Apache Solr 10 runs on a dedicated VPS behind a Caddy HTTPS proxy
(vitecsolr.evomedien.de) because the managed webserver only allows
outgoing standard ports. Credentials stay out of git: the site config
carries %env()% placeholders resolved via putenv() in the git-ignored
config/system/additional.php.
- composer: apache-solr-for-typo3/solr 14.0.0-RC1 (the only line
compatible with TYPO3 14; final 14.0.0 will arrive via composer update)
- config.yaml: read connection https/443, core_en per language,
env placeholder credentials; .gitignore covers additional.php
- setup.typoscript: config.index_enable = 1 (page indexing silently
refuses without it), solr_pi_results JSON renderer registration,
results highlighting, and content field extraction from tt_content
rows - the headless JSON output has no TYPO3SEARCH markers, so the
default page content extraction indexed an empty content field
- SearchJsonRenderer (renderer #27): JSON output for the search plugin
on /search. GET q/page in; { query, page, resultsPerPage, numFound,
totalPages, results[], suggestions } out. Disables the page cache per
request: config.no_cache is gone in TYPO3 v14, and q/page are
excluded from cHash, so cached variants would collide
- ext_localconf.php: q and page added to cacheHash excludedParameters
- SolrIndexCommand (vitec:solr-index): works the index queue from the
CLI with connection diagnostics and a --debug single-step mode -
EXT:solr 14 ships no console commands and the backend button indexes
one item per click
VITEC Import module: Products tab is now a searchable/sortable product
overview. "Edit Product" opens an XLSX upload with per-field source
mapping (component + cell), old/new preview and checkbox apply via
DataHandler; the mapping and manual matches are persisted and preselect
the next workbook. Category workbooks are detected and rejected.
- new: ProductXlsxReader (PhpSpreadsheet), ProductTextImportController,
Products/ProductTexts templates, 4 module routes
- related products: per-pair card text in new side table
tx_vitec_product_related_text (survives MM rewrites), emitted as
`cardtext` in the product JSON; missing MM relations added add-only
- card copy read from Body Copy (D) with fallback to CTA/Card Copy (E) -
the workbooks fill either depending on row type
- product detail page <title> now uses seotitle with title fallback
(provider made singleton and fed from the JSON renderer)
- per-user recent-search badges; last loaded workbook stored per product
(tx_vitec_product_workbook), Edit Product reopens on it
- composer: add phpoffice/phpspreadsheet ^5.9
- diagnostics: migrations/check_workbook.php
- vitec_usecaselist: layout + record selection like marketlist
- new plugin vitec_solutionlist (spec 7.13.3, key "solutions")
- tx_vitec_domain_model_usecase: detail_page -> resolved detailUrl
(also gives story cards in vitec_modelcard a link for the first time)
- "Show Toolbar" checkbox on product/market/solution/usecase lists
- spec bumped to v1.9
BREAKING: vitec_productlist and vitec_usecaselist now emit an object
instead of a bare array. Consumers must read products.products and
usecases.usecases. Also found: productlist had a configurable layout
that was never serialised; it is emitted now, but keeps its own 0-3
vocabulary instead of grid/list/carousel.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Closes two annex items that turned out to be the same defect seen from two
sides: markets could link to a detail page, solutions could not, and the code
that turns a page uid into a URL existed four times over.
detail_page for Solution (B-11)
- ext_tables.sql, TCA (group on pages, maxitems 1, in showitem behind slug)
and Domain\Model\Solution mirror the Market field exactly
- SolutionShowJsonRenderer emits detailUrl
- ModelcardJsonRenderer serialized no detailUrl at all: its shared
market|solution branch never carried the field. Market modelcards therefore
gain a link here too, not just solutions.
Service\LinkResolver (B-12)
- one implementation replaces four private copies (Market list/show, Usecase
list/show). The copies had drifted: Market returned null on failure, Usecase
an empty string. The contract is now explicit -- null means "no link".
