A merge that finds the other changed drops its journal so the retry plans
again. Unlink and Keep apart name the local side and leave a link that now
joins another sequence alone; Link only and the Integrate job refuse to
touch a link or a pair another integration holds. A sweep of 404s asks the
server once. To trash... trashes from the kept sequence after a swap and
every trash re-checks for new last copies. The Compare dialog and Settings
choices fit in every language, Details segments say which value they take,
twins need their sibling at the other home, and a linked row keeps three
buttons. Captures updated for the dialog and Settings scenes.
Refs #73
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The metered hold now records the jobs it paused in the index and forgets
any job the user pauses, resumes, retries or cancels, so it never resumes
a pause of the user's and still resumes its own after a restart.
Drop... reports why its dry run is blocked instead of asking first, and
dropping the version Main uses lets Main use the version that stays,
with its hero and keyframes. That version reads "In Main".
Merges:
- A re-plan no longer hashes the other's photos again.
- A resumed merge verifies photos flagged damaged while it waited.
- A merge is blocked while the other's folder holds photos the index
has not read; this is checked again under the lease before the trash.
- Choose each offers Keep both, Kept and Other, and blocks until every
differing photo is chosen.
- Classes are named by role.
The Compare dialog:
- Its headings stay over their columns.
- The Projects part scrolls.
- Every confirmation line carries its own prefix, and a merge's facts
carry none.
- A server card in a match offers Compare... alone.
The Italian strings follow the glossary.
Finding 9 (a local drop counting 0 photos) is not reproduced; a test now
keeps the count at 2.
Refs #73
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The fourth verification round of the matches work found 18 defects in
the merge, Integrate and Matches surfaces. This commit fixes them:
- An Integrate job now carries the planning account's id, so its re-plan
decides ownership the way the dialog did.
- A merge skips only a project the kept sequence holds a real copy of
(name, offset head and history content) and re-points its renders to
that copy.
- A merge runs its trash checks before the Links stage and records an
undo, so an abandoned merge gives the other sequence back its link,
session keys and pairs.
- The metered hold adopts automatic transfers that were paused before a
restart.
- Apply in the Compare dialog re-plans when the plan is pending or stale.
- The Library's Two versions... lists the second versions a Union merge
leaves, each with Drop... (dry run, confirmation, verified partner).
A merge job's Open shows that list.
- The merge's typed confirmation is the invariant count, and the prompt
names it in all 38 languages.
- Italian text uses Principale and Gemello. The Main project is named
through WorkspaceText in the Projects part and in the twin tooltip.
- The session keys tooltip is split into two keys, and To trash... and
Delete on server... say why they are disabled.
- The capture restores the editor account before the Matches scenes.
- Upload again follows SyncHost.CanSendSync (Upload with Edit).
- Matches rows trim archive names in the middle on a rebalanced grid;
the Projects part rows trim, with the full text in the tooltip.
- The sample project targets use ProjectTransplant.FreeName.
DECISIONS, CHANGELOG, STATUS 8.7, the request doc and the 16 changed
capture baselines are updated.
Refs #73
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Integrate trashes a server photo only for the sequence's owner or an admin, a
Details stage stops instead of overwriting a server change the user had kept
this PC's value against, a merge retried after a stop plans again when work
reached the other sequence, detection keeps pairs answered while it ran, and a
sync row whose match the server deleted offers Keep here, Upload again and To
trash. The Integrate and merge logs say in the UI language why a version was
kept.
The edit-picker-twins capture scene showed no Twin chip, since a capture has no
workspace: it now lists sample workspace projects of the sample matches. The
screenshot baselines gain the 38 Matches screens and take the Matches chip,
the three Settings > Server rows, the longer priority list and 1.2.0.
Refs #73
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Upload again now deletes the gone pair row once the new sequence is linked, so
the pair stops reading "deleted on server" or "kept here". A linked upload the
server refuses into its sequence (403 for an account without Edit that does not
own it) falls back to ordinary uploads instead of failing; an Integrate that
would add photos to such a sequence is blocked in its plan (NoAddRight).
Automatic photo transfers pause on a metered connection whatever the option
says. Integrate trashes server photos only against the manifest version that
checked the winner, and a merge is blocked while the other sequence holds a
damaged photo nothing else holds (DamagedPhotos), so nothing unverified reaches
the trash.
