Port the modalcloser abx-plugin as a SingleFile --browser-script rather
than a spawned Puppeteer daemon. No external extension download required.
Core behavior (singlefile.rs):
- MODAL_CLOSER_DIALOG_OVERRIDES: main-world <script> bridge (best-effort;
blocked by strict script-src CSP) that overrides window.alert/confirm/
prompt/print, nulls window.onbeforeunload, traps its setter, and wraps
window.addEventListener to no-op 'beforeunload' registrations.
- MODAL_CLOSER_POLLING_SETUP: defines _archivr_mc_run() which runs two
passes on every tick (immediate first run, then setInterval every 500 ms
matching MODALCLOSER_POLL_INTERVAL default):
Pass 1 (best-effort, CSP-sensitive): main-world <script> bridge calls
Bootstrap/jQuery/jQuery UI/SweetAlert teardown APIs.
Pass 2 (always, CSP-immune): isolated-world DOM mutations — Escape-key
dispatch (Radix/Headless UI/Angular Material), backdrop clicks, full
CSS selector hiding for 40+ named consent/overlay vendors, body
scroll-lock reset. Direct DOM mutations need no inline script execution.
- resolve_modal_closer_config(): reads ARCHIVR_MODAL_CLOSER env var
(default true); no external resource required.
- modal_closer_enabled: Option<bool> in CaptureConfig so Default::default()
yields None (follow env var) rather than false.
Schema (database.rs):
- modal_closer_enabled column in auth DB instance_settings (default 1).
- Idempotent ALTER TABLE migration for existing databases.
- get/update_instance_settings updated (SELECT col 6, UPDATE param ?7).
HTTP (routes.rs):
- modal_closer_enabled in CaptureBody and UpdateInstanceSettingsBody.
- Per-capture body overrides global setting which overrides env var.
Frontend:
- SettingsView ExtensionsTab: third ext-card for Modal & Dialog Closer.
- CaptureDialog: per-capture toggle in Advanced Options, state initialized
from server default, included in submitCapture payload.
- api.js: submitCapture forwards modal_closer_enabled to capture body.
Behavior vs original modalcloser daemon:
- Functionally identical on non-strict-CSP pages (the large majority).
- Gap: <script> bridge is subject to page CSP; page.evaluate() is not.
On strict-CSP pages framework teardown and dialog overrides are no-ops;
CSS selector hiding and Escape dispatch (Pass 2) always work.
- Gap: alert/confirm/prompt overridden to instant no-op vs CDP timed
dialog.accept() with 1250 ms delay; irrelevant for archival in practice.
Orphan cleanup bug: archiving x🧵A downloaded D/C/B/A JSONs and
media, but only registered artifacts for A. D/C/B files had no
entry_artifacts rows and were deleted as orphans.
Fix (staged scraper output, precise touched set):
- tweets::archive() stages all scraper output in temp/{ts}/tweet_stage/,
validates, then renames JSONs to raw_tweets/. Return type changed from
Result<bool> to Result<Vec<String>> (store-relative relpaths of every
produced tweet JSON, i.e. the exact touched set).
- tweets::rearchive() (new): same staged approach but always runs the
scraper. On scraper failure (tweet deleted/private), errors before
touching raw_tweets/ so existing data is preserved.
- register_tweet_artifacts() (new private helper in capture.rs): registers
every JSON in the touched set as a raw_tweet_json artifact, parses each
for media blobs, registers those too. JSON read failure is a hard error
with context, not a silent skip.
- record_tweet_entry() now accepts tweet_json_relpaths: &[String] and
delegates artifact registration to register_tweet_artifacts().
- perform_capture() passes the returned vec from tweets::archive().
Re-archive feature:
- capture::perform_rearchive(): looks up entry by uid, validates
tweet/tweet_thread, runs tweets::rearchive(), atomically swaps
entry_artifacts in a DB transaction. archived_at, title, tags,
collections are untouched.
- database: add get_entry_for_rearchive() and delete_entry_artifacts().
- POST /api/archives/:id/entries/:uid/rearchive: requires ROLE_USER,
creates capture job, returns 202 + job_uid, runs perform_rearchive in
spawn_blocking.
- Frontend: re-archive button in ContextRail for tweet/tweet_thread
entries; polls job at 500ms; refreshes entry detail on success; shows
error text on failure. Poll interval cleared before early-return on
entry deselect to prevent stale updates.
