The Favorites count is re-read 300 ms after a star changes. A reload
inside that window started the read on the leaving page, which WebKit
fails ("access control checks") and reports as a page error. After
beforeunload, the workspace clears that timer, aborts count reads in
flight and defers new ones, as well as a finished restore's reload. A
page that stays reads them on its next input or when it becomes visible,
as the legacy admin page does.
* feat(trash): a freedesktop trash backend that renames without replacing or refuses (#480)
* feat(images): remove a file's records and hide it with a removal mark (#480)
Deletes every record of one file under every recorded spelling (image links,
tags, auto-tag marks, pending review items, category, collection memberships,
favorite) in one transaction. A Needs review item that is kept keeps its
annotation rows (ON DELETE RESTRICT). The image_removals table hides an exact
file (locator, size, mtime_ns) until it changes; linking skips a marked file,
and an explicit tag, category, favorite or collection write shows it again.
* feat(images): review and apply Delete images as a reviewed plan job (#480)
POST /prompt_manager/images/delete/preview reviews both modes; the
maintenance apply takes {token, mode}. Library mode removes the records and
writes the removal mark; trash mode moves the file first and removes its
records only after a successful move. The workspace listing, the output scan,
the duplicate find and Tag output images skip marked files.
* feat(images): Delete… in the Images selection bar, inspector and viewer (#480)
The concept's Remove links… place, D159: one review per selection that asks
each time (nothing preselected) whether to remove the files from the library
only or also move them to the trash. The browser test server uses a fake
trash.
* test(images): portable delete tests, the native no-trash proof, D159 in TESTING.md (#480)
* perf(images): cache the removal marks so a listing page reads no extra table (#480)
The marks reload only when this process writes or clears one, or a restore
replaces the database. The route snapshot gains the preview route.
* test(images): the full inspector specs expect D159's Delete… in the panel and the viewer (#480)
* fix(images): the delete review starts nothing without a chosen mode, and dims its primary until then (#480)
An aria-disabled primary ignores every activation, a chosen one starts once,
and the apply helper refuses a missing mode. Tests cover a forced click, a
double click, Enter and Space on the primary, Enter in the dialog and a
submit event, and the server refusing an apply without an exact mode.
The Delete… review asks, for prompts in a saved pair, Archive (checked
on every open, D152's review) or Delete permanently, with their saved
pairs, in a D90 radio row. The choice rewrites the grid, sentences, title
and primary in place (openReviewDialog setRows/setPrimary), and the body
carries the chosen ids with their review digests. On the archived shelf,
Show archived prompts now reads the concept's way back, "← Prompt
library", as a plain button. TESTING.md records D154.
The Delete review's preview now marks which archive ids may be deleted
permanently (in a saved pair, and nobody else's negative), with a review
digest of the prompt and its pairs and the counts the review shows. The
write accepts an optional delete_paired_ids with their digests. Each is
deleted with its own saved pairs (image-pair links, then pairs, then the
prompt as #443 deletes it) only while its digest is unchanged. Any drift
archives it and counts it changed. Paired deletes are backed up and
disk-checked like deletes.
The workspace pages and every file under /prompt_manager/js and
/prompt_manager/lib went out with no Cache-Control and no validator.
ComfyUI's own cache middleware only fills a missing Cache-Control for
paths ending in .js or .css, so the workspace's .mjs modules got no
header at all, and after an update the browser could keep running old
modules until a hard refresh.
Every page and static file now carries Cache-Control: no-cache and a
weak ETag over its bytes, and a request that already holds that version
gets a 304 without a body. A normal reload picks up new code and stays
cheap when nothing changed.
Closes#478
* test: list the slowest tests on request and stop waiting 30 s on a held WAL (#487)
PM_TEST_SLOWEST=N makes tests/run_suite.py list the N slowest tests.
Three history tests restored while their own read held a snapshot, so
each publish sat out the 30 s WAL-truncate wait; they now use a short wait.
Temporary diagnostics for the Windows timings ride along in this commit.
* test(lora): check Windows folder rules without asking the network (#487)
On Windows ntpath.realpath looks up \\server\share and waited about 2.8 s
per test for a host that does not exist. The guard tests now use the
lexical NT path rules, which is what they already ran on Linux and macOS.
* ci(portability): keep Windows test files on the work disk (#487)
The Windows runner's TEMP is on C:, where a flush took about 6 ms and a test
database's create-to-delete about 111 ms; on D: the same took 0.5 ms and
23 ms. The job now points TEMP/TMP at RUNNER_TEMP and lists the 40 slowest
tests. The fake HTTPS handlers get a TLS context that reads no certificate
store, and the missing-share tests name localhost instead of an unknown
server. The temporary diagnostics are removed.