The plan and the confirmations now say what really happens: replaced versions
are no longer counted as kept, a merge lists the kept sequence's replaced
photos, plan lines use the text colour, an integrated pair that left photos out
says new photos move by hand, and a merging pair reads "merging" with its
progress. The Compare dialog gets per-side tooltips, class sizes for both
sides, policy-aware class tooltips, column headings and readable file names;
the sync table shows "deleted there" with Review and warns on skipped rows; the
Edit picker names the server, the Library card shows the match status second,
a deleted server sequence is named instead of "0 photos", and the match
tooltips stop calling a sequence a copy. The capture scenes use the server
name, the link names and consistent sample sizes.
Refs #73
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A merge could still trash photos it never verified, resume against a
pair the user had answered meanwhile, and take different unhashed
photos for copies; the dialog and the lists worded several facts
wrongly or lost them on refresh.
- Merge: a retry resumes only while the pair is still the job's;
Verify requires the other archive online and its folder in place,
stops on an unreadable photo, files a differing unhashed photo as a
second version and must cover every frame before Trash; the plan
hashes unmatched unhashed photos first, lists "Deleted in the other"
and names the relinked servers; second versions stay discoverable
(CopySides.VersionsAsync).
- Integrate: a retry after a failure before the journal runs again.
- Sync: a sequence back after a 404 stays Linked while its link holds;
a local cull no longer asks for attention.
- Block lines of both planners name the right sequence, archive,
server and space through one formatter.
- Compare dialog: one Details row per setting with worded values,
Projects headings, Choose focuses the kept-sequence radios, the
photo selection reaches rows outside the viewport, the Replace
confirmation lists the losers per side, plan lines read as queued.
- Matches and Library: counts first and grouped, integrating shown as
a share, local matches name the other archive, the Library keeps its
match and "deleted on server" statuses across a refresh.
Refs #73
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Photos not sent: automatic photos of an integrated pair that failed three
times in a row are now told with an error toast that shows the row, and the
row reads "not sent" with the last reason. The matches-found and linked toasts
of background detections run from their Windows notification too.
Maintenance now clears a merge's or an integration's journal whose job is done,
cancelled, or deleted without a cancel and untouched for the history's days.
Tests pin the matches jobs in the priority queue, the owner's defaults (ask
always, union, copy across), the dialog starting from the settings, Choose
each preselecting nothing, no automatic photos for a pair whose server copy
is gone, the repair of a corrupt local frame by an Integrate, and the owner's
whole scenario end to end against a real in-process server.
Refs #73
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The review of the found-copies work showed pairs that could stick, jobs that
could act on a pair the user had answered otherwise, and detection gaps:
- forgetting a link on an address-keyed server deletes the pair and Candidate
rows of that remote and resets the probe, so the sequence is found again
when it is back; a Linked/Integrated row with no link is asked again
- 404s count only on id-keyed servers, at most once per recheck interval
- probes store the server's newest update time (copy_probes.server_newest,
migration 0012) instead of comparing with this PC's clock; the lookup
sample carries hashed photos by size and capture time too, for servers
that have not hashed their copy
- the older-server gate no longer re-compares unchanged manifests
- a failed Integrate keeps its job id so a retry resumes only an untouched
pair; cancel and stale plans release the pair (ReleaseAsync)
- automatic photo transfers never create or retarget a sequence
- a first upload files photos the server holds elsewhere into the new
sequence
- Unlink re-compares before keeping apart; Trash here names only true last
copies; Drop... trashes one version on both sides after re-checking the
other
- a merge keeping the pair's other sequence swaps the dialog's columns and
re-keys Different choices; Upload all reads every pair
Each fix has a regression test where it makes sense.
Refs #73
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
CopyEngine puts the planners together behind ICopyEngine so the App can find,
compare, integrate, merge, link, unlink and keep pairs apart. It also moves
photos that one side deleted to the other side's trash. Each trash step
re-reads both sides when applied, and nothing moves if any photo was restored
since the dry run. The sync agent now queues the new photos of pairs that were
integrated with automatic photos, both ways, at background priority. Failures
back off, and three in a row are reported so the pair can show "not sent". On
a metered connection, integrations are held like other server transfers.
AppServices builds the engine and registers the Integrate and Merge handlers.
A CopyWatch looks for local copies after start and after jobs that change
photos. Merges wait for workspace work that has not been sent home.
Refs #73
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A local sequence and its server twin can now be integrated into one linked
pair by a resumable job instead of by hand. The planner runs the dry run of
design 5c: photo steps per policy (the wins policies decide only the Different
class; Only here and Only there always move), left out and AutoPhotos, last
copies, the one portable payload, projects copied across under free names,
space, rights and every block, and a fingerprint the job checks again.
