Commit Graph
60 Commits
Author SHA1 Message Date
dotCooCoo 501dad41a1 Merge main into fix/file-folder-param-classification 2026-11-14 08:23:36 -08:00
dotCooCoo 6b9e7e9fa6 Automation: replies hide a file or missing folder on the way to the recovery folder
When a recovery copy fails because a folder on the way to the recovery
folder is a file or is missing, the error names that folder, not the
recovery folder. Replies to a client confined to read roots hid only
the recovery folder itself, so the full path of that ancestor stayed
visible. A reply now also hides each ancestor of a recovery folder that
it writes as a whole path. The check compares text only and does not
look anything up on disk.

Follows #7287.
2026-11-13 21:16:40 -08:00
dotCooCoo 395ddae55a Merge main into fix/file-folder-param-classification 2026-11-13 20:30:48 -08:00
BFlat c1efdbdaae Merge pull request #7255 from dotCooCoo/fix/folder-prefs-out-of-recorded-steps
Actions: keep folder preferences out of saved sets and confined replies
2026-11-14 08:40:15 +07:00
dotCooCoo c1572562de Merge main into fix/file-folder-param-classification 2026-11-13 15:55:47 -08:00
dotCooCoo 1f6c9beb8c Tests: recorded steps keep a path only where the command takes it
recorded_steps_filter_only_unreadable_locations expected
object.setLiveShape's `path` to stay, but that command does not take a
`path` param, so the recorded step reads the key as a file and drops a
path outside the read root. The case now expects that, and
text.createInPath, whose `path` is the id of a path in the document,
covers a key that names part of the document.
2026-11-13 15:46:41 -08:00
dotCooCoo ac10c51381 Data Recovery: a failed copy names the folder in use and how to change it
When a recovery copy could not be written, the status bar said only
"Couldn't save recovery data: <error>", and a copy that failed on the
background worker showed as "Saving recovery data for <title> failed:
<error>". Neither said which folder was in use or how to pick another.

The message keeps the error text (and the document's title when one
copy fails), then names the folder copies go to, says whether it is
the default recovery folder or the folder set in Preferences, and
names the setting that picks another folder: Preferences → File
Handling → Data Recovery → Folder (recoveryFolder). The recovery code
writes the same line to the log. The folder is the one in use when the
round started, so a preference changed while a copy runs does not
change the folder the message names.

A copy that fails now reports its own message, so the job's result is
Ok rather than an error, and the job no longer adds "<label> failed:
<error>" over it. A store the host installed (browser storage) has no
folder to name: its failed copy shows "Saving recovery data for
<title> failed: <error>", as a copy on the worker did before, also
when the copy was written without the worker.

- engine: recovery::folder returns the folder in use and whether the
  recoveryFolder preference names it; recovery::store uses it.
- Four @msg rows in each complete-message catalog, and in the bg, bs,
  da and id catalogs, which translate the old message.
- Control replies to a client confined to read roots show a recovery
  folder outside them, and the paths under it, by the folder's name, as
  they already show other paths outside the roots. That covers the
  folder in use and the last eight folders a failure message named.
- docs/target-app-parity.md: a line next to the Data Recovery Folder
  preference.
2026-11-13 12:51:12 -08:00
dotCooCoo 8d0e45ef52 Actions: keep empty path values in ui.inspect's copy of recorded steps
Under a read root, an empty or blank value under a path param, or in a
path param's list, names no path. ui.inspect now keeps it, so a recorded
prefs.set step that clears a folder preference shows the reset.
2026-11-13 12:15:49 -08:00
dotCooCoo 1ed7785bc5 Actions: keep folder preferences out of saved sets and confined replies
A prefs.set step, recorded inside a command.batch or loaded from a
.vcactions file, can hold a folder preference such as scratchPrimary.

