* fix(forms,xfa): find every password field before a save skips scrubbing it
The quick Password-flag check in pdfcraft-forms stopped silently at its
depth limit, and visited the tree in an order where a field the full
field walk reaches could be cut off on a long branch first. It now
answers true (read every field) whenever it cannot look as far as the
field walk, and also walks below fields adopted from page widgets.
XFA password classification walked AcroForm fields only 32 levels deep,
shallower than editing and dataset synchronization, and read the
datasets packet of an XFA array without the prefixes its preamble
declares, so a typed password could stay in the saved datasets. It now
walks as deep as the field walk, refuses to clear (and so to save) when
the tree is deeper than it can check, and reads the packet the way the
datasets writer does.
* address review: check every page widget's /Parent chain for the password flag
* address review: password scan follows the parents of nameless /Fields entries; field-walk completeness checked before the datasets packet is read
* address review: the field walk's completeness is decided after every root; the subtree scan starts at the last dictionary of the parent chain
* address review: a skipped visit does not replace the reach a field was classified with
* fix(forms): a read-only radio button is not picked or left through set_value
* address review: use /V as the selection in the read-only radio check, keep reset going
* address review: radio read-only checks judge each button by the state it would get, and cover FieldValue::Widget
* AcroForm: integrate persistent timesheet script workflows
* AcroForm: address review caching, worker stacks and live metadata
* M6: address second review of AcroForm script workflows
* M6: handle top-level ASI boundaries and verify cyclic calculation results
* M6: bound script workers and isolate pending completions across edits
* M6: make form resets visible before subsequent script reads
* M6: include calculated dependents in reset comparison baselines
* M6: bound reset writeback work before native completion
* M6: charge expanded reset selections and field name bytes
* M6: bound string conversion bytes and charge reset widgets
* M6: preserve unread defaults and bound prompt text conversion
* M6: bound reset calls early and preserve calculated field identity
* AcroForm: address grouped calculations, field identity and widget ownership
* AcroForm: retire field instances and address native review edge cases
* AcroForm: route comparison Save As and Print to the source document
* M6: dispatch text widget mouse-up scripts
* M6: honor fit bounds for form button icons
* AcroForm: bound calculation work and preserve reset and save integrity
* Harden AcroForm script requests and worker recovery
* Keep form callbacks and focus consistent across edits
* AcroForm: preserve reset ordering and bound edit callbacks
* Fix reset appearance ordering and XFA focus after row removal
* Keep XFA focus indexes and stopped reset appearances consistent
* Document form script execution and reset helpers
* Keep script workers bounded and reject partial callback edits
* Keep form events available after a document script reaches VM limits
* Keep main's first-value Simple calculation and restore the calculateNow helper test
AFSimple_Calculate (JavaScript) and the native Simple calculation count each field's first
value again, as #5304 landed them; summing every selection of a multiple-selection list is
a separate behaviour change. Restores the document-helper calculateNow test and the
AFSimple_Calculate doc comment, and adds native and JavaScript multiple-selection cases.
* Patch the callback field snapshot in place instead of rebuilding it
Each field callback rebuilt all field states (3,129 on the benchmark form), and the
retained realm held the previous snapshot, so stamping instance tokens copied it again.
The runner now keeps its last snapshot and patches only the values and flags that
changed; ScriptSession::stamp_fields stamps it before dispatch, the worker copies only
when a token changes, and the realm drops the job's snapshot when the job ends (field
identity is checked against the cached keys instead).
* An oversized job script is that script's error, not a refused edit
Realm::run refused a script too long or too deep to run as an execution refusal, which
made the engine reject the whole edit (setting a 300 KiB calculation script failed). As on
main it is now a plain script error that keeps the value; the realm is not touched.
* Read a leading #! line as a comment in the chain scan; follow main's refusal and conditional limits in tests
The nesting guard from #5079 now checks brackets, prefixes and conditionals; the document
worker's preflight still bounds arrow-function and right-associative chains and skips a
leading #! line as the guard does. A refused script is an error, not a refused edit, so
the preflight test now checks the edit goes through with the error reported.