* ci(portability): spell the Windows test TEMP by its short name (#487)
The runner's own TEMP was C:\Users\RUNNER~1\..., so tests met 8.3 short
names in every temp path. The folder on D: keeps that when the disk has
short names, and the step logs which spelling it got. TESTING.md notes
that the slowest list is not printed when a job is cancelled.
* ci(portability): switch on 8.3 names for the Windows test TEMP (#487)
D: made no short names (run 40922502209 logged the long spelling), so
the step now switches them on for D: before it creates the folder, and
warns when the folder still has no short name.
The details panel reads its selected file as soon as the page lists. When the
test's source change landed before that read was answered, the page found the
file again on its own (409 -> rescope -> relist), so the card under the old
numbering the test was about to click was gone. Every test in the spec now
waits for that read before changing sources.
Closes#484
A new selection closed the shared viewer at once, resetting and removing
its video even while that video was still starting. That is the WebKit
(GStreamer) deadlock D62 guards the inline preview against, and it hung
merge-queue run 40903238387. The viewer now stays open, untouched, until
its video reports metadata or an error (or 15 s, or the user closes it),
and the URL is revoked after that. An open asked for meanwhile runs once,
latest wins, only while its selection is still current.
The test made both output/dataset and output/Dataset, which fails with
FileExistsError on Windows and default macOS. It now probes the output
folder at runtime: a case-sensitive one still gets two real files joined by
the folded key; on a case-insensitive one history's spelling already names
the watcher's file, so only the spellings differ. Same assertions, no skip.
test_a_restart_keeps_every_held_scope_and_cursor_valid registered the
second PromptManagerAPI's close with addAsyncCleanup, which unittest runs
after tearDownAsync. The base teardown then unlinked the temp database
while that API's SQLite connections were still open: Windows refuses
(WinError 32). The restart test class now closes its extra APIs in
tearDownAsync before the base teardown.
ImageAPITestCase's teardown now counts connections any live model still
holds on the test database, closes them and fails, so a handle left open
past teardown fails on Linux too instead of only on Windows.
Closes#475
* feat(vocabulary): tag search, tag sort and an Untagged prompts entry (#444)
Tags & categories gains a server-side tag search and sort (keyset pages,
cursor scoped to the query) and an "Untagged prompts" row whose Browse opens
the Library with tag_mode=untagged. Refine this view gets Tag match, saved
Library views keep it, and the routes echo the applied query so a stale
server is an error, never an unfiltered list. Recorded as D149.
Refs #444
* feat(library): filter prompts by output folder and browse tags with Any (#444)
Refine this view on the Prompt library gains the concept's Folder (folders
below an output root, read when the dialog opens) and a Tag match select
(All / Any). The listing takes folder and tag_mode, matches folders on the
normalized file_path under every spelling of each output root, echoes both
so a stale server is an error, and saved Library views keep them. Recorded
as D150.
Refs #444
* fix(library): let saved-view defaults fill the missing #444 filters
libraryViewState passed `?? 'all'` and `?? ''` fallbacks that
normalizeViewState already applies to an undefined value. The redundant
branches were never taken in the page, which dropped saved-views.mjs below
its 95% branch floor. A unit now covers a shelf without the new filters.
Refs #444
* test(saved-views): give the D153 Library view the tag_mode key (#444)
The Library saved-view state now holds tag_mode, and the API requires
every key, so main's new "Highest rated" spec posts it too.
Refs #444
* test(saved-views): give the D153 Library view the folder key (#444)
Refs #444
* fix(library): nest Refine folders by path segment; backslash separates only on Windows (#444)
Live test B1: the Folder list indented a server list sorted as full-path
text, so "-" and "_" sorted before "/" and children landed under the wrong
parent (kiko/char under kiko-v6), with case variants interleaved. The
options are now a tree built by segment: each folder directly under its own
parent, case variants distinct siblings.
On POSIX a folder named "Metadata\Klein..." was listed but filtered to no
prompts, because normalize_folder read "\" as a separator. It now does so
only on Windows (ntpath), as the listing already did.
Refs #444
History's current-file inspection now carries the same image_tags facet as
the Images read, so History's inline current file and its viewer show the
"Image tags" section added in #449. The facet reader moves to
inspection_service.image_tags_reader and both routes use it; History keys it
by the canonical target, the locator the Tag dialog stores.
Closes#457
* fix(history): join outer-whitespace raw text to the history of the prompt it saves (#438)
The Library stores text.strip() while a run keeps its exact raw text. The
history window, run detail scope id, newest pair run and prompt hints now
compare the raw text's stored form (one rule, utils/hashing.py, mirrored in
SQLite with the same code points). Hints also keep matching the exact raw
text, so older rows that kept outer whitespace stay hinted and protected.
