Upgrade the docs site to Next 16 and React 19:
- next 15.5.9 -> 16.3.0
- react / react-dom 18.2.0 -> 19.2.8
- @vercel/analytics 1.1.1 -> 2.0.1
- docsgpt-react 0.5.1 -> 0.6.3
- pagefind 1.4.0 -> 1.5.2
- typescript 5.9.3 -> 6.0.3
Two deliberate holds:
Next 16 makes Turbopack the default builder, but nextra 4.6.1 is not
compatible with it. Its loader calls transformerTwoslash() at module top
level, and twoslash gets an undefined ts.sys inside Turbopack's
webpack-loader shim, so every .mdx file fails to compile. Nextra offers no
option to disable twoslash, so the dev and build scripts opt out with
--webpack until nextra supports Turbopack.
typescript stays on 6.0.3 rather than 7.0.2 for the same reason: TS 7 is the
native rewrite and drops the ts.sys API that @typescript/vfs and twoslash
rely on, which breaks the build even under webpack. twoslash declares
typescript ^5.5.0 || ^6.0.0, so 6.0.3 is the ceiling this toolchain supports.
nextra and nextra-theme-docs were already at the latest 4.6.1.
URLCheck read the host from a regex capture group whose character class
excluded "@", so a userinfo-prefixed URL passed its userinfo off as the
host: https://allowed.com@evil.tld/ satisfied an allow_hosts entry for
allowed.com, and http://ok.com@blocked.tld/ evaded block_hosts. Match the
whole URL and take the host from the parsed authority instead. A URL that
will not parse is now redacted rather than waved through.
The PRIVATE_KEY secret pattern matched only the BEGIN header, so a redact
action masked the header and released the key material and END line in
the clear. Span the whole armored block, falling back to the header when
the block is unterminated or over the cap, and raise max_match_chars so
the streaming guard's window still covers a full PEM. The pattern also
now catches ENCRYPTED, DSA and PGP BLOCK headers, which it missed
entirely.
The documented GUARDRAILS_FLOOR example omitted "enabled": true, and the
field defaults to false, so an operator copying it got a floor that
parsed clean, merged to nothing and warned nowhere. Fix the example and
warn when a floor is set but disabled.
Lockfile-only update; package.json is unchanged.
extensions/react-widget is published as the npm package `docsgpt` and as an
HTML embed, so its production dependency tree reaches every consumer. That
tree carried 5 advisories (2 high, 2 moderate, 1 low), 7 across the full tree.
dompurify 3.4.0 -> 3.4.13 (moderate, XSS)
js-yaml 4.1.1 -> 4.3.1 (high, quadratic-complexity DoS)
linkify-it 5.0.0 -> 5.0.2 (high)
markdown-it 14.1.0 -> 14.3.0 (moderate)
svgo 4.0.1 -> 4.0.2 (high, dev)
brace-expansion 1.1.12 -> 1.1.18 (high, CVE-2026-13149; nested copies)
@babel/core 7.29.0 -> 7.29.7 (low)
npm audit goes 7 -> 0 (production tree 5 -> 0). Verified locally: clean
npm ci, `npm run build` (embed) and `npm run build:react` (library) both
build, npm run lint and tsc --noEmit pass.
chore(deps-dev): bump @babel/plugin-transform-modules-systemjs from 7.29.0 to 7.29.8 in /extensions/react-widget in the npm_and_yarn group across 1 directory
Bumps the Python backend dependency set, holding back the ones that are
resolver-blocked or that regress behaviour this repo depends on.
Notable upgrades:
cryptography 46.0.7 -> 50.0.0 (requires msal 1.37.0, which relaxes its cap)
protobuf 6.33.6 -> 7.35.1 (floats opentelemetry-* to 1.44.0)
openai 2.32.0 -> 2.53.0, anthropic 0.88.0 -> 0.121.0
google-genai 1.73.1 -> 2.17.0 (the 2.0 break is scoped to the Interactions
API, which this repo does not use)
fastmcp 3.2.4 -> 3.4.6, gunicorn 25.3.0 -> 26.0.0
starlette 1.0.0 -> 1.6.0, uvicorn 0.42.0 -> 0.52.1
sentence-transformers 5.3.0 -> 5.7.0, numpy 2.4.4 -> 2.5.1
faiss-cpu 1.13.2 -> 1.15.0, pillow -> 12.3.0 (pinned; security release)
pypdf 6.9.2 -> 6.15.0, lxml 6.0.2 -> 6.1.1 (CVE-2026-41066)
Code changes needed by the bumps:
- openai >= 2.53 rejects a falsy api_key at construction. Keyless
OpenAI-compatible backends (Ollama, llama.cpp, vLLM) legitimately have
none, and pydantic-settings yields "" for a bare `API_KEY=` in .env, so
both call sites now fall back to a placeholder.
- Flask >= 3.1.2 tears a stream_with_context request down twice, and a
ContextVar token may only be reset once. The log-context teardown hook
is now idempotent.
- sentence-transformers renamed get_sentence_embedding_dimension to
get_embedding_dimension in 5.4; use the new name with a fallback.
Held back deliberately:
torch 2.11.0 - 2.13.0 drags torchvision 0.26 -> 0.28 and the whole
docling stack; torch is only probed for cuda.empty_cache()
tokenizers 0.22.2 - transformers pins <=0.23.0 and no stable 0.23.0 exists
transformers - capped <5.9.0: 5.9+ breaks docling's PDF layout model on
Apple Silicon (MPS float64), matching docling-core's own
darwin pin
docling 2.84.0 - 2.118.1 needs rapidocr >=3.9.1 and conflicts with
transformers on macOS; wants its own PR with a
golden-corpus diff
redis 7.4.0 - 8.x defaults socket_timeout to 5s, silently capping the
blocking reads in the device broker and SSE tail
websockets 16.0 - google-genai caps <17.0
marshmallow - dataclasses-json hard-caps <4; spec tightened to match
langchain block - langchain-community 0.4.2 deletes the Qdrant vectorstore
this repo imports; langchain 1.3.x needs websockets <16
ManishMadan2882