- ui.inspect under automation read roots now checks such a step's
  value for {key, value} and each entry of values, also in a batch
  step's commands at any depth, as written and trimmed, including bare
  names, subfolders and file: URLs. An empty value and the scratch disk
  choices Startup and None stay.
- Any other step text written as an absolute path or a file: URL
  outside the read root is left out also when whitespace surrounds it.
- Save Actions refuses a set with a prefs.set step that holds a folder
  value and names the action and the preference instead of writing the
  file. An empty folder and the scratch disk choices save.
- prefs.export leaves out scratchPrimary and scratchSecondary, like the
  other folder preferences. prefs.import refuses a scratch disk set to a
  folder and still reads Startup, None and an empty value.

The folder list, the scratch disk choices and the folder check now live
in one place in the engine. Docs: docs/mcp.md, params text and a
progress line in docs/target-app-parity.md.
2026-11-13 11:37:49 -08:00
dotCooCoo 5e19c59f33 Automation: one classification of file and folder params
A param key names a file or a folder by a reviewed list of whole names,
read in the context of the command it is for (ParamsPlace): the
plug-in call check, the control server's reply hiding and recorded-step
filtering, and the sweeps' temp-folder redirection all use it.

- `url` (a web link) and `folders` (an export's grouping) no longer
  count as files or folders, and keys that only end like one
  (alignToPath, wholePaths, sortPaths, linksFolder and others) are no
  longer path params in control replies or recorded steps.
- `dir`, `directory` and the folder preferences that prefs.set takes in
  `values` (fontsFolder, pluginsFolder, recoveryFolder,
  templatesFolder) are folders everywhere, so a relative path under one
  outside the read root is taken out of a recorded step.
- A `command.batch` step, an `engine.execute` request and a recorded
  step keep their own command's exceptions below the top level, so the
  sweeps no longer replace a nested web link or export grouping.
- A new test sorts every catalog key that reads like a file or folder
  key into the reviewed lists and checks every classifier against them.
2026-11-13 09:14:58 -08:00
dotCooCoo 07522fb4e6 Tests: ui-egui tests remove the temp folders they make
Place, taskbar, opening, Save Options, Missing Fonts and one control
test left their folders in the system temp folder after every run.
Other opening and control tests removed theirs only at the end of a
passing test, so a failing one left its folder.

- A test-only Folder guard (crate::tests_folder) removes its folder
  with everything in it when dropped, also when the test panics, as
  the find-files tests' Tree guard does (#6948).
- The opening and control tests that made their own folders use it.
- Each Place test gets its own folder instead of one shared per-process
  folder, so one test's cleanup cannot remove another's files.
- The Missing Fonts helper no longer removes and remakes the folder
  vectorcraft_testkit::temp_dir made, which keeps it owner-only on Unix.
- vectorcraft_testkit::temp_dir is unchanged.
2026-11-13 07:14:08 -08:00
BFlat cea91c1374 Merge pull request #6968 from dotCooCoo/fix/followup-6910
MCP: a readable library's id keeps a hidden path that it holds after its start
2026-11-13 06:10:58 +07:00
dotCooCoo 2c3f61851f MCP: a readable library's id stays only after a delimiter
A readable library's id was kept whole after any character outside a
short list, so in `/a+/project/libs/Kit.gpl` the id
`/project/libs/Kit.gpl` was kept and the hidden `/libs/Kit.gpl` inside
it showed in the reply. The same happened after `@`, `=`, `%` and `!`.

The id now stays only at the start of the text, after a space, a
quote, `,`, `;` or an opening bracket, or in a local `file:` URL.
Written with its `loaded/`, it also stays after `=`.
2026-11-12 13:31:33 -08:00
dotCooCoo 264a6541e6 Merge main into fix/dialog-set-hidden-fields 2026-11-12 12:55:47 -08:00
dotCooCoo cbfe5248bb MCP: a readable library's id keeps a hidden path that it holds after its start
A reply hid a hidden library's absolute path wherever it appeared, also
inside another loaded library's id when the match began after the id's
start. With `/libs/Kit.gpl` hidden, the readable id
`loaded//project/libs/Kit.gpl` came back as `loaded//projectKit.gpl`,
and a request that used it failed.

