Both forms Divi writes are recognized (ruling Q11): custom_meta_<key> and the
dialog form post_meta_key + select_meta_key. Choice fields get a note about the
stored value, relation fields name IDs (F4). Only templates the user may edit
are checked or named; others are counted (Q21 pattern). Warnings only, also in
the dry run of a live body. DynamicFields extracted from lit/divi-dynamic-sources.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
TB layouts load through ThemeBuilderAccess (P12), live layouts (attached +
enabled, Q8) need a dry run and confirm (Q9) and get the shared restore point
of VisibleEdits, extracted from GlobalEdits (Q2: backup when revisions are
off, P9 refusals, P14 naming). divi/post-content only in body layouts (F22),
also for nested stored nodes. SaveScope::ThemeBuilderLayout sets Divi flags
and clears site CSS + TB cache.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
One PostIdParser for {{link:}} and select_post; post_link_url_<type> needs a
living, readable, non-attachment target of exactly that type.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
StoredNodeGuard accepts a zero-padded library id (Divi intval()s it) but
refuses instances whose opener holds more than one or an escaped
globalModule, so the entry Divi renders is the one checked. GlobalUsages
matches "globalModule":"0*<id>" with REGEXP behind a cheap LIKE prefilter.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
lottie.innerContent (src → data-lottie-src) makes the frontend load an
animation whose expressions lottie-web runs with eval; it joins the S7 gate
in ValueCleaner, which covers removals and StoredNodeGuard.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
A link without scheme and without whitespace, quotes, <, >, backslashes or
control characters passes although esc_url_raw() would prefix http://; every
value with a scheme, and protocol-relative //host, must survive url() unchanged.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
URLs must be sanitizer-stable, keys inside values follow ValueCleaner::VALUE_KEY,
instances need a digits-only globalModule; state keys reuse AttrWriter::STATES.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Divi's frontend applies the interaction effect "Attribute" with setAttribute
(e.g. an onclick handler) and sets cookies; *.decoration.interactions joins
the S7 gate in ValueCleaner, which also covers removals and StoredNodeGuard.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Refuses, for users without unfiltered_html, non-Divi blocks, raw text parts,
unreadable opaque blocks, code modules, builder-only paths (also below the
state value), strings that are not sanitizer-stable and stored $variable
references that VariableRef::canonical() refuses or would change (P2, P3).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
A contact-field item whose type is no allowed child now gets a hint that
type is the node type, with the attrs path (fieldItem.advanced.type) and
the lit-divi-update set alternative, instead of "unbekanntes Modul
divi/email". The metadata fixture gains contact-form/contact-field
(regenerated from Divi 5.2.1, additions only).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01At3V6SjwbMe6eMzVGW3Hs7
New rows no longer accept 4_5 (the builder does not offer it). On update,
the flexType of every non-desktop breakpoint (tablet, phone, …) is
format-checked against the builder list like desktop. Guide updated.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01At3V6SjwbMe6eMzVGW3Hs7
Composer::autoLabel is private (tested through compose()); the
DiviComposition getters used only by tests are marked @internal;
Limits::snapshotCount is documented as reserved; the ShortformWriter test
seam is named $modulesForTests.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01At3V6SjwbMe6eMzVGW3Hs7
One test class per Divi write tool on a shared DiviWriteAbilityTestCase;
the 30 tests move verbatim (per-test assertion counts identical). Adds the
ledger T4 test for a TreeEditor::move() StructureRules refusal.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01At3V6SjwbMe6eMzVGW3Hs7
Composer::row uses ColumnWidths::parse; rowWidthsError and its error-state
save/restore are gone. The update width check moves out of
DiviUpdateAbility into ColumnWidths::updateErrors. Behaviour unchanged.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01At3V6SjwbMe6eMzVGW3Hs7
- HtmlTokens (static, with the S45 docblock): SPLIT, RAW_OPEN, tokenize(),
structure(), closes(), unterminated() moved unchanged; new RAW_ELEMENTS list.
- StoredForms: of() (was TextReplacer::forms), mentions(), unicodeEscaped() –
the DB prefilter forms; TextReplacer::mentions() delegates, the ability's
candidate query uses StoredForms::of().