The Integrate job keeps an atomic journal and resumes at each stage: link,
photos both ways through filtered transfers that never create or retarget a
server sequence (a 404 marks the pair remote gone), trash only once the winner
is present (locally through the photo trash inside the job, on the server at
the version its dry run read), the agreed details, then projects through
ProjectTransplant. Refusals are kept and reported; a failure or a cancel sets
the pair back to Linked or its previous state.
Linked uploads now file into their target when the server announces
upload-into. The plan context travels with the job so rights and names give
the same plan on the job side.
Refs #73
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Brings in the owner review 6 parcel P4: SequenceMerger plan and
fingerprint, the journaled merge job (copy, verify, move, trash) and
the sequence payload writer. Merges without conflicts on top of P1b;
build has no warnings, Core, Data and Server suites, the copy-related
App tests, i18n check, xaml-lint and package-guard pass.
Refs #73
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A refresh no longer links copies by content behind the user's back. After the
gate is released, CopyFinder looks for copies whose ids differ: the server's
photo lookup (or 1.1's gate and the hints on an older server), session keys,
and an exact comparison by manifest, bounded per refresh and remembered in
copy_probes so nothing is compared twice while neither side moves. What a pair
becomes is the Link matches setting's call: asked by default, linked by itself
only as the setting allows, through a compare-and-set that a Keep apart wins.
Local detection finds copies inside this PC's archives and never links them.
SyncService gives every link its pair, adds Unlink (kept apart, observation
kept), marks pairs whose server sequence answers 404 (once per refresh, twice
reads deleted, never on an address key or in a 404 sweep), keeps deletion
counts current at each comparison, writes an Integrate's agreed details to both
sides by content (ApplyAgreedAsync) and queues uploads at a given priority.
CopyReviewBuilder assembles what the Compare dialog shows.
Refs #73
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
SequenceMerger plans the merge of a local pair (kept sequence, photos
copied by policy, last copies, renders, details, projects, session keys,
links, blocks and a fingerprint) and queues the MergeSequences job.
MergeJobHandler applies the confirmed plan through a journal, so every
stage resumes: copy and file the other's photos, move renders, hash-verify
every original of the other in the kept one (or list it as a last copy),
trash the kept one's losing versions, write the details
(SequencePayloadWriter), copy projects, move keys and links, then move the
other to its archive's trash only at the verified frames version. A cancel
drops the journal and returns the pair to its previous state.
Refs #73
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Brings in the owner review 6 parcel P6: Server Matches list, compare
dialog, copy settings rows, library and picker twin surfaces. The three
parcels merge without conflicts; build, tests, i18n and xaml-lint pass.
Refs #73
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Brings in the owner review 6 parcel P2: the photo lookup endpoint and
the copy path for uploads into an existing sequence, with Server tests.
Refs #73
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Server > Matches shows the pairs the copy engine found (same sequence on the
server and this PC, or twice here), attention first, with the design 8.1
actions; Keep apart and Unlink toast with Undo. The in-window Compare dialog
walks photos, details, projects and the plan, blocks until details are
chosen, and confirms only what replaces or merges (typed count for a merge).
Sync rows of an unanswered pair read "match found" and offer Compare first;
Upload all local skips suspected copies and says why. Settings > Server gains
Link matches, Integrate photos and Integrate projects. Library cards carry the
match status with Compare..., the sequence picker marks twins, and Jobs opens
Integrate and Merge targets. The sample engine and capture scenes render every
state; tests cover the text rules and the screens.
Refs #73
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Detection, Integrate and Merge need one comparison of two copies of a time-lapse, so
this adds it: CopyComparer classifies every photo (same, converted, same unhashed,
deleted here/there, differs, two versions, only here/there) with strict Differs and
Versions rules so whole-second capture times and bursts never pair, unusable frames
held as nothing and marked for repair, and the score of design 3.2.
CopyRules ports 1.1's same-originals gate exactly for "Link when half shared" and
holds the Decide table, ask-again thresholds, fingerprints, hints and the probe
sample. CopyPayload compares details in portable form (same photo at another index is
no line) and merges the chosen sides into one payload that must parse as a project.
CopySides reads both sides from the index and a server and answers the last-copy
check across every archive. SyncTombstones gains FromDtos for local pairs.
Refs #73
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Clients must recognise the same time-lapse filed under another id on the server, and an
integrated pair must get a photo into the server sequence even when another sequence there
already holds it. The server now answers both, authorised and scoped by itself.