Each whole readable id in a reply (with or without `loaded/`, or in a
`file:` URL with no host or `localhost`) is now found first, and a
hidden id or path that lies inside it stays. The same path after a
drive or in a URL with another host is still hidden.
2026-11-12 12:29:20 -08:00
BFlat b7062daa6b Merge pull request #6951 from dotCooCoo/fix/proof-view-unchanged
Color management: proof view commands redraw only when the view changes
2026-11-13 03:28:34 +07:00
dotCooCoo 344d756bf6 Color management: proof view commands redraw only when the view changes
view.proofSetup, view.proofColors, view.overprintPreview and
view.separationsPreview stored the proof view and raised every open
document's revision on each run. They did this also when the request
asked for the view that was already showing, for example Proof Colors on
while it is on, or Proof Setup with no params. Each such run redrew every
canvas. When the active document's revision changes, the canvas also
sends a tool that asks for it a new hover event, although the pointer has
not moved.

- The four commands compare the new proof view with the current one. When
  the two are equal, the commands leave the view and the revisions as
  they are. Separations Preview's visible plates are compared as a set,
  so the same plates in another order do not redraw.
- view.separationsPreview with `plates` naming a plate twice shows it
  once, so one `toggle` of that plate hides it.
- color.loadProfile with a new file of the CMYK profile that Proof Setup
  targets gives the documents new revisions, as it does for a working
  space. Before, running Proof Setup again redrew the documents through
  the new file; it now leaves an unchanged view alone.
- color.removeProfile gives the documents new revisions only when Proof
  Setup targets the removed profile.
2026-11-12 10:32:02 -08:00
dotCooCoo c8a6b4d9c3 MCP: a base64 thumbnail in a reply is sent as it is
Under automation read roots, a reply's base64 `thumbnail` (a library
item's PNG in `library.get`, a page's PNG in `document.pdfInfo`) was
searched for hidden library paths like any other text, so a short path
that happened to occur in the encoded bytes was rewritten and the image
broke. `thumbnail` now joins `dataBase64`, `pngBase64` and
`thumbnailBase64` as encoded data that is left unchanged; text under it
that is not base64 is still searched.
2026-11-12 09:12:40 -08:00
dotCooCoo 89e52eaabc MCP: a folder preference is checked as trimmed, and only a %XX escape is decoded
Under automation read roots, `ui.inspect` kept a folder preference whose
stored value was readable even when the trimmed path the app uses was
not: ` /outside/fonts` reads as a name in the working folder, while the
app uses `/outside/fonts`. A folder preference now stays only when both
forms are readable. A reply string is decoded for path matching only when
it holds a `%XX` escape, so a string such as `width="100%"` is no longer
decoded and thrown away. The library list line in the progress doc now
states when a hidden library's `loaded/<file name>` id moves.
2026-11-12 08:50:46 -08:00
dotCooCoo b4ffd4a97e MCP: a hidden path's file: URL shows as the file name, and rewritten keys keep every field
Under automation read roots, a reply replaced only the path inside a
`file:///outside/a.svg` URL, which left `file://a.svg`, a URL that names a
host. The whole URL of an absolute path, with `localhost`, another host or
no host, is now replaced with the file name, unless part of its scheme was
already replaced. A reply key rewritten to read the same as another key of
its object took that key's place and its field was lost; the rewritten key
now takes ` (2)`, ` (3)` and so on, and a key that is not rewritten keeps
its name.
2026-11-12 07:17:50 -08:00
dotCooCoo 2ec278280a MCP: ui.dock targets, shortcut keys and native backslash paths keep hidden library ids
- ui.dock's move takes a hidden panel's shown id in target, as the other
  docking destinations do.
- Shortcut overrides keyed panel:<panel id>, saved and in an open
  Keyboard Shortcuts dialog, show the panel id shown for a library
  outside the read roots, also when nothing else refers to it.
- Outside Windows, a native path is matched only as written: a sent
  /out/weird/name.svg no longer renames a hidden library at
  /out/weird\name.svg, and a reply does not read one as a spelling of the
  other. Windows paths are still matched with either separator.
2026-11-12 07:17:48 -08:00
dotCooCoo d4bbce8982 MCP: hidden library names follow the platform's path rules
- Outside Windows a `\` stays part of a hidden library's or sent path's
  file name; another platform's absolute path is still split on `\`.
- A hidden relative path is matched after leading `./` and `../` parts
  only when they start the path, so `/work/./data/icon.svg` keeps its
  text.
- `ui.inspect` and `ui.docking` replies are hidden once, so a library
  at the top of the file system shows as `loaded/<file name>` with its
