The composite index doesn't serve everything the room_id one did. In
SQLite that index is (room_id, rowid), so it also reads a room's messages
in id order, and it is the narrower one to count a room with. Keeping both,
as memberships already does, leaves those lookups as they were.
The migration was named after local time, two hours ahead of the UTC
timestamp Rails generates; it now carries a UTC one.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NuTtwJKhb77Dv7EQqv2C3C
The search box keeps only word characters, but FTS5 still reads AND, OR,
NOT and NEAR in the query as operators, so searching for "AND" or for
"salt AND" raised "fts5: syntax error" and answered with a 500. Quoting
each word makes FTS5 look for it as text, which is how it already treats
the same words in lower case.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NuTtwJKhb77Dv7EQqv2C3C
Every page of a room's messages (opening the room, scrolling back, the
refresh after reconnecting, the bot API) selects WHERE room_id = ? ORDER BY
created_at LIMIT 40. With only the room_id index, SQLite reads every
message in the room and sorts them in a temporary B-tree to return 40, so
these requests grow with the room's history.
A composite index on (room_id, created_at) lets SQLite read the 40 rows in
order straight from the index. It covers every lookup the room_id index
served, so that one goes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NuTtwJKhb77Dv7EQqv2C3C
* Document other Campfire implementations and benchmarks
* Remeasure README benchmarks with the latest Ruby optimizations
* Shorten the README benchmark context
* Preload only uncached messages and reduce rendering overhead
* Keep benchmark summaries without raw JSON results
* Use Ruby benchmark drivers and keep generated results out of the repo
The composer editor carries the lexxy-content class, so the !important
list rules for message bodies hit the menu's ul and li too, leaving the
items cramped and pushed in from the left.
The editor keeps an attachment's content as it finds it, so a hand-written
embed with a url but no href reached the editor with its markup unvalidated,
and a mention saved under Trix's editor carried the generic octet-stream
content type the editor doesn't permit and was dropped on save. Rebuilding
each attachment from its attachable renders the hardened preview partial
and restores the mention content type.
Main strips a javascript: href from a preview node before it's parsed, so
tell legacy Trix attachments apart by their filename instead. escapeHTML
moved to the string helpers, and the unfurling system test now drives
the Lexxy editor.
The companion to #279: the same per-instance `extend` appeared in
test/lib/rails_ext/action_text_attachables_test.rb. `attachable_sgid` is
`to_sgid(expires_in: nil, for: ActionText::Attachable::LOCATOR_NAME).to_s`,
so mint that directly; the minted bytes and the assertion are unchanged.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Agent worktrees under .claude/worktrees are whole extra checkouts of
this repo, and the build context picks them up. .claude is local
session state and never belongs in an image.
Moves the pin from the pre-release 910be91 (0.1.0) to 59e278c, the v0.2.0
tag Fizzy already runs. The default policy now admits IPv6 only inside
IANA-allocated unicast prefixes instead of default-allowing reserved and
unallocated space. No call-site changes: both shim entry points take the
new policy keyword's default.
The invalid-sgid test made one live Room attachable — `rooms(:pets).tap
{ |r| r.extend ActionText::Attachable }` — only to call
`attachable_sgid` on it. That method is `to_sgid(expires_in: nil,
for: ActionText::Attachable::LOCATOR_NAME).to_s`, so the test can mint
the same sgid directly and drop the per-instance extend; the assertion
(a signature that does not verify resolves to MissingAttachable) is
unchanged, and so are the bytes it tampers with.
Link previews fetch a page's OpenGraph title and description, and
Opengraph::Metadata strips tags from both before the values reach the
browser. When a field consists entirely of a markup tag, stripping
leaves it blank, the metadata fails its presence validation, and the
unfurl endpoint returns no content, so no preview is produced.
Add regression tests at the model and controller layers that pin this:
a title or description made only of a markup tag is stripped to blank
and rejected, and the endpoint answers 204. The existing sanitize tests
only cover fields that keep non-blank text after stripping, so this
blank-and-rejected path was previously untested.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0186eyzivcTn6wqjEE4Wnxdt
There were two: the one in dom_helpers escaped through a text node, which
leaves double quotes alone, so it could not have closed the hole in the
link preview's img src.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0186eyzivcTn6wqjEE4Wnxdt
Pasting a link builds the preview by interpolating the unfurled metadata
into an HTML string. The image URL went into src="..." unescaped, so a
page whose og:image carries a double quote closes the attribute early and
everything after it becomes attributes on the preview's img element.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0186eyzivcTn6wqjEE4Wnxdt
A domain name ends in a word, which is what keeps it from reading as an
address. "0x7f.0.0.1" carries a dot and a letter, so the previous shape
check let it through while a browser fetched 127.0.0.1.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0186eyzivcTn6wqjEE4Wnxdt
A browser rewrites the many spellings of an address into one before it
fetches, so "http://2130706433/rooms/1" arrives at 127.0.0.1 while a
comparison here still reads the digits. A preview names a page on the
public internet, so require its host to look like a domain name and leave
the rewriting race alone.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0186eyzivcTn6wqjEE4Wnxdt
The room caches each message's rendered presentation, and its key can't
see the link preview partial, which ActionText renders by name rather than
through a render call the digestor can follow. Without a new version, a
message already in the cache would keep its old preview.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0186eyzivcTn6wqjEE4Wnxdt
Ruby leaves a percent-escape in URI#host, so "https://%77ww.example.com"
read as a different host than the one Campfire answers on while a browser
unescaped it straight back to us. A host that carries an escape, or a
trailing dot, is now measured the way the browser will read it.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0186eyzivcTn6wqjEE4Wnxdt
A preview belongs to the page it previews, so both URLs point somewhere
else. An absolute URL on our own host passed the scheme and host checks,
and every reader's browser fetched it with their session attached, which
turns a message into a GET request made on the reader's behalf.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0186eyzivcTn6wqjEE4Wnxdt
Ruby parses "https:/rooms/1" as an HTTPS URL with no host, and a browser
resolves it against whatever origin Campfire is served from, so the scheme
check alone still let a message body aim the preview at a path here.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0186eyzivcTn6wqjEE4Wnxdt
A link preview's link and image come from attributes on the message body,
which the composer fills in from the unfurl the server performed. A body
written by hand can put anything in those attributes, and the preview
partial rendered them as they were.
Keep the link and the image only when they parse as absolute http or https
URLs, so nothing in a message body can aim either one at another scheme or
at a path on this Campfire, and render the title and the description as
text.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0186eyzivcTn6wqjEE4Wnxdt
Stanko Krtalić
David Heinemeier Hansson
Marcello Costagliola
dependabot[bot]
Sam Ruby
Jeremy Daer
Rosa Gutierrez