| Age | Commit message (Collapse) | Author |
|---|
|
|
|
Only trusting the sunshine formula would fail operations with:
Error: Refusing to load formula lizardbyte/homebrew/sunshine-beta from
untrusted tap lizardbyte/homebrew.
Note trust is not removed if we remove it from the Nix config.
|
|
error.eta carried its own copy of the HTML shell, so error pages missed the
viewport meta tag, the build-commit footer, and the link to the original
GitHub page. The renderer already passes githubUrl and buildCommit to every
template, so the layout needs nothing new from app.ts.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JZAT3YzYP85DihSvQbDKB6
|
|
A signed-out request for github.com/notifications is bounced to a login
page, so the owner route followed that redirect and left the visitor on
/login. Hand the path over to GitHub before it reaches the route.
|
|
|
|
|
|
|
|
|
|
/search?q=repo:owner/name+...&type=code matched the /:owner route and
rendered a profile page for an owner named "search", which GitHub 404s.
It is the URL GitHub's own search box produces, so it arrives whenever a
search is redirected here.
A /search route now takes it: when the query carries a repo: qualifier
and a type, it strips the qualifier and redirects to that repository's
search route, which already maps issues and pull requests to pages we
render and hands the rest to GitHub. Searches that name no single
repository -- bare queries, org:/user: scopes, a missing type -- keep
redirecting to github.com/search as they did before.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GkGX6puS45xD5KzdV9tHEA
|
|
|
|
layout.eta was two things at once: the HTML shell every page needs and
the repository chrome around it. owner.eta, being the one page that is
not about a repository, could use neither and carried its own copy of
the shell, which had drifted -- its footer named no build commit and
did not link to the original page on GitHub.
The shell moves to views/layouts/base.eta, taking its title and any
deferred scripts as layout data, and the chrome to views/layouts/repo.eta,
a layout over it that adds the clone commands, nav, search box, and the
"/" shortcut that focuses it. Repository pages point at repo.eta and
render as before; owner.eta uses base.eta directly and so gains the
footer it was missing.
Templates now live in more than one directory, so scripts/build.ts walks
views/ recursively and keys each template by its path underneath it.
Names crossing a directory are written from the views root
("/layouts/base.eta"), the one form both Eta's filesystem resolver and
the precompiled resolver in worker.ts agree on; siblings stay relative.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Ba5p88cWtCq42vf3uweweL
|
|
Every page route wrapped its scrape in tryRender, which awaited the
promise, rendered the template on success, and turned a failure into a
redirect or an error page. Hono's onError does that mapping for all of
them at once, so handlers now await their scrape inline and both
wrappers are gone. Fragments under /api/ still fail as an empty body
rather than a page; onError checks for them first, since a fragment
turns a redirect into a 404 instead of following it.
RedirectError becomes an abstract base with InternalRedirectError and
ExternalRedirectError under it, so the handler dispatches on the type
instead of sniffing the location for an https:// prefix.
One behavior change: a thrown value that is not an Error no longer
renders error.eta, because Hono only routes Error instances to onError.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QERwoRJhuqMhKzzCUW4Skv
|
|
|
|
Nothing in a commits payload says whether the path it lists is a file or
a directory: currentBlobPath, despite the name, is only the path filter,
and is set for a directory just the same. GitHub's own UI cannot tell
either, and points its browse button at /blob/<oid>/<path> so that the
redirect resolves a file to its blob. We do the same, and the blob route
already turns that redirect into one of our own tree URLs.
Breadcrumbs now stay in the view they were clicked from, so walking up
from a history lands in the history of each parent directory rather than
in its listing. The last crumb remains the page you are already on.
A path with no commits now says so rather than rendering an empty page.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01R7pRKdmTvNDgT4ZXrwkbH8
|
|
Wrap each rendered line in a <span id="Ln">, for both syntax-highlighted
and plain-text blobs, and shade the line a fragment targets.
|
|
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NWDWkWbj5oZCGQEYoybMrV
|
|
Wikis are one of the last parts of GitHub still served as plain Rails
HTML, with no embedded JSON to read, so this scrapes the markup: a page
and any older revision of it, the _Sidebar and _Footer a wiki can
define, and the page index. Every page is already in the sidebar of
every page, since the "Show N more pages" button only unhides the rest,
so linking the whole wiki costs no second request.