  separator.
2026-11-12 07:17:47 -08:00
dotCooCoo 97b77fe4fd MCP: another platform's absolute path is outside the read roots everywhere
Outside Windows, a path that starts with a drive (`C:`) or a `\` is read
as a name in the working folder, so with the working folder inside a
read root it passed the read check. ui.inspect then kept it in recorded
action steps and replies did not hide it when a request sent it. The
check now counts such a path as outside the read roots wherever the
control channel and the library commands check a path.
2026-11-12 07:17:46 -08:00
dotCooCoo 6d7a8249a7 MCP: hidden library paths are matched before the paths a request sent
A sent path whose text overlapped part of a hidden library's path could
keep that path from being hidden; the hidden libraries' ids and paths
are now matched first. A hidden relative path is also hidden where a
reply shows the absolute path it reads as from the working folder, and
a hidden library sent by that other form shows as the sent path does.
2026-11-12 07:17:46 -08:00
dotCooCoo d1e766173e MCP: a library saved with another platform's absolute path is hidden under a read root
Outside Windows, a library path that starts with a drive or a backslash counts as outside the read roots, since the file access check would read it as a name in the working folder.
2026-11-12 07:17:46 -08:00
dotCooCoo 62be2df058 MCP: replies hide each sent path the same way whatever the outcome
Each hidden id or path is replaced once, chosen from the reply as given,
so a hidden path inside another library's shown id is not replaced
again, and a readable id written with %XX escapes keeps the hidden path
it begins with. A path quoted with {:?} is matched with its quotes and
control characters escaped. A path outside the read roots that the
request sent is shown by its file name in a reply that succeeded as in
one that failed, and a path param counts with a relative path too. A
relative path is not replaced where it ends a longer path. Without a
read root, a file: URL that names another host stays in ui.inspect's
recorded steps.
2026-11-12 07:17:46 -08:00
dotCooCoo f6cabe62cc MCP: a readable library id written with LOADED/ keeps its text on Windows
The loaded/ prefix is matched in any case where the ids and paths are.
2026-11-12 07:17:46 -08:00
dotCooCoo 13bee0d6ef MCP: a readable library id keeps its text when a space follows the hidden part
A hidden id or path stays only where it begins another loaded library's whole id with no `.` or `..` part, also when that id has a space after the hidden part.
2026-11-12 07:17:46 -08:00
dotCooCoo 38184805b4 MCP: a failed request's error hides a file URL it sent that names another host
The error shows the URL's file name, as it does for a local path the request sent outside the read roots.
2026-11-12 07:17:46 -08:00
dotCooCoo 00f5a16494 MCP: replies hide library paths outside the read roots in more forms
- Every request under a read root records the hidden libraries before it
  runs, and the reply hides those and the ones hidden after it. A panic
  is answered inside the request, so its error is hidden too.
- A hidden path followed by more path text stays only when the longer
  text is another loaded library's id with no `.` or `..` part.
- Encoded file data is skipped only for a base64 string under
  `dataBase64`, `pngBase64` or `thumbnailBase64`; objects and lists under
  those keys, and other keys ending in `Base64`, are searched.
- The error of a failed request hides a path outside the read roots that
  the request sent by its file name, so it reads the same whether or not
  a library was loaded from that path.
- Paths written with `%XX` escapes are hidden, and on Windows paths in
  another letter case.
- Recorded steps in `ui.inspect` leave out `~`, paths under `~` and paths
  starting with `./` or `../` outside the read roots under any param.
- Hidden libraries are numbered in load order and keep their id from one
  request to the next.
2026-11-12 07:17:46 -08:00
dotCooCoo a2c08dec27 MCP: replies hide a library path outside the read roots wherever it appears
Under the read roots, every control reply now shows a hidden library's
id as loaded/<file name> and its path as the file name in every string
and key: status text, dialogs, the command palette query, a script's
report, recorded steps and errors. The id and path are replaced at any
boundary, so a path between spaces, brackets or quotes, followed by a
period or by more path text (<path>.bak), or quoted with escaped
backslashes is hidden too. An occurrence that continues into a longer
path stays only when the request may read that longer path, so a
readable library's id that begins with a hidden one keeps its text.
The libraries hidden after the request ran are hidden, and for a
request that names one, those hidden before it too. A reply with no
path separator in it is not searched. Encoded file data (dataBase64,
pngBase64) is sent unchanged.

