### Motivation
- Prevent concurrent GitHub Actions jobs from attempting to reserve the
same `setup-uv` cache key and failing with "Unable to reserve cache ...
another job may be creating this cache" when multiple workflows run at
once.
- Target workflows that enable `setup-uv` caching and have run
concurrently in CI: pre-commit, check-generated-models, ai-engine, and
sync_files_v2.
### Description
- Added a `cache-suffix` value to the `astral-sh/setup-uv` step in
`.github/workflows/pre_commit.yml`,
`.github/workflows/check-generated-models.yml`,
`.github/workflows/ai-engine.yml`, and
`.github/workflows/sync_files_v2.yml` to create unique cache keys
(`pre-commit`, `generated-models`, `ai-engine`, `sync-files`).
- No behavior changes beyond isolating the uv cache keys per-workflow
and no other workflow steps were modified.
### Testing
- Ran `git diff --check` which completed with no reported whitespace or
index issues.
- Verified each modified workflow is valid YAML by loading them with a
Ruby YAML parser which succeeded for all four files.
- Attempted to run `task --list` to exercise the Taskfile locally but
`task` is not installed in this environment so that check could not be
executed.
------
[Codex
Task](https://chatgpt.com/codex/cloud/tasks/task_e_6a58b399a33083259d65e0614fcc35d2)
# Description of Changes
Collapses the portal admin-roster endpoint (`getAdminSettingsData`,
`/api/v1/proprietary/ui-data/admin-settings`) from a per-user N+1 into a
constant set of queries, and adds the missing
session/user/team-membership indexes.
**Verified on H2 and real Postgres 16, 2,000-user roster:** 10,601 → 7
SQL statements, 600 → 0 writes-during-a-GET, O(N) → O(1). Portal-access
resolution is proven equivalent to the per-user check (parity test), and
a scaling guard fails the build if the endpoint ever regresses.
Also in scope (same controller / session subsystem): `getLoginData`
counts instead of loading the whole user table; `getTeamDetailsData`
fetch-joins authorities; `SessionScheduled` uses one bulk expire + a
bounded purge.
Behaviour note: the roster "active" flag now reflects *any* live session
(a strict superset of the old "newest session only") — no user who was
active is ever shown inactive.
---
## Checklist
### General
- [x] I have performed a self-review of my own code
- [x] My changes generate no new warnings
### Testing
- [x] I have run backend `task check` (spotless + full backend test
suite) — all green
- [x] I have tested my changes locally (before/after benchmark on H2 +
Postgres)
## Description
The processor home-hero and download-editor modal "Open in browser"
buttons were meant to open the editor in a new tab, but they used
`window.location.href = EDITOR_URL`, which replaces the current tab.
This switches both to `window.open(EDITOR_URL, "_blank",
"noopener,noreferrer")`, matching the existing behaviour of the "Open"
button in `EditorStatusCard.tsx`. The `noopener,noreferrer` flags mirror
that same call and prevent the new tab from getting a `window.opener`
reference.
## Changes
- `WelcomeBanner.tsx` — home-hero "Open in browser" CTA
- `DownloadEditorModal.tsx` — download modal "Open in browser" CTA
## Notes
`EDITOR_URL` can resolve to a same-origin path when the editor is the
same SPA, so opening in a new tab triggers a full page load of the
editor app. This is the expected behaviour for "Open in browser".