* fix(history): join case-variant raw text to the history of the prompt it counts toward (#460)
The prompt hash folds case, so a run of "A Lighthouse" counts toward the
saved "a lighthouse" but never joined its history. History now compares the
hash's own key (prompt_key = strip + str.lower()); SQL calls Python for it
through pm_prompt_key/pm_prompt_hash registered on every managed connection,
since SQLite's lower() folds ASCII only. Hints match by text first (#438) and
fall back to the dedup hash only when no text matches, so no existing hint
becomes ambiguous. A history position from before this change still resumes.
* test(history): a real-host case-variant re-queue joins the prompt's history (#460)
Shard 1/2 of each browser had reached its 18-minute step limit:
--shard splits by test count, so the slow concept specs all land in
shard 1. Three shards per browser, with the limits unchanged. Required
contexts do not change: only the viewerjs aggregate is required, and it
reads the matrix job's result over every shard.
Closes#473
main's only new commit is #452's squash, whose final head (e14ec03) this
branch already holds through #456. Every conflict keeps this branch's side,
so the tree equals e9d01c9's.
* feat(inspector): use recorded run history when the graph alone can't name the prompt (#386)
When a file's own graph leaves the prompt open, both inspect routes ask
recorded run history about the file (one run reported it and it was written
during that run). The save node that run reported it from anchors the walk
like a file-name match, so dataset branches that share one saver prefix
show their own prompt, labelled "Recorded run". Still open at one sampler
whose PromptManager text was recorded at execution (a language model's
output), the recorded text shows as "Execution recorded". Results the file
resolves alone never change; a failed lookup judges the file alone.
* fix(inspector): recorded texts follow history's saver; conflicts anchor nothing (#386)
Review follow-ups: the texts PromptManager recorded are looked up at the
sampler history's saver leads to, never at a sampler the unanchored walk
named; a file name naming a different saver than history anchors nothing;
outputs and texts are read in one transaction; "Recorded run" ranks below
every workflow-inferred source; unexpected lookup errors log at error level.
* fix(history): look up recorded runs by history's own spelling of the file (#386)
The history route inspects a case-normalized target on Windows, while
history matches file names exactly, so the occurrence's reference names
the file.
* refactor(inspector): share the inspector's resolve as inspection_resolution
inspect_prompt_evidence now calls inspection_resolution, which returns the
resolution to use, the anchored result and whether the recorded saver chose
it. Ingestion can call the same helper for files the default profile cannot
decide (#247).
* feat(ingestion): one per-file prompt decision for the scan and the monitor
utils/file_prompt.py holds decide_file_prompt, the ladder the output scan
runs today plus the inspector's rule (inspection profile, then the recorded
saver) before the legacy parser, with a SCAN and a MONITOR policy. Junk
text (blank, numbers, LoRA tags only, model file names) is never stored.
graph_prompt_text and parse_comfyui_prompt move here; the parser accepts
parsed metadata (#247).
* feat(ingestion): store a file's prompt in one transaction, one prompt per file
database/scanned_prompts.py links a file to the prompt of one text, creating
it when new, inside one BEGIN IMMEDIATE transaction: a file linked to
another prompt is a conflict, a failed link leaves no prompt behind and two
writers store one prompt. pair_ingestion gains file_link_rows,
link_conflict and Marks/mark_created; ingest_resolved_pair takes optional
marks for the components it creates. recorded_runs and named_prompt in
run_attribution become public (#247).
* feat(scan): decide each scanned file with the shared ladder (#247)
_ingest_scanned_file calls decide_file_prompt (scan policy) and stores the
result: pairs through ingest_resolved_pair and positives through
store_scanned_prompt, both with the scan marks. Files the default profile
leaves undecided now get the inspector's resolve, with the saver recorded
history names, before the legacy parser, so each branch of a shared-prefix
dataset run links its own prompt. Junk text is never stored, a file linked
to another prompt is a conflict instead of a second link, and a failed link
leaves no prompt behind. _BatchHistory reads the stored prompt and the
recorded saver from one recorded_runs pass. _parse_comfyui_prompt calls the
moved parse_comfyui_prompt. TESTING.md documents the decision.
* fix(ingestion): keep the file-name saver anchor out of the ladder (#390)
The #390 ruling queues a file that only its saver's file name resolves for
Needs review, never linking it. inspection_resolution gains name_anchor;
the ladder passes False, so such a file falls to the legacy floor in the
scan (and to Needs review) and is never stored by the monitor, while the
recorded saver still decides. The inspector keeps the name anchor.
* fix(ingestion): no joined PromptManager text from the legacy floor; junk keeps its step's status
The legacy parser returns prepend + text + append, so its answer is refused
when a PromptManager-family node has a prepend or append part (#343). Junk
found at the undecided-graph step keeps that step's unlinked status and
review, so an ambiguous file is still queued for Needs review. The monitor
policy's ambiguous file no longer asks for the recorded run, and the file
size may be read lazily.
* fix(ingestion): keep the legacy answer that is a PromptManager node's own text
The joined-text refusal also refused the #170 parity files whose
PromptManager prepend is a link the browser parser does not follow: its
answer is then the node's literal text, which is what the node stores.