Recorded steps in ui.inspect leave out a path outside the read roots
under path, paths, file, files, folder and any param whose name ends in
Path, Paths, File, Files or Folder (diffPath, fontFiles,
destinationFolder), and an absolute path or file: URL outside the roots
under any other param or in any list. The file: scheme is matched in
any case, a URL's %XX escapes are decoded before the check, a file: URL
inside the roots stays, and a file: URL that
names another host is always left out. file_access::is_path_key holds the list of
path params, and testkit's is_path_key uses it.
2026-11-12 07:17:45 -08:00
bflatastic 0f65f6d3f4 Revert "Merge pull request #6656 from dotCooCoo/fix/inspect-library-ids-whole"
This reverts commit 3ed0dc64a8, reversing
changes made to 70ba41bb2b.
2026-11-12 16:35:41 +07:00
dotCooCoo e6f330afd2 Automation: ui.dialog.set cannot change the command a dialog runs
ui.dialog.set took every field whose name starts with `__`, so a control
request (or MCP's answer_dialog, which goes through it) could rewrite the
`__command` and `__params` of a confirmation or the generic parameter
dialog, a plug-in command dialog's `__command`, or the paste a swatch
conflict waits to finish, after a person opened the dialog. The person's
OK then ran something else.

A `__` field is now settable only when it is a control of that kind of
dialog: a section or tab shown, a link between two fields, a name or
search being typed (for example Save PDF's `__section`, Transform Each's
`__constrain`, the Preferences `__search`). Every other `__` field, held
or not, is refused with "the `<kind>` dialog cannot set protected field
`<field>`". The request sets no field, and with `action: "confirm"` OK
does not run. Flags a button sets together with other fields are refused
too: the style dialogs' `__reset*`, Artboard Options' `__fit`,
`__sizeChoice`, `__refx` and `__refy`, and Scribble's `__scribbleSetting`.

Recolor's Merge takes a row listed twice in the selection once, and
Randomize keeps a seed at u64::MAX instead of overflowing it.