---------
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
* Fix#5211: persist JavaScript field colours
* Read current document state in field scripts
* Redraw form fields after validation changes
* Format validation redraw regression test
* Keep field list alive in redraw test
* Refresh fields only after same-field scripts
* Format conditional field refresh
* Borrow accepted field value for storage
* Redraw calculated fields from current appearance
* Format JavaScript field state merge
* Restore JavaScript runner document state
* Group JavaScript runner form inputs
* Read current field display state in scripts
---------
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
* Fix#5211: make calculateNow synchronous
* Bound and cache calculation execution
* fix(engine): invalidate field calculation cache
* perf(js): build calculation scripts lazily
* Refresh calculation cache after page insertion
* Load calculations for document script helpers
* Assert imported calculation result
* Initialize formatted field-change state
* Update formatted calculation test expectations
* Include the referenced field in calculation test
---------
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
crates/forms/src/values.rs: the read imports take both listed_indices and
selected_options (list selection) and main's listing_fields_named (radio
fields sharing a name).
Main's DataOp::Value now replaces a multi-select list's value-group picks in the
packet and in the data copy, so the reset writes through it and the separate
DataOp::Reset (and Want::picks) is dropped. The reset keeps the PR's indexed
targets and its pass over fields without widgets, beside main's floating-field
pass (which the unshown pass leaves embedded fields to), and finish() lays out
again for either an owed relayout or a changed embedded value. Parity notes and
evidence take both sides.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Pq7HyCHD1R3rywQBPCHXZA
Refs #2354. Thai typed into these tools is shaped and drawn from the Thai face through Add text's writer (now pdfcraft_fonts::shaped_text), searchable and copyable, instead of being written as ?????; a build without a Thai face refuses it with a message.
- fill: the first same-named field's read-only flag still applies when a later one draws the
choice, and no radio field of the name is written when any of them is read-only.
- enumerate: a nameless /Fields widget its parent already claimed is not read again as a nameless
field; one its parent's /Kids omits is still read on its own.
- walk: only kids whose /Subtype is /Widget replace a merged field's own widget.
- inspection: a kid whose /Parent is another field is named from that chain even under a direct
listing field, and widget annotations beside named kids are listed as a field of the parent's
name (ISO 32000-2 §12.7.4), as the form reader lists them.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Pq7HyCHD1R3rywQBPCHXZA
ISO 32000-2 §12.7.4.2: fields with the same fully qualified name have
the same value. Choosing a radio option by the name now writes the
choice as /V of a same-named twin that draws no button for it (its
buttons stay off), and of a field whose /Kids lists them without being
their /Parent, so it shares their name (issue15096.pdf's group, whose
/V was a stale Choice2). Exactly one button stays drawn on.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Pq7HyCHD1R3rywQBPCHXZA
A field's fully qualified name is built from the partial names of the field
and its ancestors, and a child names its parent with /Parent (ISO 32000-2
§12.7.4.2, Table 226). A /Kids entry with /T whose /Parent does not name the
field listing it is now named up its own /Parent chain, in the form reader
and in document inspection, as PDFium and MuPDF name it: pdf.js's
issue15096.pdf lists its two radio fields as Group1, not Group1.Group1.
Fields with the same fully qualified name are one field (§12.7.4.2), so a
radio option that a later one of them draws is now chosen by the name: it is
written through the first field that draws it and the others turn off.
Before, only the first field's options were accepted.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Pq7HyCHD1R3rywQBPCHXZA
* Forms: date fields name months in the date language, not always English (#5303)
`AFDate_FormatEx("dd-mmm-yy")` wrote "18-Oct-26" in an Italian interface,
where a localized Acrobat writes "18-ott-26". Form date formats,
`util.printd` and `util.scand` now take their month and weekday names
from the date language (Preferences > Date format > Language, which
follows the interface language until one is picked), and still read
English names typed in. The names are process-wide, so their test runs
in its own binary.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Forms: dates read back in every date language (#5303)
Trying every date language showed what a field wrote not always reading
back: names holding digits ("10月"), spaces ("tháng 10"), dots ("ม.ค.")