Refuse only an answer that is not such a node's own text.
* fix(scan): one prompt per file for recorded-run links; writes survive a library move
The recorded run's stored prompt now links in one transaction with the file's
link rows, so a file linked to another prompt is a conflict instead of a
second link. The scan's stores run through survive_library_switch. A single
file asks history once (a one-file _BatchHistory), the scan passes
in_output=False and reads the file size only for files with prompt
metadata. A conflict row reads "Kept: already linked to a different
prompt". TESTING.md records the rules.
* feat(tools): reviewed repair of the foreign dataset links (#247)
tools/repair_foreign_links.py judges every link row of a file that only
foreign runs reported with the shared ladder (scan policy, recorded saver).
The dry run judges a copy, writes a JSON plan and a Markdown list, and
rehearses the apply on that copy to audit the prompts that would lose
every link (deleting them is off unless chosen). --apply refuses while
ComfyUI runs, takes a backup, re-judges every row inside one BEGIN
IMMEDIATE transaction, creates the targets with the scan's marks, relinks
and drops pair links, keeps UNIQUE(prompt_id, file_path), and rolls back
with exit 2 on any drift.
* fix(tools): the repair list keeps whole file paths and shows each target hash
* fix(tools): harden the #247 repair apply after review
- refuse while any other process holds the library, its -wal or -shm open
(tools/comfyui_stopped.py), also in a rehearsal; without /proc only with
--i-stopped-comfyui; require main.py in --comfyui-dir; check again just
before BEGIN IMMEDIATE
- the plan records its library (normcase realpath) and max row id; apply
refuses on another library
- refuse when opening the library changes its schema (run from a checkout
at the live code version)
- verify the backup (integrity_check and counts) and never overwrite one
- target_text is a drift key and must match its hash; keep compares the
linked text's hash, not the stored column
- the dry run rehearses the orphan deletes when --delete-orphans is chosen
- a .md plan path is refused; the list is <stem>.list.md
- any error during apply rolls back, names the backup, exits 2
- relpath on another drive falls back to the full path
* fix(ingestion): match a file's link rows under its resolved spelling too
The scan stores realpath spellings and the monitor the watcher's raw path,
so a file reached through a symlinked or junction output folder could get a
second generated_images row. file_link_rows now queries the path and its
realpath (each normalized and raw) and dedupes by row id. TESTING.md notes
the reverse order is matched only when the raw spelling is the realpath.
* fix(ingestion): a switched-off PromptManager node behind a pass-through stores nothing (#90)
The #90 graph walk follows only prompt-like inputs, so a PromptManagerText
with Save to library off reaching the encoder through another input name
(easy showAnything.anything, a custom node's value) was invisible to it,
and the inspector's resolve stored its text. The ladder now checks the
positive every step names: a resolved or inspected positive whose path
passes a node that is off, or a legacy answer that is such a node's own
text, makes the file not_saved.
* fix(ingestion): any switched-off node upstream of a positive stores nothing (#90)
The ladder walks every input upstream of each sampler or guider positive
(not negative, model, clip, latent or image inputs) for PromptManager nodes
whose Save to library is off or wired from an unreadable node, so the
sampler-less step, an off node with a prepend part and a PromptManagerPair
behind a pass-through no longer store its text. The legacy-floor guard reads
a node's resolved text. Every guard reports not_saved for a switch that is
off and not_saved_wired for one it cannot read.
* fix(ingestion): the switch walk skips negative inputs by #90's own rule
The upstream walk treated only names containing "negative" as negative, so
an off node on a neg, neg_text or text_neg input blocked an enabled
positive. prompt_graph gains is_negative_input (the _NEGATIVE_KEY rule, also
used by follows_input) and the walk uses it. The guards accept a node with
null inputs and a side whose provenance is not an object.
* fix(ingestion): the switch walk also skips any input named with "negative"
#90's rule matches only whole underscore-separated words, so negative2,
negativeprompt and NegativePrompt counted as positive inputs and an off
node wired only to one of them blocked the file. The walk skips those too;
follows_input is unchanged.
* test(tools): the #247 repair never relinks a switched-off node's text (#90)
The plan decides every file through decide_file_prompt, so a not_saved or
not_saved_wired file is a skip. Cover an off node behind a pass-through, an
off node in the recorded saver's branch (literal and wired switch), an off
node on another branch, and an off node wired only to the negative (still
repaired).
* feat(inspector): use recorded run history when the graph alone can't name the prompt (#386)
When a file's own graph leaves the prompt open, both inspect routes ask
recorded run history about the file (one run reported it and it was written
during that run). The save node that run reported it from anchors the walk
like a file-name match, so dataset branches that share one saver prefix
show their own prompt, labelled "Recorded run". Still open at one sampler
whose PromptManager text was recorded at execution (a language model's
output), the recorded text shows as "Execution recorded". Results the file
resolves alone never change; a failed lookup judges the file alone.