Docs: docs/control-protocol.md, docs/mcp.md, docs/target-app-parity.md.
2026-11-11 23:59:18 -08:00
dotCooCoo 55570c10d1 Merge main into fix/inspect-library-ids-whole 2026-11-11 23:31:03 -08:00
dotCooCoo c9ddbfe7cb MCP: ui.dock targets, shortcut keys and native backslash paths keep hidden library ids
- ui.dock's move takes a hidden panel's shown id in target, as the other
  docking destinations do.
- Shortcut overrides keyed panel:<panel id>, saved and in an open
  Keyboard Shortcuts dialog, show the panel id shown for a library
  outside the read roots, also when nothing else refers to it.
- Outside Windows, a native path is matched only as written: a sent
  /out/weird/name.svg no longer renames a hidden library at
  /out/weird\name.svg, and a reply does not read one as a spelling of the
  other. Windows paths are still matched with either separator.
2026-11-11 22:42:59 -08:00
dotCooCoo c74821031e MCP: hidden library names follow the platform's path rules
- Outside Windows a `\` stays part of a hidden library's or sent path's
  file name; another platform's absolute path is still split on `\`.
- A hidden relative path is matched after leading `./` and `../` parts
  only when they start the path, so `/work/./data/icon.svg` keeps its
  text.
- `ui.inspect` and `ui.docking` replies are hidden once, so a library
  at the top of the file system shows as `loaded/<file name>` with its
  separator.
2026-11-11 21:56:02 -08:00
dotCooCoo e83c36785e Commands: pointer events refuse pen values and spaces of the wrong type
ui.pointer and the headless pointer call read a pointer event's pressure,
holdMs, tiltX, tiltY and rotation with a default for any value that was not a
number, and treated any space other than screen as document space. Each event
is now checked before the packet runs: those values must be numbers and space
must be doc or screen, with null keeping the default. docs/mcp.md says so.
2026-11-11 21:04:01 -08:00
dotCooCoo d5044de8b1 MCP: another platform's absolute path is outside the read roots everywhere
Outside Windows, a path that starts with a drive (`C:`) or a `\` is read
as a name in the working folder, so with the working folder inside a
read root it passed the read check. ui.inspect then kept it in recorded
action steps and replies did not hide it when a request sent it. The
check now counts such a path as outside the read roots wherever the
control channel and the library commands check a path.
2026-11-11 19:59:09 -08:00
dotCooCoo 7c28138ce3 Commands: pointer packets are read whole before any event runs
A headless or control channel pointer packet now checks every event's kind,
coordinates and mods before it sends any of them, so a malformed later event
no longer leaves the earlier ones applied. A mods value that is not an object,
such as an array, is refused. Rotate View takes null mods as omitted. The MCP
docs name the params that refuse null.
2026-11-11 19:57:39 -08:00
dotCooCoo 9d9cb52017 MCP: hidden library paths are matched before the paths a request sent
A sent path whose text overlapped part of a hidden library's path could
keep that path from being hidden; the hidden libraries' ids and paths
are now matched first. A hidden relative path is also hidden where a
reply shows the absolute path it reads as from the working folder, and
a hidden library sent by that other form shows as the sent path does.
2026-11-11 19:44:43 -08:00
dotCooCoo 3e0d5a0aaf MCP: a library saved with another platform's absolute path is hidden under a read root
Outside Windows, a library path that starts with a drive or a backslash counts as outside the read roots, since the file access check would read it as a name in the working folder.
2026-11-11 18:59:31 -08:00
dotCooCoo 68cbf041a4 MCP: replies hide each sent path the same way whatever the outcome
Each hidden id or path is replaced once, chosen from the reply as given,
so a hidden path inside another library's shown id is not replaced
again, and a readable id written with %XX escapes keeps the hidden path
it begins with. A path quoted with {:?} is matched with its quotes and
control characters escaped. A path outside the read roots that the
request sent is shown by its file name in a reply that succeeded as in
one that failed, and a path param counts with a relative path too. A
relative path is not replaced where it ends a longer path. Without a
read root, a file: URL that names another host stays in ui.inspect's
recorded steps.
2026-11-11 18:50:00 -08:00
dotCooCoo a8989aff6d Commands: app and control channel params of the wrong type are refused
view.setZoom, view.screenMode, window.panel, file.revert, file.close,
ui.newSwatch, ui.symbolOptions, ui.batchDialog and file.exportSelection
read a number or switch of the wrong type as its default and went on, so
`view.setZoom {"zoom": "200"}` set 100%. They now refuse it and change
nothing.

