt builds

[ErikBjare]

Study-readiness follow-ups for the Research Edition (Qt bundle is the study deliverable; participants are Windows/macOS):

• New scripts/patch_research_edition_awqt.py: research builds flip aw-qt's new autostart_on_first_run default (aw-qt#131, merged) to true, so participant machines register start-at-login on first launch with zero setup steps. The user's later tray choice is always respected (marker file; disable is never overridden). Fails closed if the aw-qt pin predates the option — a research build can't silently lose reboot-survival.
• Wired into both Qt build jobs (build-qt and build-qt-manylinux-2-28); build-tauri untouched since it doesn't bundle aw-qt.
• Bumps the aw-qt submodule to include aw-qt#131 (atomically with the fail-closed step, so research builds keep working).
• getversion.sh: exclude *-research tags from the dev-version fallback — master dev builds currently report v0.14.0b4-research.dev-<sha> because git describe picks the research tag as nearest. Exact-match is deliberately untouched: building on a research tag still reports the research version.

Verified locally: patch script success/fail-closed/missing-file paths, getversion output on master (v0.14.0b4.dev-0a9d836), workflow YAML parses, aw-qt test suite green on the bumped pin (103 passed).

[greptile-apps]

Greptile Summary

The PR enables first-run start-at-login for Research Edition Qt bundles while preserving the standard and Tauri build paths, and prevents research tags from becoming the base of ordinary development versions.

• Adds a fail-closed patcher for aw-qt’s autostart_on_first_run default.
• Runs the patch before both Qt build paths when producing Research Edition artifacts.
• Updates the aw-qt submodule to a revision supporting the new option.
• Excludes *-research tags only from non-exact development-version fallback.

Confidence Score: 5/5

The PR appears safe to merge with no concrete blocking or independently actionable non-blocking issues identified.

The research-only workflow gating, pre-build patch ordering, fail-closed marker check, canonical tag exclusion, and untouched exact-tag behavior align with the repository’s Qt packaging and release-version contracts.

Important Files Changed

Filename	Overview
.github/workflows/release.yml	Applies the new fail-closed aw-qt patch in both research-only Qt build paths before bundle construction.
aw-qt	Advances the Qt desktop submodule to the revision that introduces the first-run autostart configuration option.
scripts/package/getversion.sh	Excludes canonical research tags from development-version ancestry while retaining exact research-tag identity.
scripts/patch_research_edition_awqt.py	Performs a narrow first-occurrence default replacement and aborts research builds when the expected source marker is unavailable.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart LR
  Trigger[Release workflow] --> Edition{Research edition?}
  Edition -- No --> Standard[Build unchanged Qt artifacts]
  Edition -- Yes --> Window[Patch research watcher defaults]
  Window --> Qt[Patch aw-qt first-run autostart default]
  Qt --> Build[Build Qt bundle]
  Build --> Artifact[Research Windows, macOS, or Linux artifact]
  Tags[Git tags] --> Exact{HEAD exactly tagged?}
  Exact -- Yes --> ExactVersion[Keep exact tag, including research suffix]
  Exact -- No --> Fallback[Find latest non-research tag]
  Fallback --> DevVersion[Generate development version]

Loading

_{Reviews (1): Last reviewed commit: "build(deps): updated submodule (aw-qt: a..." | Re-trigger Greptile}

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-09-01T20:45:16.782519Z	4912d09	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.