* feat: uBlock Origin Lite integration for ad-blocking during WebPage captures
- singlefile.rs: when ARCHIVR_UBLOCK=true and ARCHIVR_UBLOCK_EXT is set,
archivr owns Chrome's lifecycle (--headless=new, --remote-debugging-port,
--load-extension); single-file connects via --browser-server instead of
launching its own Chrome. Falls back to old behaviour with ublock_skipped=true
when the ext path is missing or invalid.
- capture.rs: thread ublock_skipped through CaptureResult
- database.rs: add notes_json TEXT column to capture_jobs (DDL + idempotent
ALTER TABLE migration); update_capture_job_status gains notes_json param
- archive.rs: expose notes_json in CaptureJobSummary
- routes.rs: store {"ublock_skipped":true} in notes_json on completed captures
- ToastStack.jsx: warning toast variant (toast--warning) with Details expander
and Ignore button
- CaptureDialog.jsx: fire warning toast when poll result has ublock_skipped
- App.jsx: sessionStorage-backed Ignore suppression for ublock warnings
- styles.css: .toast--warning (amber left border) + .toast-warning-detail
- flake.nix: ublockLite derivation fetches uBOLite_2026.705.2152.chromium.zip
(pinned SHA256) from uBlockOrigin/uBOL-home; sets ARCHIVR_UBLOCK_EXT in both
archivr and archivr-server wrappers
Env vars:
ARCHIVR_UBLOCK=true (default) — enable uBlock during WebPage captures
ARCHIVR_UBLOCK_EXT — path to unpacked uBOL extension dir (set by Nix)
* feat: Extensions settings tab + capture dialog redesign with Advanced options
Settings/Extensions tab (admin-only):
- New 'Extensions' tab between Cookies and Storage
- ExtensionsTab component: shows uBlock Origin Lite card with pill toggle
- Reads ublock_enabled from instance settings; patch via existing PATCH endpoint
- Shows ublock_ext_available status from server (whether ARCHIVR_UBLOCK_EXT is set)
Instance settings:
- Add ublock_enabled BOOLEAN (default true) to instance_settings auth DB table
- Idempotent ALTER TABLE migration in initialize_auth_schema()
- get/update_instance_settings include ublock_enabled
- GET /api/admin/instance-settings now also returns ublock_ext_available (computed
from ARCHIVR_UBLOCK_EXT env var at request time)
- PATCH /api/admin/instance-settings accepts ublock_enabled
Per-capture override:
- CaptureBody gains ublock_enabled: Option<bool>
- CaptureConfig gains ublock_enabled: Option<bool>
- singlefile::save() gains ublock_enabled_override: Option<bool> param
- Capture handler resolves: body override > global instance setting > env var
- submitCapture(aid, loc, qual, extensions) in api.js passes ublock_enabled
Capture dialog redesign:
- Archive button: full-width, 13px padding, min-width 220px, primary CTA
- Cancel: full-width but text-style, below Archive
- ‹Advanced options› chevron toggle (rotates on open)
- Expanded panel shows uBlock toggle for this capture session
- Loads global ublock_enabled default from instance settings on mount
Styles:
- .ext-toggle pill switch (44×24 and 36×20 small variant)
- .ext-card for Settings Extensions tab
- .capture-advanced + .capture-advanced-panel + .capture-chevron
- .capture-ext-row / .capture-ext-label / .capture-ext-name / .capture-ext-desc
- .form-hint utility class
* fix: remove ublock_enabled from INSERT OR IGNORE in DDL batch
The INSERT ran before the ALTER TABLE migration added the column,
causing 'table instance_settings has no column named ublock_enabled'
on existing databases. The INSERT OR IGNORE for the default row only
needs the original columns; the migration's DEFAULT 1 handles the
new column for existing and new rows alike.
* feat: Reader mode via Mozilla Readability.js
Adds an opt-in 'Reader mode' advanced option to the capture dialog.
When enabled, Readability.js is injected as a browser script during
SingleFile capture; it fires on single-file-on-before-capture-start,
replaces the page body with the distilled article content, injects a
clean typographic stylesheet, and adds a header with title/byline/site.
Falls back silently if Readability fails (e.g. non-article pages).