* fix(inspector): recorded texts follow history's saver; conflicts anchor nothing (#386)
Review follow-ups: the texts PromptManager recorded are looked up at the
sampler history's saver leads to, never at a sampler the unanchored walk
named; a file name naming a different saver than history anchors nothing;
outputs and texts are read in one transaction; "Recorded run" ranks below
every workflow-inferred source; unexpected lookup errors log at error level.
* fix(history): look up recorded runs by history's own spelling of the file (#386)
The history route inspects a case-normalized target on Windows, while
history matches file names exactly, so the occurrence's reference names
the file.
* refactor(inspector): share the inspector's resolve as inspection_resolution
inspect_prompt_evidence now calls inspection_resolution, which returns the
resolution to use, the anchored result and whether the recorded saver chose
it. Ingestion can call the same helper for files the default profile cannot
decide (#247).
* feat(ingestion): one per-file prompt decision for the scan and the monitor
utils/file_prompt.py holds decide_file_prompt, the ladder the output scan
runs today plus the inspector's rule (inspection profile, then the recorded
saver) before the legacy parser, with a SCAN and a MONITOR policy. Junk
text (blank, numbers, LoRA tags only, model file names) is never stored.
graph_prompt_text and parse_comfyui_prompt move here; the parser accepts
parsed metadata (#247).
* feat(ingestion): store a file's prompt in one transaction, one prompt per file
database/scanned_prompts.py links a file to the prompt of one text, creating
it when new, inside one BEGIN IMMEDIATE transaction: a file linked to
another prompt is a conflict, a failed link leaves no prompt behind and two
writers store one prompt. pair_ingestion gains file_link_rows,
link_conflict and Marks/mark_created; ingest_resolved_pair takes optional
marks for the components it creates. recorded_runs and named_prompt in
run_attribution become public (#247).
* feat(scan): decide each scanned file with the shared ladder (#247)
_ingest_scanned_file calls decide_file_prompt (scan policy) and stores the
result: pairs through ingest_resolved_pair and positives through
store_scanned_prompt, both with the scan marks. Files the default profile
leaves undecided now get the inspector's resolve, with the saver recorded
history names, before the legacy parser, so each branch of a shared-prefix
dataset run links its own prompt. Junk text is never stored, a file linked
to another prompt is a conflict instead of a second link, and a failed link
leaves no prompt behind. _BatchHistory reads the stored prompt and the
recorded saver from one recorded_runs pass. _parse_comfyui_prompt calls the
moved parse_comfyui_prompt. TESTING.md documents the decision.
* fix(ingestion): keep the file-name saver anchor out of the ladder (#390)
The #390 ruling queues a file that only its saver's file name resolves for
Needs review, never linking it. inspection_resolution gains name_anchor;
the ladder passes False, so such a file falls to the legacy floor in the
scan (and to Needs review) and is never stored by the monitor, while the
recorded saver still decides. The inspector keeps the name anchor.
* feat(config): gallery.monitoring.save_foreign_runs, default on (#247)
* fix(ingestion): no joined PromptManager text from the legacy floor; junk keeps its step's status
The legacy parser returns prepend + text + append, so its answer is refused
when a PromptManager-family node has a prepend or append part (#343). Junk
found at the undecided-graph step keeps that step's unlinked status and
review, so an ambiguous file is still queued for Needs review. The monitor
policy's ambiguous file no longer asks for the recorded run, and the file
size may be read lazily.
* fix(ingestion): keep the legacy answer that is a PromptManager node's own text
The joined-text refusal also refused the #170 parity files whose
PromptManager prepend is a link the browser parser does not follow: its
answer is then the node's literal text, which is what the node stores.
Refuse only an answer that is not such a node's own text.
* feat(monitor): store the certain prompt of a foreign workflow's output (#247)
A foreign output (no positive PromptManager node) is decided by the shared
ladder with the MONITOR policy. Certain files are stored at settle with
RUN_MARKS through survive_switch; files only recorded history can decide
wait in their own bounded set and are re-checked once after the run's
outputs are sealed, with the sealed reference. The setting
GalleryConfig.SAVE_FOREIGN_RUNS is read for every file.
* fix(scan): one prompt per file for recorded-run links; writes survive a library move
The recorded run's stored prompt now links in one transaction with the file's
link rows, so a file linked to another prompt is a conflict instead of a
second link. The scan's stores run through survive_library_switch. A single
file asks history once (a one-file _BatchHistory), the scan passes
in_output=False and reads the file size only for files with prompt
metadata. A conflict row reads "Kept: already linked to a different
prompt". TESTING.md records the rules.