# Description of Changes
- Patch CVEs in engine Python dependencies (43 alerts): `starlette`
1.3.1, `cryptography` 49.0.0, `pyjwt` 2.13.0, `urllib3` 2.7.0, `aiohttp`
3.14.1, `python-multipart` 0.0.32, `langchain-core` 1.4.8, `langsmith`
0.9.1, `authlib` 1.7.2, `requests` 2.34.2, `idna` 3.18, `pytest` 9.1.1,
`pygments` 2.20.0, `pydantic-settings` 2.14.2
- Cap `pydantic-ai` `<2.0.0` and bump to 1.107.0 (1.99.0 patches
CVE-2026-46678; 2.0 is a separate major migration)
---
## Checklist
### General
- [x] I have read the [Contribution
Guidelines](https://github.com/Stirling-Tools/Stirling-PDF/blob/main/CONTRIBUTING.md)
- [ ] I have read the [Stirling-PDF Developer
Guide](https://github.com/Stirling-Tools/Stirling-PDF/blob/main/DeveloperGuide.md)
(if applicable)
- [ ] I have read the [How to add new languages to
Stirling-PDF](https://github.com/Stirling-Tools/Stirling-PDF/blob/main/devGuide/HowToAddNewLanguage.md)
(if applicable)
- [x] I have performed a self-review of my own code
- [x] My changes generate no new warnings
### Documentation
- [ ] I have updated relevant docs (if functionality has heavily
changed)
- [ ] I have read the section Add New Translation Tags (for new
translation tags only)
### UI Changes (if applicable)
- [ ] Screenshots or videos demonstrating the UI changes are attached
### Testing (if applicable)
- [ ] I have run `task check` to verify linters, typechecks, and tests
pass
- [ ] I have tested my changes locally
## What
CI cost/routing cleanup. Four changes, each reversible with no code
deleted.
### 1. Disable Depot repo-wide (reversible)
Depot ran on trusted (non-fork) triggers via the `is_fork` output of
`_runner-pick.yml`, driving both the `depot-*` runner selection and the
Depot docker build actions. It's now disabled everywhere behind a single
kill-switch:
- `_runner-pick.yml` gains a dedicated `use_depot` output, forced
`false` via `DEPOT_ENABLED=false`. `is_fork` stays truthful for trust
gating (e.g. `build-enterprise` skipping on forks).
- All `runs-on:` and `USE_DEPOT:` expressions now key off `use_depot`,
so every job falls back to `ubuntu-latest` + buildx.
- `settings.gradle` Depot remote build cache (`cache.depot.dev`) gated
behind `depotCacheEnabled = false`.
**Switch back on:** set `DEPOT_ENABLED=true` in `_runner-pick.yml` (and
`depotCacheEnabled = true` in `settings.gradle`). Depot then reactivates
on trusted triggers exactly as before.
### 2. arm64 PR docker build only on Dockerfile changes
`test-build-docker.yml` was building `linux/amd64,linux/arm64/v8` on
every PR matching the broad `project` filter. With Depot off, the arm64
leg runs under slow QEMU emulation on every code PR. New `dockerfiles`
path filter (`docker/**/Dockerfile*`) gates the arm64 leg: normal code
PRs build amd64 only; PRs that touch a Dockerfile still build amd64 +
arm64. arm64 is still fully exercised on the base-image publish and on
release.
### 3. Tauri PR build -> Linux only, unsigned, deb-only
The PR path built the full 3-OS matrix (Windows + macOS-universal +
Linux), plus the flaky Linux AppImage pass (#6127). PRs now build Linux
only (fastest + cheapest to compile) via a new `minimal` input on
`tauri-build.yml`: Linux deb only, no rpm, no AppImage. The full signed
multi-OS matrix still runs on release, and nightly still warms the Rust
cache with all-OS defaults (unchanged).
Tradeoff: Windows/macOS desktop build breaks are caught by nightly
(all-OS) rather than the introducing PR.
### 4. CI self-testing routing
Editing `build.yml` only matched the `project` filter, so a change to
how e2e / enterprise / tauri / engine jobs are dispatched didn't
actually run those jobs. Added a `ci` anchor (`build.yml` +
`.github/config/.files.yaml`) that every job-gating area filter now
includes, so editing the router or the filter config runs every job.
Also added the orphaned reusable workflows (`e2e-*`,
`frontend-validation`, `docker-compose-tests`, `test-build-docker`,
`check-openapi`, `check-licence`) to their area filters so editing a
reusable workflow self-tests.