- vendor/readability/Readability.js Apache 2.0, Mozilla, v0.6.0
- singlefile.rs: embed READABILITY_JS + READER_MODE_WRAPPER_JS via
include_str!; write both to temp dir when reader_mode is true;
base_single_file_cmd now accepts &[&Path] for multiple --browser-script
- capture.rs: CaptureConfig.reader_mode: bool
- routes.rs: CaptureBody.reader_mode: Option<bool> (defaults false)
- api.js: submitCapture passes reader_mode in payload
- CaptureDialog.jsx: Reader mode toggle in Advanced options (off by default)
* fix: diagnose single-file no-output-file error + prevent stdout dumping
- Add --dump-content=false to every single-file invocation to prevent
the Docker-detection heuristic from routing HTML to stdout instead of
the output file (the heuristic can trigger in some macOS environments)
- Improve the no-output-file error message to include: temp dir contents,
stderr, and first 200 chars of stdout — this gives enough context to
diagnose any remaining cause without re-running
* fix: switch uBlock loading from --browser-server to --browser-args
The --browser-server (CDP) path caused 'Unexpected server response: 404'
on macOS Chrome because simple-cdp's WebSocket upgrade to the debugger
endpoint failed after Chrome started — likely a version-specific CDP
endpoint shape mismatch.
New approach: single-file always manages Chrome. When ARCHIVR_UBLOCK_EXT
is set, --headless=new, --load-extension, and --disable-extensions-except
are injected via --browser-args. single-file's browser.js prefix-strips
its own conflicting flags before appending ours, so --headless=new
overrides the default --headless (enabling extension support in headless).
Removes allocate_free_port, wait_for_chrome_ready, run_single_file_with_server
(all dead code now). Docblock updated to reflect actual behaviour and notes
the --single-process caveat: uBOL's declarativeNetRequest static rulesets
are expected to work (network-stack level, not service-worker), but this
has not been mechanically verified under --single-process.
Smoke tested on macOS (this machine): capture with --load-extension + all
three browser-scripts (strip, Readability, reader-mode wrapper) produces
output file correctly. Ad-blocking verification deferred to manual test
with a tracker-heavy URL.
* fix: use correct single-file hook event (single-file-on-before-capture-request)
Prior scripts listened on 'single-file-on-before-capture-start' which
does not exist in single-file-core 1.1.49. The real hook is:
single-file-on-before-capture-request (dispatched by initUserScriptHandler
after receiving single-file-user-script-init; userScriptEnabled defaults
to true in args.js so it always fires when --browser-script is passed)
Changes:
- strip-scripts: -start -> -request (no preventDefault needed; synchronous)
- READER_MODE_SCRIPT: -start -> -request; add 'installed' meta marker at
script-evaluation time so artifact inspection can distinguish 'script
not injected' / 'hook never fired' / 'Readability parse failed'
* fix: correct singlefile.rs docstring (scripts.js concatenates, not isolates)
* fix: dispatch single-file-user-script-init so request hook fires
single-file's initUserScriptHandler (in single-file-bootstrap.js) listens
for 'single-file-user-script-init' and only then installs
_singleFile_waitForUserScript. Without that dispatch our scripts'
'single-file-on-before-capture-request' listeners were never reached,
so neither strip-scripts nor reader-mode Readability applied.
Dispatch the init event at the top of strip-scripts (always present) and
redundantly in READER_MODE_SCRIPT. Verified end-to-end: artifact for
run_b3181d6d276e4e56a1a6c356ef9bbe8f has
meta content="applied", max-width:680px CSS, 0 script tags.
* feat: cookie consent extension support (ARCHIVR_COOKIE_EXT)
Mirrors the uBlock Origin Lite integration exactly:
Backend:
- singlefile.rs: resolve_cookie_ext_config() reads ARCHIVR_COOKIE_CONSENT
(default true) + ARCHIVR_COOKIE_EXT path; extension paths comma-joined
into --load-extension / --disable-extensions-except so uBlock and cookie
ext can coexist; SaveResult.cookie_ext_skipped tracks miss
- database.rs: cookie_ext_enabled column on instance_settings (DEFAULT 1);
idempotent ALTER TABLE migration; get/update wired through
- capture.rs: CaptureConfig.cookie_ext_enabled: Option<bool>; threaded to
singlefile::save(); cookie_ext_skipped surfaced in CaptureResult
- routes.rs: CaptureBody + UpdateInstanceSettingsBody get cookie_ext_enabled;
capture handler resolves effective value (body overrides global); notes_json
only includes skipped fields that are true; GET instance-settings includes
cookie_ext_available from env path check
Frontend:
- api.js: submitCapture forwards cookie_ext_enabled
- SettingsView.jsx: 'I Still Don't Care About Cookies' card in Extensions
tab; always-active toggle (user can disable even when ext not installed);
amber 'Not configured' hint + ARCHIVR_COOKIE_EXT guidance when unavailable
- CaptureDialog.jsx: 'Block cookie banners' toggle in Advanced options;
always shown with amber hint when ext not configured; defaults from
global setting
Operator setup: download + unzip the extension from GitHub releases, set
ARCHIVR_COOKIE_EXT=/path/to/unpacked/ext. No Node daemon needed.