* feat(settings): save_foreign_runs through POST /settings, applied live (#247)
GET /settings reports settings.save_foreign_runs; a save takes a strict
bool, writes gallery.monitoring.save_foreign_runs and publishes it to
GalleryConfig in the commit, with no restart field. The diagnostics
report lists it. The legacy contract snapshot gains the added key (an
addition, no legacy key moves).
* feat(system): Save prompts from other workflows in Sources & scans (#247, D147)
A host setting row below Watch for new outputs drafts save_foreign_runs
and saves with the one Save row; the server applies it at once, so no
restart is named. While monitoring is drafted off the row is disabled and
says it applies while Watch for new outputs is on. The concept comparison
maps the inserted row and its mutation check covers it.
* test(comfyui): a workflow of ComfyUI's own nodes creates and links its prompt (#247)
The real-host attribution spec gains the foreign case (CLIPTextEncode,
KSampler, SaveImage): its prompt is created and linked at settle and a
second run links to the same prompt. The fixture route lists the waiting
foreign files. TESTING.md records the monitor's foreign path and D147.
* test(vendor): the execution matrix counts the #247 real-host case
* fix(monitor): read a foreign file's size only when the ladder needs it (#247)
Uses PR 1's lazy file size and pins that a foreign file linked to another
prompt keeps its one link on the pair and the positive store paths.
* fix(ingestion): match a file's link rows under its resolved spelling too
The scan stores realpath spellings and the monitor the watcher's raw path,
so a file reached through a symlinked or junction output folder could get a
second generated_images row. file_link_rows now queries the path and its
realpath (each normalized and raw) and dedupes by row id. TESTING.md notes
the reverse order is matched only when the raw spelling is the realpath.
* fix(monitor): wait for the run unless history answered; monitoring off stops foreign writes (#247)
Review fixes: at settle a certain answer ends the wait only when the
lookup found a record or the answer is the inspector's or the recorded
saver's; the legacy floor waits for the re-check. A record without a
certain prompt ends the wait at once. Foreign writes need both
save_foreign_runs and monitoring_enabled, read again just before the
store. Dropped foreign files are logged outside the lock (evictions as
evictions, due ones too), and a drained re-check never drops a newer
reference.
* fix(system): the foreign row can be set while the monitor still watches; its note describes it (#247, D147)
Review fixes: while monitoring is saved off but a source still reads
watching until restart (D128), the row stays enabled with its note; the
note is the switch's aria-describedby. The browser spec restores
save_foreign_runs on the shared server. TESTING.md moves D147 into order
and records the monitor's review rules.
* test(comfyui): each branch of a shared-prefix foreign run links to its own prompt (#247 CUJ 1)
* fix(ingestion): a switched-off PromptManager node behind a pass-through stores nothing (#90)
The #90 graph walk follows only prompt-like inputs, so a PromptManagerText
with Save to library off reaching the encoder through another input name
(easy showAnything.anything, a custom node's value) was invisible to it,
and the inspector's resolve stored its text. The ladder now checks the
positive every step names: a resolved or inspected positive whose path
passes a node that is off, or a legacy answer that is such a node's own
text, makes the file not_saved.
* fix(ingestion): any switched-off node upstream of a positive stores nothing (#90)
The ladder walks every input upstream of each sampler or guider positive
(not negative, model, clip, latent or image inputs) for PromptManager nodes
whose Save to library is off or wired from an unreadable node, so the
sampler-less step, an off node with a prepend part and a PromptManagerPair
behind a pass-through no longer store its text. The legacy-floor guard reads
a node's resolved text. Every guard reports not_saved for a switch that is
off and not_saved_wired for one it cannot read.
* fix(ingestion): the switch walk skips negative inputs by #90's own rule
The upstream walk treated only names containing "negative" as negative, so
an off node on a neg, neg_text or text_neg input blocked an enabled
positive. prompt_graph gains is_negative_input (the _NEGATIVE_KEY rule, also
used by follows_input) and the walk uses it. The guards accept a node with
null inputs and a side whose provenance is not an object.
* fix(ingestion): the switch walk also skips any input named with "negative"
#90's rule matches only whole underscore-separated words, so negative2,
negativeprompt and NegativePrompt counted as positive inputs and an off
node wired only to one of them blocked the file. The walk skips those too;
follows_input is unchanged.
* fix(monitor): a foreign not_saved decision logs its skip and never waits (#247, #90)
The foreign path stored nothing for a ladder not_saved decision, but logged
it as "No certain prompt" and, when only the legacy answer was a
switched-off node's text, kept the file waiting for a seal that could only
store nothing. Such a decision now ends the wait at settle, and both settle
and the re-check log the opted-out path's own line (a wired switch says so).
Tests cover an only-PromptManager node switched off (opted out, not
foreign), a switched-off PromptManagerText behind a pass-through into a
core CLIPTextEncode, a wired switch, a negative-only switched-off node,
recorded history holding the off text, and foreign saving off.