## Validation
- All workflow YAML + `.files.yaml` parse; anchor resolution verified
(every job-gating filter resolves to include the `ci` paths).
- Gradle evaluates `settings.gradle` cleanly; `spotlessGradleCheck`
passes.
# Description of Changes
Removes the feature flags for enabling policies on both the backend and
frontend. We shouldn't be releasing another self-hosted release that
doesn't include policies, so it makes sense to do this now. Builds that
don't have the Processor will just not run policies because they won't
have any. Beyond that, the API should always be available, but checks
whether the user actually has the entitlements to run policies (whether
they have credits/a payment method available)
# Description of Changes
The policies stored in the DB are currently encrypted at rest, because
one version in the past included S3 keys. These are now stored properly
in the credentials system and I've manually removed the only policy that
used S3 (it was very recently released). Since there's no S3 (or other)
credentials in the policies stored JSON now, we might as well just
decrypt them. This PR pairs with #7034 to fix the issues - #7034 makes
it resilient to crashing when attempting to load encrypted JSON that's
been encrypted with the wrong key, and this makes it so if it does load
any encrypted policies, they'll be re-saved decrypted, so we should have
a vanishingly small number of encrypted policies over time.
## What this does
Consolidates the frontend's colour/theme system into a small,
well-defined token layer and reworks the theme picker. The goal was a
minimal, scalable set of semantic tokens that the editor **and** the
Processor/portal (and Storybook) all share, plus a theme model that's
easy to reason about.
## Token architecture (`core/theme/`)
A four-file layer, imported once via `index.css`:
| File | Role |
|---|---|
| `primitives.css` | The raw palette — the **only** place literal
colours live (neutral ramps `--p-gray-*`/`--p-zinc-*` + status hues). |
| `colors.css` | ~21 semantic `--c-*` tokens (surfaces, text, borders,
primary, status) mapped from primitives per theme. **Reference these.**
|
| `compat.css` | Legacy names (`--bg-*`, `--text-*`, `--color-*`)
aliased onto `--c-*` via `:root:root` so ~200 existing files keep
working. |
| `dimensions.css` | All non-colour tokens (spacing, radius, z-index,
type, motion) — single source, resolving prior collisions. |
A blocking linter (`scripts/lint/theme-lint.mjs`, run in
`frontend:lint`) enforces "literals only in `primitives.css`" within
`core/theme/`, and has a non-blocking WCAG contrast report. See
`core/theme/README.md`.
## Theme model
- **Mode** (`light` / `dark` / `system`) and **accent** are independent.
Each mode has its own accent (`lightPrimary` / `darkPrimary`).
- The editor is always `data-app-theme="custom"`; `ThemeProvider`
injects the accent as `--user-primary` and sets `data-accent`.
- **Two accent states:**
- A **colour** (preset or custom hex) → tints every surface that hue
(whole-app theming).
- The **`default`** sentinel → neutral surfaces (white/grey light, zinc
black/grey dark) with blue buttons, no tint. (`data-accent="default"`
opts surfaces out of the tint.)
- Accent contrast guardrails (`utils/customPrimary.ts`): lightness
clamps so an accent can't collapse into the base, a contrast-picked
on-primary foreground, and an accent-as-foreground variant so accent
text is never dark-on-dark.
## Theme picker (Settings → General)
- 3×5 grid: a distinct **Default** icon chip (not a colour) + 14 curated
accents, in a dropdown per mode.
- **Custom** colour via the shared `ColorInput`, with a live gamut clamp
(`clampValue`) that refuses white/grey/black — the picker handle sticks
at the boundary and preserves the working hue at achromatic extremes.
- "Restore theme to default" resets both modes.
## Other
- Dark mode is a true neutral zinc (no navy "midnight" tint); the
Mantine dark ramp and Tailwind dark channels were neutralised to match.
- Pre-paint inline script in `index.html` applies theme + accent before
first paint (no FOUC); portal and editor now share the same
`preferences.theme` source of truth.