- Usecase keeps a thin resolvePageUrl() wrapper appending ?? '', because it
concatenates paths and a null would tear the strings apart.
- LocationsJsonRenderer and NewsJsonRenderer stay out on purpose: they resolve
a full parameter construct resp. slug paths plus canonical, which is not
page-uid-to-URL and does not fit the same contract.
Requires on the server, the column exists in code only:
vendor/bin/typo3 database:updateschema "*.add,*.change"
vendor/bin/typo3 cache:flush
New "VITEC Import" module under Web: one import page per domain model
(v1: Market, Solution, Product - registry-driven, adding a model is one
config entry). Workflow: upload a CSV (delimiter and encoding are
auto-detected, including German-Excel semicolon/Windows-1252), map CSV
columns to DB fields, persist the mapping per model together with the
identity field used for matching (new table tx_vitec_import_mapping,
no TCA - pure tool configuration), review a unified list of CSV rows
matched against the DB records (new / update with differing fields /
unchanged / db-only), then apply the checked rows through DataHandler.
Each importable row carries an editable JSON payload textarea - what
is written is the textarea content, not the raw CSV, so editors can
fix values right in the review step. The parsed CSV travels through
the form as a hidden JSON field: no session state, no temp files.
Importable fields are derived from TCA at runtime (scalar types only;
files, categories and other relations are excluded - a flat CSV
cannot carry them). Payloads are whitelisted against that field list
on apply; new records require a storage pid (prefilled from existing
records). BE user permissions apply via DataHandler.
The module ships its own CSS (backend-import.css, loaded only by
this module) using the frontend button palette from _vitec.scss:
orange #f47937 for primary actions, navy #26358c for secondary
actions and structure. The stray <h2>Hi</h2> debug leftover in the
shared backend layout is removed (also affects the OG Image module).
A fourth tab "SEO Research" handles the recurring keyword-research
CSV. It is deliberately not an import mask - the file carries research
only (no meta title/description yet). Each upload is persisted as a
delivery (tx_vitec_seo_research, never deleted) and evaluated: diff
against the previous delivery keyed by URL, a structure check of the
CSV tree against TYPO3 (pages by slug path; market/solution/product
rows against the domain tables, matched by slug then normalized
title), and the three work lists from the SEO flags (quick wins by
GSC impressions, shared terms grouped by keyword, already ranking).
CsvReader now deduplicates repeated header names, which that CSV has.
New CLI command vitec:create-markets: creates the market records the
structure check reports as missing, sourced from the stored delivery
and matched through the same SeoResearchService - what the module
lists is what the command creates. Each market gets a sys_category of
the same title, found anywhere under the auto-detected market category
root or created; sub-market categories are created under the parent
market's category, so the category tree carries the hierarchy the
flat market model cannot. Idempotent, dry-run first.
New CLI commands vitec:create-markets and vitec:market-dummy-image.
create-markets creates the market records the structure check reports
as missing (root detection three-staged: option, auto-detect, find or
create a "Markets" category). market-dummy-image assigns a shared
placeholder (white logo on brand navy, fileadmin/placeholders/) to
every market without an image - one sys_file for all, replacing the
file restyles every placeholder at once. Both idempotent.
Deliberately out of v1: import log with three-way compare (protection
against overwriting manual edits), images/relations, multiple saved
mappings per model.
New "VITEC Import" module under Web: one import page per domain model
(v1: Market, Solution, Product - registry-driven, adding a model is one
config entry). Workflow: upload a CSV (delimiter and encoding are
auto-detected, including German-Excel semicolon/Windows-1252), map CSV
columns to DB fields, persist the mapping per model together with the
identity field used for matching (new table tx_vitec_import_mapping,
no TCA - pure tool configuration), review a unified list of CSV rows
matched against the DB records (new / update with differing fields /
unchanged / db-only), then apply the checked rows through DataHandler.