- POST /api/v1/photos/lookup (Read right): per probe, the sequences of the served archive
that hold its hashes in a frame, tombstoned them (deleted or replaced by a conversion),
match its unhashed photos by size and capture millisecond, or carry its session key; only
ids and counts, batched raw SQL, 413 too_many beyond 200 probes or 20,000 items, a 4 MB
body limit so a full request fits.
- Uploads with IntoTarget: sequenceId required, never with Restore, the target must be the
served archive's and the caller its owner, an admin or an editor. Only a copy in the target
stands in; a tombstone keeps the photo deleted. Otherwise the server copies an intact file
of another sequence of the served archive (never another archive of the index) into the
target: quota and room checked, verified copy, hash compared (a damaged source falls back
to a normal upload), committed refusing removed photos, used bytes charged only for a new
frame, frames event published, answered copied.
- server/info announces photo-lookup and upload-into (the desktop's hosted share too).
Refs #73
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Owner review 6 asks Chronos to recognise the same time-lapse held under different ids on the
server and on this PC, compare the copies, and keep them apart or integrate and sync them. This
lays the shared foundation the later parcels build on, without any UI or detection yet.
- Migration 0012 adds found_copies (local-server and local-local pairs with evidence, scores,
answer, policy and the 404 record) and copy_probes; every existing link with a local sequence
becomes a Linked pair, so a server deletion and the per-link policy have a place.
- FoundCopyStore keeps the pairs: upserts that never touch the answer, a compare-and-set of the
state with an Integrating backstop (a failed or cancelled job reads as the previous state),
link promotion and demotion, the once-per-refresh 404 count, the merge's links stage
(session keys, links reset, pairs moved, refused when both are linked on one server), the
local overlap queries and the probes. SyncLinkStore.RekeyAsync moves pairs and probes with
the links in the same transaction.
- Contracts for detection, review, plans, the Integrate and MergeSequences jobs (kinds 25/26,
High level, new detail steps) and the Upload all rule that keeps suspected, kept-apart,
deleted, busy and server-born sequences here.
- Wire: POST photos/lookup, uploads into a target (IntoTarget/Copied, trailing and optional),
frames/delete on the client, capabilities photo-lookup and upload-into, cached per remote.
- The trash plans and deletes accept the job that holds the sequence (its own hold is not
Busy) and a frames version checked under the lease; a Background job never interrupts a
running one.
- Settings for linking, the photo policy and projects; the en and it keys of design section 9.
Refs #73
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Windows hands-on test of v1.1.0 on the development PC (#70).
- The App suite runs whole on Windows again. Five stale screen smoke tests are brought up to date.
- The Reorganize capture gives the sample sidecars.
- Forget fits in the Bulgarian certificate rows.
- The job log shows local time. It stays UTC on disk, and lines are dated backwards from the job's end.
- The Jobs and Server › Jobs Type column takes a 1* share (90–170 px).
- Builds without a tag say 1.1.1.
Closes#70
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Alberto Balbo <alberto.balbo.96@gmail.com>
Owner review 4 and owner review 1 phase 6.
- Work board in Videos (desktop and admin panel): Published and Workspace lanes (#29).
- Anonymous usage counts removed: Chronos sends nothing about its use (#30).
- The archive's whole folder tree is created inside any archive root, on the server and the desktop (#31).
- Video cards start with the preset; sizes keep number and unit together (#28).
- Deploy guide docs/deploy (Docker, Unraid, desktop), BRIEF §24, original documentation in docs/reference/original (#22).
Closes#22#28#29#30#31
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Alberto Balbo <alberto.balbo.96@gmail.com>
Owner review 1 R3, R8, R9. A grouped Export step and Render menu; encoding, playback (reverse, boomerang, loops, fades, hold) and still frames; one videos library on the desktop and the server with a verified trash and copies to a videos folder.
Closes#21
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Alberto Balbo <alberto.balbo.96@gmail.com>
Owner review 1 R5 + R6. Camera rotation, mirror, flip, straighten and crop handles locked to the output; an Exposure Look group; Export output size ("proportions win"); settings format 2 with older-node gates. Also fixes project text and hashes differing between Windows and Linux.