* fix: surface cookie_ext_skipped warning toast in CaptureDialog
* feat: package istilldontcareaboutcookies in flake, wire ARCHIVR_COOKIE_EXT
Add isdcac derivation mirroring ublockLite:
- Fetches ISDCAC-chrome-source.zip v1.1.9 from GitHub releases
- Validates manifest.json at extension root before install (guard against
nested-folder zip regressions in future releases)
- Sets ARCHIVR_COOKIE_EXT in both archivr and archivr_server wrappers
Verified: nix build .#archivr-server and .#archivr both succeed;
wrapper scripts export correct store paths; manifest.json present at root.
* fix: gate consent-overlay cleanup on cookie_ext; reset overflow; narrow selectors
- Strip overflow:hidden from body/html only when cookie_ext is active for
the capture — prevents mutating legitimate pages when the feature is off
- Remove .fc-dialog (Google Funding Choices), .qc-cmp2-*, .sp-message-container,
#sp-cc, #usercentrics-root as fallback for CMPs the extension misses
- Removed overbroad [class^="uc-"] and [id^="usercentrics"] selectors
that could match real page content
* fix: remove ad placeholders when uBlock active; kept height causes blank gap
uBlock Origin Lite blocks ad network requests but first-party placeholder
elements (ins.adsbygoogle, #aswift_* iframe hosts) retain their computed
height (e.g. 280px for a top banner), leaving a large blank space at the
top of captured pages.
Gate cleanup on ublock_ext.is_some(): remove ins.adsbygoogle, aswift_*
iframes, and google_ads_* iframes before SingleFile serialises. Also
collapse the parent container if it becomes empty after removal.
* fix: walk up to .top-ad/.google-auto-placed ancestor before removing ad slot
Removing only the inner ins.adsbygoogle left the outer .container.top-ad
wrapper (with pb-4 padding) in the layout, preserving the blank gap.
Now walk up via closest() to the nearest ad-slot container class before
removal so the whole slot including padding collapses.
Adds per-instance cookie rules (admin-only) that are injected into
every network touchpoint during capture.
Storage:
- New cookie_rules table in the auth DB (idempotent migration)
- Rules have pattern_kind (global/wildcard/regex), optional url_pattern,
and cookies_json (validated as string-only JSON object)
Matching (resolve_cookies_for_url):
- Global rules always apply
- Wildcard: * and ? with full metacharacter escaping; matched against
hostname via reqwest::Url when pattern has no ://, full URL otherwise
- Regex: matched against the full URL
- Later rules in ordinal order override earlier ones per cookie name
All six network touchpoints receive resolved cookies:
- http::probe_url_kind and http::download: Cookie request header
- singlefile::save: Netscape cookie file -> --browser-cookies-file
- ytdlp::fetch_metadata and ytdlp::download: Netscape cookie file -> --cookies
- tweets::archive: semicolon credentials file -> --credentials-file
(only when both ct0 and auth_token are present; otherwise falls back
to ARCHIVR_TWITTER_CREDENTIALS_FILE)
Security:
- Cookie files written 0o600 (owner read/write only)
- Exact parsed hostname used as cookie domain (no PSL stripping)
- Files deleted unconditionally before any error propagates,
including spawn failures (hold-result-then-delete pattern)
- No cookie values in process args (no --add-header exposure)
API: GET/POST /api/admin/cookie-rules, PATCH/DELETE /api/admin/cookie-rules/:uid
Frontend: Cookies tab in Settings (admin only) with rule list,
inline edit, pattern-type selector, client-side JSON validation
CLI: CaptureConfig::default() - no behaviour change
254 tests passing (4 new cookie-rule handler tests)
* feat: add YouTube Music and Spotify source detection
- Add Source variants: YouTubeMusicTrack, YouTubeMusicPlaylist,
SpotifyTrack, SpotifyAlbum, SpotifyPlaylist
- ytm:ID shorthand → music.youtube.com/watch?v=ID (audio-only, forced
in core regardless of caller quality hint)
- ytm:playlist/ID and music.youtube.com/playlist URLs detected but
fail with 'not yet implemented' via fail_run
- Spotify URLs/shorthands detected and fail fast with clear DRM error
via fail_run (after run item created, so status is visible in /runs)