The control channel's ui.key, ui.move, ui.wheel, ui.click, ui.drag,
ui.pointer, ui.resize, ui.screenshot, ui.render and ui.set, and the
headless session's key, pointer and render calls, do the same for their
numbers, switches and mods, before any input is queued. Omitting a param
or giving null keeps its default.
2026-11-11 18:22:01 -08:00
dotCooCoo 386c444ff4 MCP: a readable library id written with LOADED/ keeps its text on Windows
The loaded/ prefix is matched in any case where the ids and paths are.
2026-11-11 17:31:25 -08:00
dotCooCoo 027035668b MCP: a readable library id keeps its text when a space follows the hidden part
A hidden id or path stays only where it begins another loaded library's whole id with no `.` or `..` part, also when that id has a space after the hidden part.
2026-11-11 17:07:30 -08:00
dotCooCoo cce2a685bc MCP: a failed request's error hides a file URL it sent that names another host
The error shows the URL's file name, as it does for a local path the request sent outside the read roots.
2026-11-11 16:43:49 -08:00
dotCooCoo db8726e94b MCP: replies hide library paths outside the read roots in more forms
- Every request under a read root records the hidden libraries before it
  runs, and the reply hides those and the ones hidden after it. A panic
  is answered inside the request, so its error is hidden too.
- A hidden path followed by more path text stays only when the longer
  text is another loaded library's id with no `.` or `..` part.
- Encoded file data is skipped only for a base64 string under
  `dataBase64`, `pngBase64` or `thumbnailBase64`; objects and lists under
  those keys, and other keys ending in `Base64`, are searched.
- The error of a failed request hides a path outside the read roots that
  the request sent by its file name, so it reads the same whether or not
  a library was loaded from that path.
- Paths written with `%XX` escapes are hidden, and on Windows paths in
  another letter case.
- Recorded steps in `ui.inspect` leave out `~`, paths under `~` and paths
  starting with `./` or `../` outside the read roots under any param.
- Hidden libraries are numbered in load order and keep their id from one
  request to the next.
2026-11-11 16:33:37 -08:00
dotCooCoo 6a7214df4c MCP: replies hide a library path outside the read roots wherever it appears
Under the read roots, every control reply now shows a hidden library's
id as loaded/<file name> and its path as the file name in every string
and key: status text, dialogs, the command palette query, a script's
report, recorded steps and errors. The id and path are replaced at any
boundary, so a path between spaces, brackets or quotes, followed by a
period or by more path text (<path>.bak), or quoted with escaped
backslashes is hidden too. An occurrence that continues into a longer
path stays only when the request may read that longer path, so a
readable library's id that begins with a hidden one keeps its text.
The libraries hidden after the request ran are hidden, and for a
request that names one, those hidden before it too. A reply with no
path separator in it is not searched. Encoded file data (dataBase64,
pngBase64) is sent unchanged.

Recorded steps in ui.inspect leave out a path outside the read roots
under path, paths, file, files, folder and any param whose name ends in
Path, Paths, File, Files or Folder (diffPath, fontFiles,
destinationFolder), and an absolute path or file: URL outside the roots
under any other param or in any list. The file: scheme is matched in
any case, a URL's %XX escapes are decoded before the check, a file: URL
inside the roots stays, and a file: URL that
names another host is always left out. file_access::is_path_key holds the list of
path params, and testkit's is_path_key uses it.
2026-11-11 14:23:06 -08:00
dotCooCoo d9e9adbf45 MCP: answer_dialog sets the fields and answers the dialog in one request
answer_dialog sent ui.dialog.set and then ui.dialog.confirm or
ui.dialog.cancel as two requests. A dialog another control connection
opened between them was confirmed or canceled without the fields.

ui.dialog.set takes `action: confirm | cancel`, checked before any field
is set, and confirms or cancels the dialog the fields were set in within
the same request, replying {dialog, result}. answer_dialog sends set and
action together; an app whose reply has no `dialog` key did not answer
it, and the dialog is then confirmed or canceled with its own request.
2026-11-10 14:51:50 -08:00
dotCooCoo fd3482c272 Merge main into fix/shape-builder-merge-off-ui-thread 2026-11-08 11:28:00 -08:00