Routes mirror GitHub's and resolve in its order, so an underscore name
reserved for its own actions (_pages, _history, _new, _edit) can never
be mistaken for a page; the ones we don't serve fall through to the
usual redirect. A slug reaches us decoded and has to be encoded again
rather than interpolated the way a ref is, or a page named C# would
truncate at the fragment.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013rQG1io1i4GJQW5RFcXNGd
|
|
pnpm 11 no longer reads the "pnpm" field in package.json and warns that
it ignored onlyBuiltDependencies, which would leave workerd's build
script unapproved and unrun on a fresh install. Its new home is
pnpm-workspace.yaml, which pnpm accepts without a packages key for a
single-package repo.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PbJ1wy2dNdjk7LpME4ud3J
|
|
GitHub moved commitGroups from the top of the embedded JSON into a
commitsRefRoute, alongside the code view's own routes, so every commits
page threw. The groups themselves are unchanged.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BErPR92xppzGFXkPqXPYj2
|
|
GitHub renders a PDF with a viewer of its own and so ships no content for
one, leaving the blob view with nothing to show. Its raw bytes come as
application/octet-stream, which browsers download rather than display, so
serve them from a new embed route that retypes them and point an <object>
at that.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BErPR92xppzGFXkPqXPYj2
|
|
Swap Deno for nodejs-slim_26, which runs server.ts directly. Drops
DENO_DIR and CacheDirectory, which only existed to give Deno's module
cache a writable home under DynamicUser.
Listen on a unix socket like clippersnip rather than a TCP port.
|
|
|
|
|
|
|
|
Coexists with ytdl-sub rather than replacing it, downloading into its own
tree under /srv/media/incoming so the two never manage the same files.
Podcast clients cannot authenticate, so the feed, cover art, episode art
and media stream routes bypass oauth2-proxy with auth_request off. Those
are exactly the routes pinchflat serves unauthenticated itself, each
addressed by an unguessable UUID. The OPML endpoint is exposed as well but
stays token-protected by the app.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
|
|
Fixes socket permission issue. Drop the RuntimeDirectoryMode workaround:
nginx no longer traverses /run/miniflux.
Also drop the now-redundant explicit socket dependencies from bjj-booker
and clippersnip; systemd adds those implicitly.
|
|
|
|
|
|
Binary files with no preview (a .ttf, say) have no rawLines, no
colorizedLines and no richText, so the blob view's final branch threw on
a null textLines and the page 500ed. Offer the raw file instead.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
|
|
The meta refresh page renders as a broken image when the request came
from an <img> on one of our own pages. Such requests are already past
the redirect-loop hazard the refresh works around, so send them a real
redirect instead.
|
|
GitHub rebuilt the repository sidebar, so the selectors parseRepoInfo used
matched nothing and the About box rendered empty. Everything but the latest
release is in the page's embedded payload as sidebarAbout; the latest release
is only in the _sidebar JSON already fetched for languages, so it arrives with
that fragment and is moved beside the release count.
|
|
|
|
|
|
A browser extension redirects github.com to cgithub with declarative net
requests. Its rules match our HTTP redirects back to github.com too, so
those bounce straight back here and loop forever. A meta refresh makes
cgithub the initiator origin instead, which the extension's rules can
exclude, and its delay leaves a page the user can stop on.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
|
|
Workers can't shell out at request time, so the commit has to be baked
in before wrangler bundles the sources. compile-views.ts grows a second
job and becomes scripts/build.ts: alongside the precompiled views it
writes src/commit.generated.ts with `git describe --tags --always
--dirty`, which yields a bare short hash until the repo is tagged and
marks builds made from an unclean working copy.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
|
|
The Node server entry, esbuild bundle, and Makefile rsync/systemd deploy
are gone; wrangler dev and wrangler deploy are the only paths now, so the
dev:cf/deploy:cf scripts take over the plain dev/deploy names. @hono/
node-server and esbuild drop out with src/index.ts, and the README's
systemd socket and nginx snippets go with them.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
|
|
|
|
The owner/name header in the layout now links to the owner page and the
repository root. tryRender passes the request's path and query through
as githubUrl — our URLs mirror GitHub's, so the same path is the page we
scraped — and the footer links there with rel="noreferrer", matching the
no-referrer policy used on redirects to github.com.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
|
|
GitHub no longer server-renders the repository sidebar; its React UI
lazily fetches /:owner/:repo/_sidebar as JSON, which carries the
language breakdown (name, percentage, color). getGitHubLanguages() uses
that endpoint, and a new /api/:owner/:repo/languages fragment route
renders the bar and legend the same way refs and latest-commit already
work, so repo page loads still cost a single scrape.
The latest-commit placeholder loader in refs.js is generalized over
.placeholder[data-src] so both placeholders share it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
|
|
|
|
|
|
|
|
Requires /var/secrets/typetype to be populated by hand:
- env with DATABASE_PASSWORD and POSTGRES_PASSWORD (set the same value)
- youtube_remote_login_internal_token
- youtube_session_encryption_key
Postgres data lives at /var/lib/typetype/postgres as a bind mount rather
than a Docker volume.
Downloads (typetype-downloader + Garage) and SSO are not set up yet.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
|
|
|
|
|
|
|
|
|
|
See https://community.home-assistant.io/t/all-automations-show-no-traces-found/389304
and https://www.home-assistant.io/faq/unique_id
|
|
|
|
|