- TextReplacer keeps problems(), apply() and the splice/safe() verification;
safe() still runs the tokenizer and HtmlContext as two independent scans, in
the same order.
HtmlTokensTest asserts RAW_ELEMENTS === HtmlContext::RAW and that RAW_OPEN opens
exactly those elements. TextReplacerTest and HtmlContextTest unchanged.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01At3V6SjwbMe6eMzVGW3Hs7
enum ValueKind: string { Html, Code, Url, Text } with strictness() and
tightens(self $routed) (the S33a rule formerly AttrWriter/ValueCleaner::tightens
over STRICTNESS). ValueCleaner::clean() takes ?ValueKind, kindOf()/kind()/
unknownKind() return it, the sanitizer match is exhaustive over the four cases;
the "Feldart" message prints ->value, so its text is unchanged. ShortformWriter
drops the identity map CLEAN_KIND for ValueKind::tryFrom($spec['kind']) ??
ValueKind::Text; Composer and ShortformWriter pass ValueKind::Text/Url.
The installed PHPCompatibility sniffs do not know enums: two sniffs are disabled
in ValueKind.php only, with the reason.
Tests: ValueCleanerTest passes/asserts ValueKind cases instead of strings;
+ValueKindTest (full tightens matrix, case values).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01At3V6SjwbMe6eMzVGW3Hs7
The choke point moves out of AttrWriter into ValueCleaner (catalog, the
composition's VariableRef, sanitizer): clean() (was AttrWriter::cleanValue),
kindOf(), guardPath() and the private cleanNode/kind/unknownKind/tightens with
VALUE_KEY, URL_KEY, URL_SEGMENT, STRICTNESS, BUILDER_ONLY (its docblock now sits
on it; STRUCTURE stays in AttrWriter). Code of the pipeline moved unchanged.
- ValueCleaner's docblock is the one canonical description: steps 0-8 with their
rulings in code order, invariants I1-I3; the step comments in cleanNode carry
the same numbers. AttrWriter/ShortformWriter/Composer/TextReplacer docblocks
point there and keep only their own rules.
- ValueCleaner::textReplacementFor($module) holds the D10 rule that was a
closure in DiviReplaceTextAbility::replace (text/html only, clean() by kind).
- AttrWriter, ShortformWriter and Composer take the ValueCleaner; DiviComposition
builds one per composition (cleaner()).
Tests: the eight AttrWriterTest methods that only exercised cleanValue/kindOf
moved verbatim to ValueCleanerTest (clean/kindOf, ComposeFixture::cleaner());
mixed tests call the cleaner in place; +1 test for textReplacementFor. Every
provider row and per-test assertion count unchanged.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01At3V6SjwbMe6eMzVGW3Hs7
- DiviComposition::references() builds the one VariableRef with the entry's
live map (S22/S31) and passes it to AttrWriter; Placeholders only needs the
design system now, Placeholders::references() is gone (no reach-through).
- DesignSystem::colorIdFor()/variableFor() replace the dead colorId()/variable()
and carry Placeholders' lookup incl. the S13a ambiguity check (byLabel moved,
messages unchanged); isColorVariable() decides color vs content for {{var:}}.
- DesignSystem::colorReferenceIds() (VariableRef color names) and
contentVariableIds() (DynamicContentRules design names) replace the two
inline computations; same lists, same order.
Tests: fixtures/tests get the VariableRef from ComposeFixture::references() or
build it with the map; DesignSystemTest uses the new lookups (+2 tests).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01At3V6SjwbMe6eMzVGW3Hs7
Without /D, "$" also matches before a trailing "\n", so "custom_meta_telefon\n",
a color setting "5\n" (stored as sent) or a whole reference followed by "\n"
passed VariableRef. Added /D to every ^...$ pattern that validates client or
admin input:
- VariableRef META_KEY, NUMBER, WHOLE (canonical/typeOf; Placeholders::field
shares META_KEY)
- Cursor::decode, GuideRepository topic names (MCP input and filter keys)
- ExposurePolicy::sanitize ability keys and group ids, LogTab list filters
Not changed: Placeholders::PATTERN (runs on trim()), Warnings hex checks (hints
only), NodeSummary/AttrPathIndex/LogListTable (stored or Divi data, no client
input). Each fixed pattern has a test with a trailing "\n" that is rejected.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01At3V6SjwbMe6eMzVGW3Hs7
- DiviFixtures: markup(), givenPost(), schema(), designSystem(), catalog(),
dynamicContentRegistry(), maliciousRefs() - moved verbatim out of
DiviPageLoaderTest, DesignSystemTest, ModuleCatalogTest, DynamicContentMapTest
and VariableRefTest (its malicious() provider now returns maliciousRefs()).