Closes#19Closes#20
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Alberto Balbo <alberto.balbo.96@gmail.com>
Originals in the workspace, local render and send home with provenance, the workspace clean-up, size limit and trash, moving the workspace folder, and the complete Settings › Workspace page (owner review 2, R8, phase 5c). One adversarial review round (35 findings, 32 fixed), a copy-cancel race fixed, 184 strings in all 38 languages.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Alberto Balbo <alberto.balbo.96@gmail.com>
A Chronos Server keeps its sequences' projects, as a PC's archive does: Main and named projects with versions under /api/v1/sequences/{id}/projects, the projects capability and the projectChanged event. The desktop opens a server's sequences in the workspace (proxies downloaded, hash-checked, resumable) with Save version, Sync, conflicts, History, Restore and the project actions against the server. Renders can be of a project version. The admin panel lists a sequence's projects and versions. Strings in all 38 languages.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Alberto Balbo <alberto.balbo.96@gmail.com>
Clients could not reach a server in a container whose ports are published as other host ports: the panel showed the container's ports, the template had no public HTTP port, and a bare host meant port 443. Public ports become panel settings, the panel shows the client address, both clients default to 8443, and an address that answers as another service is named as not a Chronos server (its certificate never offered for trust). The screenshot baselines are captured on the development PC and committed; measured values are fixed while the capture runs.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Alberto Balbo <alberto.balbo.96@gmail.com>
Edit opens a sequence as a project of a workspace folder outside the library. Projects have linear version histories kept as sidecars in .chronos/projects/, Main mirrors the sequence's settings, and sync merges setting by setting. Migration 0009, JobKind.WorkspaceCopy = 20, Settings › Workspace, and 241 new strings in 38 languages.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Alberto Balbo <alberto.balbo.96@gmail.com>
Convert photos already in the archive to JPEG, TIFF-16 or Linear DNG, from the panel's Library and the desktop's Archive and Library. Each run asks Replace (same frames, originals to the trash, recoverable) or New sequence. A dry run lists what converts, what is skipped and why, size before and after, and free space; the job converts exactly the confirmed photos, resumes, and commits all or nothing per sequence. A conversion never changes what Chronos reads from a photo. Replaced originals are erased only on the archive's own word that their conversion is gone. Desktop TIFFs that WIC showed black or as noise now go to the managed reader; the Edit picker no longer spins a core.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Alberto Balbo <alberto.balbo.96@gmail.com>
- Reorganize… on the web panel and the desktop: folder, file and title templates applied to existing sequences, with a dry run and an explicit Apply that runs as a job.
- One naming rule (FrameNaming) for import, upload and Reorganize.
- Crash-safe plan and file journals, held plans, stale-plan refusal; exports move with their folder.
- Writers refuse a folder another program moved or gave to another sequence.
- Server: POST /api/v1/archive/reorganize with a bounded plan cache; template rules.
- Phase 3 docs and translations merged; strings in all 38 languages.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Alberto Balbo <alberto.balbo.96@gmail.com>
- Web panel Library: browse, view, delete sequences and photos, mass operations (analyse, proxies, thumbnails, verify, retag, delete, CSV). It never edits the video.
- Photo trash with tombstones (migration 0008): crash-safe delete and restore, hash-checked restore with reasons, and emptying only by an explicit fingerprinted job. Sync and uploads keep deleted photos deleted; an explicit "Upload again" brings one back.
- Upload from the web panel: in-house SHA-256 in a worker with a fail-closed self-test, capture time from EXIF, chunked, resumable; no folders.
- Desktop trash dialog with photo entries and Empty.
- Admin files are revalidated (no-cache + ETag).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Alberto Balbo <alberto.balbo.96@gmail.com>
- Jobs > Scheduled on the desktop and the admin panel. Every timer that queued work is now a visible system schedule: archive scan, upload clean-up, index backup, job-history and audit clean-ups, desktop housekeeping. New actions: verify, thumbnails, proxies, analysis, maintenance, backup, clean-ups, reports.
- The scheduler: migration 0007, in-house recurrence, compare-and-set claim, catch-up, scan at start.
- Log settings: levels, retention, size parts, total limit, daily or hourly rotation, audit retention, diagnostics share; the server writes Debug/Trace to its files again; logging never fails a job.
Upgrade: CHRONOS_LOG_LEVEL set in an existing container locks both levels; migration 0007 is one-way.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Alberto Balbo <alberto.balbo.96@gmail.com>
- :latest follows the highest stable release, or the highest pre-release while none is stable (Chronos.Tools release-channels); bad tags are refused before anything is built.
- Unraid: CHRONOS_SPACE_PATH and the template's Array view show the array's free space; the write volume (cache) is shown apart; completions check the archive volume (507).
- Settings is a gear, docked at the foot of the admin sidebar; About is the last part of Settings on the desktop and the panel.