or combining vowel signs (Thai, Hindi, Telugu) fell apart word by word,
and Italian, Spanish, Romanian and Albanian share some short weekday and
month names ("mar"). The language's names are now found whole in the
text first, the longest at each place, and read as their English names;
the format's order of `mmm` and `ddd` settles a shared name. Both
AFDate parsing and util.scand do this, on text up to 1 KB. A test reads
back every language, each name format, and hostile text.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Forms: date names no field font can draw are drawn in English (#5303)
With the names in the date language, a Russian dd-mmm-yy field drawn in
Helvetica showed "18-????.-26": no /DR font held "окт." and the WinAnsi
stand-in drew each letter as ?. The same held for Japanese, Chinese,
Greek, Arabic, Thai and others, and for Latin names outside WinAnsi
(Czech "říj"). Before, the names were English and always drew. An
AFDate/AFTime field now formats its value again with English names for
a widget whose fonts can't draw the localized text; a custom Format
script is run again with util.printd writing English names for such a
widget, and its result kept only when it draws whole. A field whose
font holds the names keeps them, and scripts otherwise still print the
date language.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Forms: an English weekday before the month keeps the month (#5303)
With Italian dates, "Tue 17 Mar 2026" in a "ddd d mmm yyyy" field read
as no date: the language's names were found in the text, but not
English ones, so "Tue" left the format's weekday unfilled and "Mar",
which Italian shares between Tuesday and March, was taken as the
weekday. Both AFDate parsing and util.scand lost the month, and
Spanish, Romanian and Albanian ("mar") did the same. English names are
now found too and fill their field of the format as localized ones do.
A test reads English and mixed weekdays before the month in those
languages.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Forms: an English date run is drawn even when it prints the value (#5303)
A Helvetica field storing "18-Oct-26" whose Format script prints it
again as dd-mmm-yy still showed "18-???.-26" in Russian. The English
run printed "18-Oct-26", but it was dropped for equalling the stored
value, which was how a failed run was told apart: a failing script
handed back the value unchanged. The English run now says whether it
ran (`None` when the script failed or the runner can't write English
dates), and what it made is drawn whenever it draws whole. A test
reprints the stored value under Russian, and checks that a failing
script keeps its text and is reported undrawn.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Forms: a short weekday typed without its dot is the weekday (#5303)
With French dates, "mar 6 janv 2026" in a "ddd d mmm yyyy" field read
as no date through util.scand, and "mar 06/01/2026" in "ddd dd/mm/yyyy"
through AFDate too. French writes Tuesday as "mar.", which a name
search only found with its dot; since English names are found as well,
the English "Mar" (March) was the only name of that length, so the
weekday became a month. Romanian "mar" did the same. A short name
ending in a dot is now also found typed without it, as the same kind of
name, so the format's order of name fields decides between it and an
equal name of the other kind. Both copies of the name search change. A
test reads French and Romanian weekdays typed without their dot.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Forms: a shared name is the month when no weekday is typed (#5303)
With Italian dates, "17 mar 2026" in a "ddd d mmm yyyy" field read as
no date, nor did "mar 17, 2026" in "dddd, mmmm d, yyyy". AFDate leaves
weekdays out of what it reads (they are shown, not typed), but the name
search still expected the format's weekday first, so "mar", which
Italian shares between Tuesday and March, filled it and left no month.
Spanish and Albanian "mar" and Albanian "sht" (Saturday, September) did
the same, and the keystroke check refused those dates. Where such a
read finds no date, the names are now read again with no weekday
expected, so the shared name is the month; a typed weekday still comes
first ("mar 17 mar 2026"). util.scand reads its fields in order and
needs the weekday, so it is unchanged. A test types those dates with the
weekday left out.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Forms: Turkish names typed in capitals are read (#5303)
With Turkish dates, "1 MAYIS 2026" and "1 EKİM 2026" in a d mmmm yyyy
field read as no date through AFDate and util.scand, and "SALI 6 OCAK
2026" through util.scand. Names were compared lower-cased, which turns
I into i, not the dotless ı of "Mayıs" and "Salı", and İ into i with a
combining dot above, which "Ekim" doesn't have. Names and typed text
are now folded alike, the capital I and the dotted and dotless i all as
i and the combining dot dropped, one piece for each char so positions
in the text stay aligned: in both copies of the name search, the AFDate
month reader and util.scand's name matcher. English reads as before. A
test types Turkish names in capitals and English ones as a check.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Forms: Greek names typed in capitals are read again (#5303)
Folding names char by char turned a capital Σ into σ, so "ΙΑΝΟΥΆΡΙΟΣ"
no longer matched "Ιανουάριος", whose last letter is the final ς. The
fold now makes ς a σ on both sides, in the forms and JS parsers alike.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
* M6: forms decodes values, renamed calculations and detected pages within shared budgets
Part of #5755: the forms crate's four unbudgeted stream reads go through a
`DecodeBudget`, so an operation's decoding is bounded however often the file
names the same stream.
