Switch to make fmt for formatting. The issue with zig fmt --check ./ is that it
ends up checking zig-pkg. Really need to give an explicit list, which is were
the make target comes in (also updated the github action so use that same target
to keep everything consistent)
Depends on https://github.com/lightpanda-io/demo/pull/311 and on enabling perf
on lpd-bench-hetzner (1). Also relies on perf-fmt update (there's a PR).
The adds a whole build for the PRs parent, to generate a comparison between the
PR and its parent. But it should be built in parallel, so hopefully it doesn't
actually slow anything down.
(1)
```
sudo apt install linux-tools-common linux-tools-$(uname -r)
echo 'kernel.perf_event_paranoid = 2' | sudo tee /etc/sysctl.d/99-perf.conf
sudo sysctl --system
```
I looked at profiles and couldn't find a different. I let claude work on this
for hours, and it's convinced it's an environment change. Runs started to
consistently get slower after Nov 4 ~12:45 UTC, and no code change explains it
(and builds with the exact same code in the CI, at that time, ran faster before
than after).
The action moves from v0.4.8 to v0.7.16. The badge page has read
"unknown" since July although every run reports the score as
published; the current release publishes in the format the page
serves, and it also analyses the composite actions under
.github/actions.
.plumber.yaml becomes an overlay on the CLI's built-in defaults, so
the controls added since v0.4.8 apply without copying the whole
template again. The trusted action list stays, the nix installer
included. The perf-fmt image is exempt from the forbidden tag check
while it tracks :latest, an exemption the CLI added in v0.6.4. The
rustup download inside .github/actions/install and the archived CLA
action keep their exceptions, with the reason next to each. The gate
moves from the all-or-nothing default to 85 points, so one small
finding never blocks a pull request, and the job gets the actions read
scope the action documents.
Six workflows had no permissions block and ran on whatever the
repository default grants. They get contents: read at the workflow
level. The three perf-fmt jobs pull ghcr.io/lightpanda-io/perf-fmt, a
private image, with GITHUB_TOKEN, so they keep packages: read on top of
it. No job in these workflows writes to the repository.
actions/checkout leaves the job token in .git/config by default, where
every later step and tool in the job can read it. Every checkout opts
out except the three whose job pushes a branch with that token: the
orderfile regen and the two nix dependency updates keep theirs.
The Debian package build ran in debian:stable-slim, a tag that moves
with every Debian point release. It is pinned to the current index
digest, with the tag kept next to it for readability.
Read browser tool schemas from Schema.all() instead of re-parsing them,
format table rows through Anchor/Cell formatters, share one flag writer
between traits and hints, and use mcp/tools.zig's browser_tool_list
directly. Group headings move next to the generator, and an addGenerator
helper in build.zig drives both the skills and docs steps.
Render MDX partials for the docs site from the registries the runtime
serves, so the browser tools, MCP tools and agent REPL command references
can't drift from the code:
- browser-tools.mdx: every tool, grouped, with its description, an
argument table from its input schema, and its script/navigation/page
state traits.
- mcp-tools.mdx: titles and annotation hints per tool, linking to the
browser tools reference, plus the MCP-only tools and resources.
- agent-commands.mdx: meta, LLM and browser slash commands in the syntax
the REPL parses.
Tool.group() assigns each tool its docs section as an exhaustive switch,
so a new tool can't ship undocumented. The release workflow attaches the
partials to each release as lightpanda-docs.tar.gz for the docs repo to
pull.
I know this was suggested before, and I was against it. But this take uploads
the debug symbols to s3, so that, as long as we know the version of the binary,
we can reverse it. (Also, we get far fewer crashes now).
Dependencies:
- v8: zig-v8-fork 0.17 branch (lightpanda-io/zig-v8-fork#218).
- sqlite3: build the amalgamation directly; the allyourcodebase wrapper
has no 0.17 support yet.
- pcre2: master commit with 0.17 support (no release yet).
- translate_c package (2.0.0) replaces the deprecated addTranslateC.
default_init is set to keep 0.16's zero-initialized struct fields.
Build:
- b.pathFromRoot/build_root/args/sysroot are gone: resolve paths from
b.root, use addPassthruArgs, declare configure-time file and directory
dependencies, and poison the configure cache for the git version.
- Drop curl config values the template never used (now an error).
- Drop the 0.16 zip-fetch workaround from the Makefile and CI.
- -Doptimize=ReleaseFast -> -Doptimize=fast.
Language and std:
- @intFromEnum/@enumFromInt -> @backingInt/@fromBackingInt.
- Struct-of-arrays @typeInfo (field_names, field_types, decl_names,
param_types, error_names).
- `a ** n` removed: @splat for arrays, string.repeat for strings.
- errdefer captures removed: split WebDriver.run and fulfillRedirect.
- @hasDecl only sees pub decls: make StyleManager's Spec.finish pub,
otherwise it silently stops running.
- SafeAllocator replaces DebugAllocator; the test runner now has to
initialize std.testing.allocator_instance itself.
- std.fmt.allocPrint/bufPrint -> Allocator.print/std.mem.print,
dupeZ -> dupeSentinel, std.builtin -> std.lang, builtin.os ->
builtin.target.os, zon.parse arena API, BufferFirstAllocator, and the
remaining renames (getLastOrNull, bit_set, ascii.find*, meta.Int).