- High-visibility surfaces migrated to tokens (FAB, landing upload
buttons, portal hero banners); scattered per-component colour swaps were
intentionally **left for a follow-up** to keep this PR focused.
## Testing
- `task frontend:check:all` (typecheck all variants + eslint + prettier
+ colour-lint) green.
- Verified light/dark, default vs tinted accents, and the custom clamp
via computed styles in the dev preview.
> Note: the `prerender-og` build step failing in the e2e/deploy jobs is
unrelated to this diff — it's in `vite.config.ts` (untouched here) and
builds cleanly locally.
## Overview
Adds a **Classification policy** to the processor's policy catalogue,
set up the same way as the Security policy. This moves classifier
configuration out of the editor (where the labels UI landed in #6898 and
was then removed with the rest of the editor's policy-management surface
in #6932) and into the processor, which is now the single place policies
are configured.
## What it does
- **Classification card** in the processor policy catalogue. Always
shown, but **setup is locked until the backend reports the AI engine is
on** — so admins can see the capability they're missing rather than it
being hidden entirely.
- **Setup wizard** mirrors Security: the workflow step shows the team's
**classification label editor** (reused
`LabelsEditor`/`LabelsEditorModal` — add box, chip grid, per-label icon
picker, import/export, reset) instead of tool toggles, since classify is
a single non-configurable step.
- On enable, the team's label vocabulary is **seeded with the 268
built-in defaults** (clobber-safe: only when the team has none). On
upload the document is classified against the team's labels and tagged;
on SaaS with the engine on, files group by category in the editor
sidebar.
## Reuse & consolidation
- Reuses the existing labels table, `labelsFile` helpers, and default
vocabulary. Labels read/write through the processor's own
`apiClient.local` (not the editor's axios client) so auth/base routing
stays explicit; the wire shape is shared.
- Consolidates policy-category icons into a shared, **id-keyed**
`policyCategoryIcon` util (outline glyphs) used by both the editor and
the processor, replacing the processor's emoji-glyph map (and the stray
`schedule` key that rendered a bare dot).
## Testing
- `task frontend:typecheck:{core,proprietary,portal}`,
`frontend:lint:eslint`, `frontend:test` (156 files / 1305 tests) — all
green.
- Verified in Storybook: the Classification card renders, the setup
wizard shows the label editor (268 defaults), and the full labels editor
opens with icons/import/export/reset. Added an MSW handler for the
app-config + labels endpoints and a `Classification` wizard story.
## Notes for reviewers
- The AI-engine gate reads the public `/api/v1/config/app-config`;
classification labels use `/api/v1/classification/labels` (team-scoped,
team-lead/admin-gated, `policies.enabled`); the classify step hits
`/api/v1/ai/tools/classify-and-label` — all pre-existing backend from
#6898.
- Known parity behavior (matches the editor hook): a transient failure
loading team labels falls back to showing the defaults; not changed here
to avoid diverging the two hooks.
# Description of Changes
The Policies page and all the frontend logic for running Policies is not
making use of the bidirectional type mappings that we now have to safely
convert from frontend to backend param models and vice versa. This
changes the way we track the types throughout so we use the mappings
properly.
Because of this, the Add Watermark settings in Policies now actually
pre-populate with the defaults instead of with nothing like they
previously did.
<img width="791" height="725" alt="image"
src="https://github.com/user-attachments/assets/cbdf4ae0-35af-4792-bf64-89216e48d304"
/>
# Description of Changes
Some of the tool settings make use of editor preferences indirectly, but
the Processor never gets that provider, so it crashes when trying to
load them.
Merges the `hotfix/v2.14.2` branch into `main`.
on the hotfix branch:
### What this actually changes on `main`
- **Version bump 2.14.1 → 2.14.2** `build.gradle`, `tauri.conf.json`,
both AUR `PKGBUILD`s, and the two `serverExperienceSimulations.ts`
test-config files.