- A reading of the form decodes its fields' value streams (/V, /DV) within one
64 MiB budget, as it already does their scripts. Each field got a fresh
1 GiB allowance, so 300 fields naming one stream decoded it 300 times.
- Renaming a field rewrites the calculations that read it within one script
budget; past it, a calculation is left as it is, like one that won't decode.
- Field detection reads a page's content within one page's allowance
(`DecodeBudget::page`), as the content walks do, however often /Contents
lists a stream.
- The PDF/A declaration is read from at most 8 MiB of XMP, as preflight's
Standards panel reads it, and only once /NeedAppearances isn't already set.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AmsdeUCdTaD72jkDw2pJKY
* M6: the forms README says which metadata declares PDF/A
declares_pdfa reads pdfaid:part from XMP metadata that decodes within
8 MiB, as preflight's Standards panel reads it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AmsdeUCdTaD72jkDw2pJKY
* M6: a test pins that metadata without pdfaid:part declares no PDF/A
cargo-mutants found `==` turned to `!=` in declares_pdfa's search unpinned:
no case had metadata within the cap that declares nothing.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AmsdeUCdTaD72jkDw2pJKY
* M6: field detection reads operators one at a time; the PDF/A check reads XMP as preflight does
From the Resources review of 66415d737:
- page_shapes collected every parsed operator, about 80 bytes for each
decoded byte of dense content, so a page within its 256 MiB allowance
could take some 20 GB. It now reads them one at a time
(pdfcraft_content::ops) and drops the pieces once joined.
- declares_pdfa decoded through a DecodeBudget, which counts every filter
stage against the 8 MiB cap, so XMP stored through a chain of filters
stopped declaring PDF/A while preflight's Standards panel, which caps
each stage, still read it. It reads the metadata as preflight does.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AmsdeUCdTaD72jkDw2pJKY
* M6: say what declares PDF/A to forms, and why its metadata read stays direct
From the Legal-forms review of c00e3d2fb: the README and doc comments said
forms reads the declaration "as preflight's Standards panel reads it". The
two share the 8 MiB cap, applied to each filter stage, but forms matches the
bytes pdfaid:part and preflight the identifier by its namespace, as on main.
The sweep's direct-/DecodeParms hit is answered at the site: preflight and
organize read metadata the same way, and the readers move together.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AmsdeUCdTaD72jkDw2pJKY
* forms: rename a shared calculation once; keep unread defaults on Clear form; cap q in detection
- A Calculate action that several fields share by reference is read and
rewritten once, in place, so every calculation reads the new name and
the script is decoded once, not skipped past the budget.
- Clear form leaves a field alone when its /DV is a stream the value
budget didn't reach, instead of emptying it.
- Field detection keeps at most 256 saved graphics states; deeper q are
counted so each Q still restores its own.
- The dense-content test checks that the budget is spent; exhausted() was
true from the start with no form visits.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* forms: restore the matrix past the save cap; reset a value that wasn't read
- Field detection keeps the matrix of the first save past MAX_SAVES and
restores it when the last of them closes. A shape painted after a Q
that would have undone a cm inside those saves is dropped, as the
matrix there isn't known, instead of being reported scaled.
- Clear form skips a field only when its value is known: one whose /V
is a stream the value budget didn't reach is reset, not kept as if it
equalled an empty default.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* forms: keep detection's saved states in a Saves type
shapes_within went past clippy's cognitive complexity limit (32/30) with
the save-cap tracking inline. The same logic now lives in Saves (save,
restore, moved, unknown).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: FredApps <153173854+FredApps@users.noreply.github.com>
The click queued the clicked list box's own option positions under the
shared field name, but set_value reads ChoiceIndices in the first field of
that name. A twin listing options that share an export value in another
order therefore gave the first field the option at the same position, not
the one clicked. The UI now maps the twin's picks to the first field's
positions by export value and text (Field::positions_in, the same matching
the engine uses to copy /I to twins), falling back to the export values
when an option has no match there.
(review)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Pq7HyCfwzyRXj4VfQAVVdv
The reset of objects an opening script added before they have widgets
missed three kinds of value, and its name matching grew with the product
of objects and names (review):
- Exclusion groups were skipped, so a radio value set and reset in a new
row stayed in the data. A group now resets to the on value of its
default button, or to nothing.
- A choice list reset to the text it shows, not the item value it saves.
The default now goes through Field::saved_value, as the live form
reads it.