Each importable row carries an editable JSON payload textarea - what
is written is the textarea content, not the raw CSV, so editors can
fix values right in the review step. The parsed CSV travels through
the form as a hidden JSON field: no session state, no temp files.
Importable fields are derived from TCA at runtime (scalar types only;
files, categories and other relations are excluded - a flat CSV
cannot carry them). Payloads are whitelisted against that field list
on apply; new records require a storage pid (prefilled from existing
records). BE user permissions apply via DataHandler.
The module ships its own CSS (backend-import.css, loaded only by
this module) using the frontend button palette from _vitec.scss:
orange #f47937 for primary actions, navy #26358c for secondary
actions and structure. The stray <h2>Hi</h2> debug leftover in the
shared backend layout is removed (also affects the OG Image module).
Deliberately out of v1: import log with three-way compare (protection
against overwriting manual edits), images/relations, multiple saved
mappings per model.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
New "VITEC Import" module under Web: one import page per domain model
(v1: Market, Solution, Product - registry-driven, adding a model is one
config entry). Workflow: upload a CSV (delimiter and encoding are
auto-detected, including German-Excel semicolon/Windows-1252), map CSV
columns to DB fields, persist the mapping per model together with the
identity field used for matching (new table tx_vitec_import_mapping,
no TCA - pure tool configuration), review a unified list of CSV rows
matched against the DB records (new / update with differing fields /
unchanged / db-only), then apply the checked rows through DataHandler.
Each importable row carries an editable JSON payload textarea - what
is written is the textarea content, not the raw CSV, so editors can
fix values right in the review step. The parsed CSV travels through
the form as a hidden JSON field: no session state, no temp files.
Importable fields are derived from TCA at runtime (scalar types only;
files, categories and other relations are excluded - a flat CSV
cannot carry them). Payloads are whitelisted against that field list
on apply; new records require a storage pid (prefilled from existing
records). BE user permissions apply via DataHandler.
Deliberately out of v1: import log with three-way compare (protection
against overwriting manual edits), images/relations, multiple saved
mappings per model.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
New CLI command vitec:import-downloads: reads the JSON export of the
old site (www.vitec.com/import), cuts it out of the surrounding page
template by brace counting, and imports the downloads idempotently by
slug. Scope is deliberately narrow: only files under
/fileadmin/downloads/Collateral/, records without title or file are
skipped, records sharing one file_url are merged. Files are fetched
resumably into fileadmin/downloads/Collateral/ and stored on pid 29.
Legacy filenames are normalized on fetch so every imported file
matches the <prefix>__<filetype>__<NN>-<letter> version convention:
__NN_A -> __NN-A, ___NN -> __NN, __NNA -> __NN-A, and a bare __NN
gets -A appended as its initial revision. Verified against the full
export: all 179 names match afterwards, none of the already-correct
ones change.
tx_vitec_domain_model_download gains a second category field "type"
(display taxonomy: Datasheet, Success Story, Brochure, Software) next
to the filename-driven filetype categories. No SQL change needed -
the core generates columns for category fields. The three download
renderers emit the field as a string analogous to filetype; MM rows
are distinguished by fieldname, existing queries filter on it and
remain unaffected.
Import values are matched via the custom columns sys_category.filetype
and .type (falling back to title); missing categories are created
(filetype under --category-parent, default 4; type under a "Download
Type" parent).
getDownloadFile() in the three download renderers gains a filepath
fallback (FAL -> convention -> filepath) as a safety net for any
future filename that escapes the convention - previously such records
emitted file: null.
Architecture spec bumped to v1.7.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Resolves a cascade of TYPO3 v14 breaking changes that left every VITEC
plugin's JSON output silently empty:
1. `list_type` column dropped — all renderer page-discovery queries now
filter by `CType = 'vitec_X'`.