- source_metadata: youtube_music/music/audio and spotify/music/audio
(entity_kind='music' for UI pill, representation_kind='audio' stored)
- locator_to_ytdlp_url includes YouTubeMusicTrack for probe endpoint
- generate_entry_title: 'Title — Artist' for YTM tracks
- Frontend: isVideoSource handles ytm: and music.youtube.com/watch;
Spotify returns false (no probe, clear server error on submit)
- Placeholder updated to include ytm:ID
- SOURCE_ICONS: youtube_music (red disc) and spotify (green waves)
- 14 new tests covering all new sources (163 total, all pass)
* fix: prevent yt-dlp playlist expansion and stalled run recovery
- Add --no-playlist to ytdlp::download and fetch_metadata: URLs with a
list= parameter (e.g. music.youtube.com/watch?v=ID&list=RDAMVM…) no
longer cause yt-dlp to expand the full playlist and hang; both the
metadata probe and the download are now single-item only
- Fix fail_stalled_capture_jobs to also recover archive_runs and
archive_run_items: capture_jobs.run_uid is NULL at crash time so a
join is unreliable; instead fail all archive_runs/items still
in_progress directly, then recount failed_count via subquery.
Startup recovery now makes the Runs UI reflect the correct failed
state after a hard shutdown
- Expand fail_stalled_jobs_on_restart test to assert archive_run and
archive_run_item rows are also marked failed, not just capture_jobs
* fix: use play triangle for youtube_music icon
- database.rs: add has_active_capture_jobs(), list_orphaned_blob_rows(),
all_referenced_file_relpaths(), delete_orphaned_blob_rows()
- routes.rs: GET/DELETE /api/archives/:id/blob-cleanup (ROLE_ADMIN)
- GET returns {orphaned_blob_rows, deletable_files, total_bytes}
- DELETE has two active-capture guards (before and after disk walk)
to prevent deleting files mid-capture; walks raw/ and raw_tweets/
- Referenced set = entry_artifacts.relpath ∪ live blobs' raw_relpath,
so a file is never deleted if any artifact still points at its path
- api.js: scanOrphanBlobs(), deleteOrphanBlobs()
- App.jsx: pass archiveId to SettingsView; add 'storage' to SETTINGS_TABS
so /settings/storage survives refresh/back navigation
- SettingsView.jsx: new Storage tab (admin-only) with idle→scanning→
scanned→deleting→done/error state machine; shows file/record counts
and human-readable byte sizes before a btn-danger confirm
Tests (14 new, 248 total passing):
- database.rs: has_active_capture_jobs for pending/running/completed,
list_orphaned_blob_rows, all_referenced_file_relpaths edge cases,
delete_orphaned_blob_rows preserves referenced rows
- routes.rs: auth (401), active-capture 409 on GET and DELETE,
end-to-end delete preserving referenced file and removing orphan
blob file + extra disk-only file
Store how many bytes of each entry's artifacts are already on disk from
an earlier entry (content-addressed blob deduplication means shared
blobs are only stored once).
Design
------
- Add `cached_bytes INTEGER NOT NULL DEFAULT 0` to `archived_entries`
- Precompute at capture time via `database::refresh_entry_cached_bytes`
called after all artifacts are saved for every capture path
(web page, generic URL, tweet, yt-dlp/local)
- One-time migration in `initialize_schema`: detects missing column via
PRAGMA table_info, ALTERs the table, then back-fills all existing rows
with the correlated subquery
- `database::cascade_cached_bytes_after_delete` ready for when entry
deletion is implemented; designed to run asynchronously after the
delete is acknowledged to the user
- `cached_bytes` included in `EntrySummary` and all four SELECT paths
(list_root_entries, search_entries, list_entries_for_collection,
entries_for_tag) via the shared ENTRY_SELECT_COLS constant
Frontend
--------
- `EntryRow` shows a `% cached` sub-line under the size when non-zero,
with a tooltip showing the raw cached byte count
- No separate API endpoint or extra fetch — value rides in the existing
entries list response at zero extra query cost per read
- fix(hash): hash raw bytes instead of lossy UTF-8; add hash_bytes
- feat(database): add get_blob_by_sha256 lookup
- feat(font_extractor): extract embedded font data-URIs from archived HTML