- A multi-select list whose picks the data holds as <value> elements
compared as empty and could not be written, so its picks survived. A
new DataOp::Reset compares and replaces those picks (none for an empty
default); DataOp::Value still leaves nodes with children alone.
- Reset names are indexed (ResetTargets), so matching an object costs
the same however many names a script passes, and forms::reset looks
named fields up through sets instead of scanning the list per field.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Pq7HyCfwzyRXj4VfQAVVdv
- A same-name twin gets the copied /I mapped to its own options by export
value and text, so options listed in another order keep the option picked;
one with no match there leaves the twin selecting from /V, never part of /I.
- A position picked twice is written once in /V and /I, so they still match
item for item.
- Shift-click extends from the row clicked last, kept in the view: with rows
sharing an export value the value can't say which came last.
- A single-select list loaded from an array /V writes one value when its
shown row is clicked again.
(review)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Pq7HyCfwzyRXj4VfQAVVdv
Options are tracked by position as well as export value, so two options
sharing an export value (ISO 32000-2 Table 234, /I) show and pick the right
row; a single-select list given several values holds and shows the first;
/I matches /V item for item, repeats included; values match options in
linear time.
Rebased onto main as one commit: main split the forms crate's lib.rs into
read.rs and values.rs and moved forms_ui.rs into tools/. Folds the branch's
four commits.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Pq7HyCfwzyRXj4VfQAVVdv
A check box field with kids whose on states are A and B could only ever
tick A. With A ticked, clicking B cleared the field, so B's tick flashed
and vanished, and with nothing ticked, clicking B ticked A. The click sent a
field-wide Check(!on), and write_value always turned on the first kid's on
state, even for a value naming B.
§12.7.5.2.3 makes the field's /V the on state of the kid that is on (/AS
deciding where they differ), and Table 230 lets kids be checked
independently.
- A click on a kid that is off sends its on-state name: it (and any kid
sharing that state) turns on and the others off. A click on a kid that is
on clears the field. "On" is the kid's /AS (Field::widget_on).
- write_value: a named state or /Opt export ticks the kids with that state.
Check(true) keeps the kid that is on, else takes the first.
- Field::check_box_value maps a stored value to the state it names. Reset to
/DV, script value changes and FDF/XFDF import use it.
- The display-only fallback ticks only the kids that are on.
- form_fields lists a check box's options and the checked one when its kids
have more than one on state. form_fill takes the name.
Kids sharing an on state still tick together (EXAMPLE 2).
Tests: forms unit, ui-egui end-to-end clicks, automation form_fill and
xfdf import. Each fails with its fix reverted.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0175FWj4xFhMhSH3UpMU8DSC
Adopt upstream's split set_props and explicit import lists. The list box
scroll after a property edit moves into scroll_list, keeping set_props under
the new cognitive_complexity limit.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SqmHYx8FiWM5PcsoiErgp9
A widget's own /TI is read before the field's, so each one scrolls by its
own top row and size (ListRows::top_showing). Before, the widgets were only
moved, all to the first widget's new row, when the first widget itself
scrolled: a second widget scrolled past the remaining options, or away from
a new value the first already showed, kept its stale row.
Test each_widget_with_its_own_top_row_scrolls_by_itself covers both
set_props (options shrink) and set_value; restoring the first-widget gate
in either caller makes it fail.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SqmHYx8FiWM5PcsoiErgp9
/V is inheritable like /DV (ISO 32000-2 §12.7.4.1). Clearing a text or
choice field's value removed its own /V, so a parent field's /V showed
through: Reset Form on a kid with no default refilled it with the parent's
value. When an ancestor sets /V, an empty value is now written instead of
removing the key. Check boxes and radio buttons already write /Off.
Test clearing_a_kids_value_does_not_bring_back_its_parents covers a text
kid and a list box kid through reset and through set_value; each half of
the fix was switched off in turn and the test failed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SqmHYx8FiWM5PcsoiErgp9
Changing a list box's options could leave /TI (the first visible row)
pointing at a row that moved or no longer exists. A widget's own /TI is
read before the field's, and one past the end draws no options at all.
set_props now runs the same ListRows::top_showing rule set_value uses
(ISO 32000-2 12.7.5.4): /TI stays put while the selected rows are still
visible, otherwise it scrolls to them, clamped to the last full page.
The field's /TI is updated, and so is every widget that carries its own.
The scroll helpers move out of set_value so both paths share them.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SqmHYx8FiWM5PcsoiErgp9