- **Fix Postgres user settings for some users** removes `@Lob` from
`User.java that broke settings for some Postgres users.
- **Release workflow: stop msiexec hang in Windows signature verify**
---------
Co-authored-by: Ludy <Ludy87@users.noreply.github.com>
Co-authored-by: James Brunton <jbrunton96@gmail.com>
Co-authored-by: ConnorYoh <40631091+ConnorYoh@users.noreply.github.com>
Co-authored-by: LFdev <146497073+LFd3v@users.noreply.github.com>
Co-authored-by: EthanHealy01 <80844253+EthanHealy01@users.noreply.github.com>
# Description of Changes
Redesign S3 connections based on feedback from #6948. Also redesigns the
UI for Sources to make them more like the Pipelines page which improves
UX quite a bit. There's still plenty more UI/UX work for Sources and S3
but moving in the right direction.
# Description of Changes
- Replaced the Tauri application icon with an RGBA-formatted PNG.
- Added a root `.imgbotconfig` that excludes the Tauri icon from
automatic image optimization.
- Fixed the `desktop:test` compilation failure caused by
`tauri::generate_context!()` rejecting the previous non-RGBA icon.
- Prevented ImgBot from potentially converting the icon back to an
unsupported indexed PNG while optimizing its file size.
- Verified that the current icon uses PNG Color Type 6 (`Truecolour with
alpha`).
```sh
[desktop:test] error: proc macro panicked
[desktop:test] --> src/lib.rs:202:12
[desktop:test] |
[desktop:test] 202 | .build(tauri::generate_context!())
[desktop:test] | ^^^^^^^^^^^^^^^^^^^^^^^^^^
[desktop:test] |
[desktop:test] = help: message: icon /Users/runner/work/Stirling-PDF/Stirling-PDF/frontend/editor/src-tauri/icons/icon.png is not RGBA
[desktop:test]
[desktop:test] error: could not compile `***-pdf` (lib) due to 1 previous error
[desktop:test] warning: build failed, waiting for other jobs to finish...
[desktop:test] error: could not compile `***-pdf` (lib test) due to 1 previous error
task: Failed to run task "desktop:test": exit status 101
Error: exit status 101
```
---
## Checklist
### General
- [ ] I have read the [Contribution
Guidelines](https://github.com/Stirling-Tools/Stirling-PDF/blob/main/CONTRIBUTING.md)
- [ ] I have read the [Stirling-PDF Developer
Guide](https://github.com/Stirling-Tools/Stirling-PDF/blob/main/DeveloperGuide.md)
(if applicable)
- [ ] I have read the [How to add new languages to
Stirling-PDF](https://github.com/Stirling-Tools/Stirling-PDF/blob/main/devGuide/HowToAddNewLanguage.md)
(if applicable)
- [ ] I have performed a self-review of my own code
- [ ] My changes generate no new warnings
### Documentation
- [ ] I have updated relevant docs on [Stirling-PDF's doc
repo](https://github.com/Stirling-Tools/Stirling-Tools.github.io/blob/main/docs/)
(if functionality has heavily changed)
- [ ] I have read the section [Add New Translation
Tags](https://github.com/Stirling-Tools/Stirling-PDF/blob/main/devGuide/HowToAddNewLanguage.md#add-new-translation-tags)
(for new translation tags only)
### Translations (if applicable)
- [ ] I ran
[`scripts/counter_translation.py`](https://github.com/Stirling-Tools/Stirling-PDF/blob/main/docs/counter_translation.md)
### UI Changes (if applicable)
- [ ] Screenshots or videos demonstrating the UI changes are attached
(e.g., as comments or direct attachments in the PR)
### Testing (if applicable)
- [ ] I have run `task check` to verify linters, typechecks, and tests
pass
- [ ] I have tested my changes locally. Refer to the [Testing
Guide](https://github.com/Stirling-Tools/Stirling-PDF/blob/main/DeveloperGuide.md#7-testing)
for more details.
Anthony Stirling