- Import > Folder > Whole folder: one sequence per format (desktop only).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Alberto Balbo <alberto.balbo.96@gmail.com>
Gitea writes a release's html_url and browser_download_url with its ROOT_URL,
which on the owner's server is the plain-http LAN address
http://192.168.30.23:3000. The update check accepts only https links, so it
ignored v0.1.0-rc.1 entirely. The page and download links now come from
ChronosReleasesUrl, the address the build names for its users; only the
tags, file names and sizes are read from Gitea's answer.
The lease renewal test's timer periods follow TestTiming too: on the shared
CI host a 100 ms renewal timer came 700 ms late once.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Alberto Balbo <alberto.balbo.96@gmail.com>
The owner renamed the product and moved the repository to the self-hosted
Gitea (https://gitea.encke-hake.ts.net/BalboSoft/Chronos) for more control,
and asked for Actions and a first automatic release with the server package
and the Windows exe (docs/requests/2026-10-23-chronos-gitea.md).
- Chronos everywhere: a byte-exact rename of contents and paths, then a
per-language review of the forms the old name's ending had dictated.
- Gitea instead of GitHub: the update check, licence, issue and support
links read ChronosReleasesUrl; the update check takes the highest stable
version, offers a pre-release build its final release and reports a
repository it cannot see as a failed check. EULA, security policy and
Gitea issue forms live at the repository root.
- Gitea Actions for the Linux act_runner (.gitea/workflows): CI, release,
nightly image and performance gates in the .NET SDK container with git
checkout and the system SQLite. The release cross-compiles the Windows
programs, checks their icon, version and DPI manifest, packs installers
and zips, publishes the Gitea release and pushes the server image
(REGISTRY_TOKEN). Windows-only tests run in an opt-in job on a Windows
runner (WINDOWS_RUNNER).
- New Chronos.Tools commands: zip, gitea-release, postgres-ci (replaces the
docker-exec script, which Gitea's service containers do not allow; the
throw-away server runs without durability) and pe-check.
- Timed tests take their budget from TestTiming (CHRONOS_TEST_TIME_FACTOR,
3 on the shared runner; the weekly performance gates keep theirs).
BREAKING CHANGE: programs, settings, data folders, file types, environment
variables and the Docker image use the new name; nothing is migrated from
Titano, which never had a release.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Alberto Balbo <alberto.balbo.96@gmail.com>
- Implement tests for picker behavior, including loading sequences, searching, and handling empty libraries.
- Validate sequence opening via Enter key and double-click actions.
- Ensure proper handling of unsaved edits and sequence deletion scenarios.
- Verify accessibility features for picker rows across different languages.
- Introduce a dialog counting mechanism to manage unsaved edit confirmations.
Owner review 1, phase 3a (docs/requests/2026-10-15-owner-review-1.md, R1).
The desktop rail is 200 px wide, so the section names fit in every
language, and collapses to 48 px of icons with its ≡ button or Ctrl+B.
Collapsed, each icon's tooltip names its section before saying what it is
for, while screen readers keep the name and the description. The choice is
written to the settings file the moment it changes (nothing else restarts)
and survives Settings › Reset all and Import; Ctrl+B waits behind the
first-run dialog like Ctrl+1…6, and a held key toggles once.
The admin panel's sidebar is 256 px and collapses to icons the same way,
with its own ≡ button or Ctrl+B (on the physical key where the layout's B
types no Latin letter, so Cyrillic, Greek, Hebrew and Arabic layouts work).
The choice is kept in the browser, for the page when the browser refuses
storage, and the tables fit their columns again after each toggle. At
900 px and below it stays the top bar and Ctrl+B stays the browser's.
The wider rail leaves 49-72 px less for every screen at the 1200 px
minimum. Six adversarial review rounds found what that cut, and in every
shipped language no action is cut off now: the Jobs columns share their
width (Target and Progress give way first, Kind/Where/State capped on wide
windows), the editor's preview tools take a second line, a server job's
Cancel has its own header column, the Archive tree keeps a year's label,
the Import rows and result, the certificates and cache rows and the frame
table fit. Headers, titles and cells that may be shortened end in an
ellipsis and show their text in their tooltip; state tooltips start with
the state, and their explanations were reworded not to repeat it (a queued
job "waits for its turn", jobs running one at a time). The crop-result
tool sits beside Crop mask, disabled with its reason while the mask is
off. Tests check these rows in all 38 languages at 1200 px; the App test
program now shuts WPF down in order, so its exit no longer crashes after
every test passed.