* fix(monitor): an undecided not_saved waits for its run's record (#247, #90)
A not_saved from the ladder's undecided step is not final before a record
exists: the recorded saver may still resolve another branch. The settle
lookup now keeps such a file waiting; with the record it stores nothing and
logs the opted-out line. An early not_saved (a switched-off node upstream of
any sampler's positive) never reaches the lookup and stays final.
Tests: the undecided case (literal and wired switch) waits, then skips; a
switched-off node feeding one branch of a shared-prefix run never waits and
stores nothing, not even the branch the recorded saver names.
* docs(system): renumber the foreign-runs switch deviation to D158 (#247)
#430 took D147 on main; the Save prompts from other workflows row is D158.
* test(monitor): a switched-off node in the recorded saver's branch is never foreign (#247, #90)
PR 3's repair skips a file whose recorded saver's branch reads a
switched-off PromptManager node. On the monitor the same shape never reaches
the foreign path: a literal switch is the opted-out path's and a wired one
the node's own run's. Nothing is stored at settle or after the run, even
from a recorded run holding the off text.
The archived prompt page's notice always said "This prompt and its
negative variants are hidden", also for a saved-pair prompt with none.
It now says the prompt is archived because it is in a saved pair, and
counts its negative variants only when there are any. From the live
test of #470.
Live test B1: the Folder list indented a server list sorted as full-path
text, so "-" and "_" sorted before "/" and children landed under the wrong
parent (kiko/char under kiko-v6), with case variants interleaved. The
options are now a tree built by segment: each folder directly under its own
parent, case variants distinct siblings.
On POSIX a folder named "Metadata\Klein..." was listed but filtered to no
prompts, because normalize_folder read "\" as a separator. It now does so
only on Windows (ntpath), as the listing already did.
Refs #444
A prompt in a saved pair with no negative variants was described in the
Delete… review as "1 prompt with negative variants: hidden, restorable".
Paired prompts are archived because of the pair, so the Archive row now
reads "N prompts in a saved pair: hidden, restorable" and the first
sentence explains the pair. Found in the live test of #470.
The Library stores text.strip() while a run keeps its exact raw text. The
history window, run detail scope id, newest pair run and prompt hints now
compare the raw text's stored form (one rule, utils/hashing.py, mirrored in
SQLite with the same code points). Hints also keep matching the exact raw
text, so older rows that kept outer whitespace stay hinted and protected.
* feat(vocabulary): tag search, tag sort and an Untagged prompts entry (#444)
Tags & categories gains a server-side tag search and sort (keyset pages,
cursor scoped to the query) and an "Untagged prompts" row whose Browse opens
the Library with tag_mode=untagged. Refine this view gets Tag match, saved
Library views keep it, and the routes echo the applied query so a stale
server is an error, never an unfiltered list. Recorded as D149.
Refs #444
* test(saved-views): give the D153 Library view the tag_mode key (#444)
The Library saved-view state now holds tag_mode, and the API requires
every key, so main's new "Highest rated" spec posts it too.
Refs #444
Keeps both features: #444's folder filter, tag match (all/any/untagged),
tag search and sort with #443's archive/delete and Show archived shelf.
The ranked tag sort and the Untagged prompts count now hide archived
parents, as the other tag counts and the shelf already do.
* fix(inspector): give image tags their own labelled section (#449)
The inspector drew a file's tags as a bare chip row right under the negative
prompt, so they read as part of it. The side panel, the viewer details and
History's current file now draw them in labelled sections in the prompt
sections' frame: "Image tags · N" for the file's own tags, then "Prompt tags
· N" or "Recipe tags · N" for the linked tags. Each count is true: repeats are
removed within a source only. Recorded as D154.
* feat(inspect): choosing a file opens the viewer with its details (#448)
On Inspect a file, choosing or dropping a file now opens the shared viewer
on it at once, details showing, instead of previewing it and waiting for
"Open file metadata". Details show on this page even when "Collapse viewer
details by default" is on. An unreadable file or an AVI/WMV note stays on
the page; the heading action reopens any selection. A new selection that
starts while the viewer shows the previous file closes it first, so its
blob URL is never revoked under an open element. Recorded as D155.
* feat(viewer): full-screen toggle beside the close button, and a System default (#450)
Every viewer gets a button before the close button that switches between
Centered (today's framed dialog) and Full screen (the whole browser window,
edge to edge); it names the view it switches to. In full screen the details
stay toggleable. System > Appearance gains a Viewer surface with "Open viewer
in: Centered / Full screen" and the moved "Collapse viewer details by
default"; the choice is a browser display preference (viewerOpenMode). The
button changes only the current view and hides at 700px and below, where the
viewer is already full-bleed. Recorded as D156.