2. `#[AsAllowedCallable]` attribute added to every public render() —
without it v14 throws AllowedCallableException, which headless
silently filters via its "Oops, an error occurred!" guard.
3. `$GLOBALS['TSFE']->id` is null inside JSON cObj context — page id
now read from the `frontend.page.information` request attribute
(TSFE fallback kept for legacy entry points).
4. page.tsconfig wizard items migrated to `CType = vitec_X` directly
(legacy `CType=list, list_type=...` no longer exists in v14).
5. ContainerChildrenProcessor + ContentElementResolver dispatch the
PLUGIN_RENDERERS map by CType (primary) with list_type fallback.
Beyond the bug fix this commit also contains:
- Datasheets renderer rewritten to "products with newest datasheet"
semantics (filtered via Download.hideondatasheets).
- New vitec/card content block — multi-purpose card with backend
preview and layout/background/aspect/alignment/border/shadow options.
- New vitec_container CType (b13 single-column container, FlexForm
cssClass propagated into the JSON envelope).
- ContentElementResolver service for contentelement/contentelementcta
link resolution; ContainerChildrenProcessor for nested plugin
resolution inside b13 containers.
- Vitecset setup.typoscript: ContentElement import moved to top so
lib.contentElement is defined before VITEC's tt_content blocks.
- Cleanup: 98 .bak.<timestamp> debugging backups removed; *.bak.* and
/public/_assets_install/ added to .gitignore.
Resolves a cascade of TYPO3 v14 breaking changes that left every VITEC
plugin's JSON output silently empty:
1. `list_type` column dropped — all renderer page-discovery queries now
filter by `CType = 'vitec_X'`.
2. `#[AsAllowedCallable]` attribute added to every public render() —
without it v14 throws AllowedCallableException, which headless
silently filters via its "Oops, an error occurred!" guard.
3. `$GLOBALS['TSFE']->id` is null inside JSON cObj context — page id
now read from the `frontend.page.information` request attribute
(TSFE fallback kept for legacy entry points).
4. page.tsconfig wizard items migrated to `CType = vitec_X` directly
(legacy `CType=list, list_type=...` no longer exists in v14).
5. ContainerChildrenProcessor + ContentElementResolver dispatch the
PLUGIN_RENDERERS map by CType (primary) with list_type fallback.
Beyond the bug fix this commit also contains:
- Datasheets renderer rewritten to "products with newest datasheet"
semantics (filtered via Download.hideondatasheets).
- New vitec/card content block — multi-purpose card with backend
preview and layout/background/aspect/alignment/border/shadow options.
- New vitec_container CType (b13 single-column container, FlexForm
cssClass propagated into the JSON envelope).
- ContentElementResolver service for contentelement/contentelementcta
link resolution; ContainerChildrenProcessor for nested plugin
resolution inside b13 containers.
- Vitecset setup.typoscript: ContentElement import moved to top so
lib.contentElement is defined before VITEC's tt_content blocks.
- Cleanup: 98 .bak.<timestamp> debugging backups removed; *.bak.* and
/public/_assets_install/ added to .gitignore.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- New Classes/Service/ContentElementResolver.php: parses TYPO3 typolink
(t3://record?identifier=tt_content&uid=N, t3://page?uid=P#N, plain numeric)
to a normalised tt_content envelope ({id,type,colPos,sorting,appearance,
data}), recursively resolving any nested VITEC list-plugin children.
- ProductListJsonRenderer / ProductShowJsonRenderer: contentelement and
contentelementcta now emit the resolved object instead of the raw
typolink string (breaking change at those two keys).
- Product model: new field videofile (FAL, single video upload in tab
Images and Videos, mp4/webm/ogv/mov/m4v); new bool field showdatapath
(toggle at end of General tab, controls the "Datapath is now Vitec"
graphic in the frontend). ext_tables.sql, TCA, model property +
getter/setter, and both renderers updated accordingly.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