Phase 3 lands as two commits: R7 (the Edit section) edits the same shell
files and follows. Also updates BRIEF (12.1, 14.8, 24), DECISIONS,
CHANGELOG, STATUS, the request file, DESIGN and the mockups; every new or
reworded string is in all 38 languages.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Owner review 1, phase 2 (docs/requests/2026-10-15-owner-review-1.md, R2).
Jobs now run one at a time on the desktop and on the server. Parallelism
(Settings > Rendering, TITANO_PARALLELISM) is how many photos the running
job processes at once: thumbnail and proxy passes draw that many photos
together and still write the pack and checkpoints in frame order, decode
tiers are probed that many at a time, the deep analysis measures that many
frames and a render decodes that many ahead. Wherever originals are
developed whole (TIFF, PNG, RAW without a usable preview), fewer run at
once within the memory budget, which the running job now has whole; the
server's frame window is sized to the host instead of 1 GiB.
A new job waits its turn and pausing the running job starts the next one.
An optional priority queue gives each job kind a level (renders, imports
and downloads high, verifies and proxies low by default): a queued job of
a higher level pauses the running one without asking, and that job goes
back to the queue with its checkpoint and resumes by itself afterwards.
Jobs of the same level, and a job's own follow-ups, never interrupt it.
The Jobs screen and the admin panel show the parallelism and the queue;
Settings > Rendering and the admin panel's Settings set them, at once.
TITANO_WORKERS and the Workers key of server.json still set the
parallelism (1 to 64, with a warning at start for the variable).
Titano.Tools gains i18n prune, which removed the renamed keys from the 36
machine-translated languages; the new strings are translated in all 38.
Found while testing: uploads and imports failed when a pass read the
sidecar while they wrote it (atomic writes now retry the rename while a
reader holds the file, and never for a read-only one), and taking a lease
just released while another process read it failed on Windows (the create
now waits for the reader). Five adversarial review rounds checked the
phase before this commit.
Also updates BRIEF (12.8, 14.6, 14.8, 14.9, 24), DECISIONS, CHANGELOG,
ARCHITECTURE, STATUS, the request file and the compose/Unraid templates.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Owner review 1, phase 1 (docs/requests/2026-10-15-owner-review-1.md).
Thumbnail and proxy jobs over several sequences counted sequences while
they checkpointed inside one: the rate decayed until the time left
overflowed TimeSpan, the job failed while publishing progress and again
while saving it, and stayed Running. The throughput is now seeded by the
first window with progress, the time left is bounded (JobEta), saving
progress never throws, image passes count frames, and job speeds use the
right unit (bytes for verify and import).
The debugger stops in WicDecoderProvider.cs when presets changed were
cancelled preview decodes leaving synchronous thread-pool delegates; the
editor's sources, LatestRunner and the render node's proxy task now keep
them in their own tasks, and a cancellation is never an unreadable frame.
GoPro DNGs converted by Adobe software were classified GPR and read only
through their 8-bit preview, and Windows' preview-only DNG decoder was
reported as a native decode. A DNG is GPR only when its raw image is
VC-5, WIC decodes RAW only through the Raw Image Extension and never DNG,
analyses and previews take an adequate embedded preview by choice while
renders develop natively, and frames already indexed as GPR are
reclassified by the next light analysis.
Also records the owner's request (docs/requests), the BRIEF amendments
(§8.1, §8.2, new §24), DECISIONS, CHANGELOG, STATUS and ARCHITECTURE.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The orphans fix and scans treated a folder whose only new photo could not be
read yet (held open, no permission, an offline placeholder) as changed: the
sidecar was rewritten, the covering flags cleared and analyses queued again at
every click. SequenceAdopter.Reconcile now returns null when it indexed nothing
and changed no flag, as its contract says.