* fix(inspect): open a video's viewer only once its preview settles (#448)
Replacing a video while both the inline preview and the auto-opened viewer's
video were still starting deadlocked WebKit (D62). The viewer now opens for a
video once its inline preview reports metadata or an error (or after 15 s).
* test(viewer): pin the D97 view-switch spec to the desktop width (#450)
* test(inspect): cover the AVI note and an open dialog holding the auto-open back (#448)
Frontend coverage for inspect.mjs fell below its floor (branches 91.04 < 92,
functions 92.85 < 93): no spec reached the AVI toast or the open-dialog guard,
and the video wait had a page-end callback no spec reached. The wait now ends
through its own callback at page end, and two specs cover the AVI note and a
selection landing under another dialog.
The shelf laid out each loaded page only in its next animation frame. A
browser that holds frames (a background tab, headless WebKit under CPU
load, seen holding them for more than 5 s) left the grid empty while the
caption already showed the count. A loaded batch is now laid out at once
when the grid is in the document; measurements still flush per frame.
The regression test holds every animation frame and fails before the fix
on all four projects.
* fix(inspector): give image tags their own labelled section (#449)
The inspector drew a file's tags as a bare chip row right under the negative
prompt, so they read as part of it. The side panel, the viewer details and
History's current file now draw them in labelled sections in the prompt
sections' frame: "Image tags · N" for the file's own tags, then "Prompt tags
· N" or "Recipe tags · N" for the linked tags. Each count is true: repeats are
removed within a source only. Recorded as D154.
* feat(inspect): choosing a file opens the viewer with its details (#448)
On Inspect a file, choosing or dropping a file now opens the shared viewer
on it at once, details showing, instead of previewing it and waiting for
"Open file metadata". Details show on this page even when "Collapse viewer
details by default" is on. An unreadable file or an AVI/WMV note stays on
the page; the heading action reopens any selection. A new selection that
starts while the viewer shows the previous file closes it first, so its
blob URL is never revoked under an open element. Recorded as D155.
* fix(inspect): open a video's viewer only once its preview settles (#448)
Replacing a video while both the inline preview and the auto-opened viewer's
video were still starting deadlocked WebKit (D62). The viewer now opens for a
video once its inline preview reports metadata or an error (or after 15 s).
* test(inspect): cover the AVI note and an open dialog holding the auto-open back (#448)
Frontend coverage for inspect.mjs fell below its floor (branches 91.04 < 92,
functions 92.85 < 93): no spec reached the AVI toast or the open-dialog guard,
and the video wait had a page-end callback no spec reached. The wait now ends
through its own callback at page end, and two specs cover the AVI note and a
selection landing under another dialog.
* test(comfyui): the Inspect sender spec takes the viewer choosing a file opens (#448)
The real-host spec chose a file and then pressed "Open file metadata"; with
D155 the viewer is already open, so the press waited behind the modal until
the test timed out in the merge queue (classic-start and nodes2-start).
The M1 test let the rename answer at once and changed the hash only after the
toast, so on a loaded host the rename's 300 ms count read could start before
Fresh's page read; the page then waited on that one read and the poll for two
held reads failed. Hold the rename's answer until Fresh's read is held.
The star is only for favorites (maintainer ruling 2026-10-28). The workspace
no longer shows or edits a prompt rating: no Rating field in the Library
editor, no stars on Library cards, no "Highest rated" sort on Images or the
Library, and no rating column or rated-images total in System > Statistics.
The editor never sends `rating`, so a workspace save keeps a stored rating
as it is. A saved view or Library visit with the retired rating sort opens on
Recently added. Data and API are unchanged; legacy pages keep their ratings.
Recorded as deviation D153 (amends D14, D23, D96); concept comparisons map
the concept through it.
A starved renderer (headless WebKit under CPU load, a background tab) can
hold requestAnimationFrame for seconds. The shelf laid out each loaded page
only in its next frame, so the grid stayed empty after the first load or a
delete's reload while the caption already showed the new count. A loaded
batch is now laid out at once; measurements still flush per frame.
The regression test holds every animation frame and fails before the fix.
libraryViewState passed `?? 'all'` and `?? ''` fallbacks that
normalizeViewState already applies to an undefined value. The redundant
branches were never taken in the page, which dropped saved-views.mjs below
its 95% branch floor. A unit now covers a shelf without the new filters.
Refs #444
- A Default layout or Thumbnail size saved in System now reaches the next Images
visit in an open tab (this tab or another): the last visit records the
defaults it started from, and a value its toolbar never changed follows newer
defaults. A toolbar choice stays the current view's own.
- Images opens in Masonry when the browser has no stored Default layout
(gallery-settings.js DEFAULT_SETTINGS.workspaceLayout); stored choices are kept.
- Grid/List card art, Library card art, the inspector cover and the Tag image
preview crop with object-position 50% 20%, so portrait images keep their top.
- D157 in TESTING.md; concept comparisons start from the concept's stored Grid.
Vito Sansevero