- ComposeFixture moves to tests/Support.
- DiviWriterFixture: modules(), designReader(), composition(), a recording saver(),
log() and writer(&$saved, &$failures, unfiltered, loader) for the five tests that
wired the Divi write stack by hand.
No test cases or assertions changed (1397 / 5834).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
TextReplacer replaces only in plain-text segments (S8 split), outside
script/style bodies, inserts the replacement cleaned per field kind and
refuses any result whose tags, entities or references differ from the
original plus the inserted text. Unchanged strings are never cleaned.
Per page all or nothing with the tree_rev lock; Throwables are reported
per page (S10).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01At3V6SjwbMe6eMzVGW3Hs7
DiviSaver/DiviWriter/DiviComposition/DiviChange, WpValueSanitizer (S4),
WpMediaResolver, DiviFlags (S11), DynamicContentMap from Divi's live
registry (S22), S31 wiring, S35 control strip at apply() entry.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
TreeEditor (position/insert/move/remove/duplicate on the BlockTree) and
Warnings (post-write quality checks), plus S18's outline breadth limit
on the 2a Outline/DiviOutlineAbility.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01At3V6SjwbMe6eMzVGW3Hs7
Control characters now go on the raw value in AttrWriter::cleanNode and
VariableRef::setting; after the sanitizer only rejection remains ('<' + VT/FF
no longer turns into a tag). VALUE_KEY anchored with D, exact column-width
fractions, nested rows in columns, specialty columns refused, fullwidth by
catalog category, preset ambiguity in DesignSystem::modulePresetId().
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01At3V6SjwbMe6eMzVGW3Hs7
Labels, background_color and preset ids go through the choke point (S13);
ambiguous preset names fail with their ids (S13a); fullwidth + rows is an
error. Carry-ins: control characters stripped before the marker check (S32),
kind override may only tighten, reference alt texts dropped, color check via
VariableRef::typeOf (S33).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01At3V6SjwbMe6eMzVGW3Hs7
ShortformWriter routes every short-form string through AttrWriter::cleanValue()
with the curated field kind (new optional kind override), exposes the color
check for background_color and guards builder-only paths on removal too.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01At3V6SjwbMe6eMzVGW3Hs7
Critical-1/S28: sanitizers that strip <x>, NUL or %XX could rebuild the
$variable( marker after the raw check; every cleaned string (AttrWriter
at every depth, VariableRef settings) is checked again. S29: custom_meta_
map names and *meta_key setting values obey META_KEY. Minor-2: malformed
dynamic content maps fail with InvalidArgumentException.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
S26: every css.* path is gated like S7 for users without unfiltered_html.
S27: Divi ignores settings on global variable and theme font references,
so any settings there are rejected.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
I-1: reject )$ in every decoded string setting and in the re-serialized
JSON. S21: string settings cleaned and re-serialized. S22: injectable
dynamic-content allow map (name => setting => field type), loop_* only via
the map. S23: unknown content.innerContent is rich text. S24: keys below
scalar kinds are unknown, ...Url path segments are URLs. S25: isVariable
removed. M-1: real \u escapes in tests, malicious suite as editor, fixture
html() raw for unfiltered_html. M-2: S7 gate at any depth plus
css.advanced.customCss.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
VariableRef parses, validates and re-serializes every $variable(...)$ value
(S1/S2/S12); AttrWriter routes strings by catalog kind incl. per-subName
composite kinds and key-name URL fallback (S3/S3a), rejects odd value keys
(S13) and Divi HTML/attribute fields without unfiltered_html (S7);
ambiguous design-system labels fail with their ids (S13a); R2 finding
recorded (S5).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