A deep pass counted as current by frames version alone, so a pass whose verdict
a proxy made stale overwrote the report a later pass made from the proxies,
then dropped the deep status. It now counts as current only when its report
describes the frames as they are, and otherwise replaces only a saved report
that does not describe them either.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012iLMBNwBE2GnphAsu4cjaC
A desktop scan queues a background deep analysis of the sequences whose
frames it changed when Settings > Analysis > After a scan is on; photos the
Archive screen's orphans fix indexed got only the light analysis, which then
dropped the deep status. ArchiveService.ReconcileAsync takes that setting
and queues the deep analysis too, at background priority. A follow-up is
left out when a queued job of its kind already covers the sequence, whatever
else that job targets, as the scan's follow-ups check.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012iLMBNwBE2GnphAsu4cjaC
A deep pass overtaken by a frames change that keeps the frame count, with a
light analysis queued, leaves an analysis.json of that count in place rather
than save its own. It then still recorded its own report: the deep status,
summary, warning count and Auto values described a report that is not on
disk, and the queued light analysis kept that summary until the next deep
pass. SaveReport now says whether it saved; when it did not, the pass reads
the saved report as a light analysis would and records that one.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012iLMBNwBE2GnphAsu4cjaC
Each index of a shared archive bumps a sidecar put back behind its own row
to one version past that row, which can be the version another index holds
for other frames (a desktop rebuilding a damaged sidecar while the server is
one upload ahead). That index then took the sidecar's frames as unchanged:
it kept the project's positions and what covered the frames, and clients and
editors holding that version saw nothing. When the versions match, a scan
now also compares the row's frame ids and order with the sidecar's; other
frames are read again under the lease and written one version past the row.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012iLMBNwBE2GnphAsu4cjaC
The Archive screen's orphans fix indexed photos copied in by hand but left
AnalyzedLight, ThumbsBuilt and ProxiesBuilt set and recorded the folder's
stamps, so no scan came back to it and the light analysis was never redone:
Auto values, warnings and preview-only tiers stayed those of the old frames.
It now clears the three flags in the sidecar and the row as a scan's
reconcile does, and queues the light analysis on the host's engine (on a
server, the thumbnails and proxies too).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012iLMBNwBE2GnphAsu4cjaC
A deep pass saved analysis.json as soon as it finished, before it checked
that its frames were still the sequence's. A pass overtaken by an upload, or
by another Titano sharing the archive, replaced the report a later pass had
made of the frames as they are: the report endpoint answered 404 and the
next light analysis dropped that pass's deep status. The report now waits in
memory and is saved under the sequence lock, when its frames are current or
the saved report describes the frames as they are no better; the checkpoint
is deleted only then.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012iLMBNwBE2GnphAsu4cjaC
Each tag upsert locks its row until commit, and tags were written in the
order the caller listed them. A PATCH listing them in another order than
an upload, a scan or another edit of a sequence sharing two of them could
lock the rows in opposite orders, and PostgreSQL aborted one writer with a
deadlock that nothing retries. Tags are now written in the order of their
keys, and a set the sequence already carries is not written again.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012iLMBNwBE2GnphAsu4cjaC
ApplyLocallyAsync recorded the folder's scan stamps unconditionally after
writing the server's title, tags, rating or notes into the sidecar. Photos
another process (a desktop sharing this archive) had listed there since
this index last synced the folder then stayed out of the index until the
folder changed again, and the row's frames version stayed behind the
sidecar's, so analyses here recorded nothing for the sequence meanwhile.
It now asks ArchiveIndexer.InStepAsync before its sidecar write, as the
other partial writers do, and records the stamps only when that held.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012iLMBNwBE2GnphAsu4cjaC
An analysis compared only this host's index row with the frames version it
read at its start. On a desktop sharing a server's archive, the server's
upload bumps only the shared sidecar until the desktop's next scan, and a
desktop scan clears nothing. A desktop pass that recorded in that window
wrote Auto values and statuses for the old frame count over the server's
analysis in the shared project.json and sidecar, and nothing analysed the
photos again.
Under the sequence lock the analyser now reads the sidecar first. When its
frames version is not the row's, the photos changed in another process and
this index does not hold them yet: the pass records nothing, logs
Advisor.Log.FramesChangedElsewhere and queues nothing, since a light
analysis here would read the same index. The process that changed them
analyses them, and this host's next scan indexes them (a server's scan
then queues their analysis). A moved row still queues a light analysis
unless one is queued, in which case the pass records as before.
Tests cover the shared-archive case with a second index and engine, and
the queued-light escape on a one-worker engine.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012iLMBNwBE2GnphAsu4cjaC
FrameRepository.UpdateLightStatsAsync updated every frame of a sequence
in batches without locking the sequence row. On PostgreSQL a scan's or
rebuild's sync holds the row and rewrites every frame in its own order,
so a deep analysis writing its statistics at the same time could hold
some frames while waiting for others the sync held, and PostgreSQL
aborted one of the two jobs with a deadlock.
On PostgreSQL the method now locks the rows of its frames' sequences
first, in the same transaction, like every other writer of a sequence's
frames. SQLite's write lock already serialises writers.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012iLMBNwBE2GnphAsu4cjaC