Help for Developers

What changed in the source code most recently: the main branch of se7enxweb/exponential, read from its public feed by this server.

Hide merge commits

1. Updated: A code block keeps its line breaks, tabs and indentation whe… (opens in a new window) · 11/05/2026 03:39 pm

   Message

   Updated: A code block keeps its line breaks, tabs and indentation when the online editor saves it
   
   eZXMLInputParser::process() removes every line break of HTML input (they are no content in HTML) and makes
   every tab a space, before it parses. Inside , which becomes , the white space is the content:
   every code block was stored as one long line, its indentation left as runs of spaces in the middle of it.
   This hit every block an editor saved in the online editor and every page an import wrote through it.
   
   preformattedWhitespace() now writes the line breaks of a  block (LF, CRLF, CR) as , which the
   literal handler stores as line breaks, and its tabs as 	, which stay tabs. The line break right after the
   opening tag is no content, as in HTML. Input without  is unchanged, and text outside code blocks loses
   its line breaks as before.
   
   Tests without a database: lines and indentation, empty lines, tabs, leading spaces, CRLF and CR,  from
   the editor, entities, the block's class, text outside code blocks, and a block opened in the editor and
   saved again unchanged.

   6bb303b456
2. Updated: cond gives its last parameter as given without template comp… (opens in a new window) · 11/05/2026 12:39 pm

   Message

   Updated: cond gives its last parameter as given without template compilation, also an empty array, hash, 0, '' or false
   
   The processing interpreter took the value for "no condition holds" only when it was true, so cond( ..., array() )
   gave no value with TemplateCompile=disabled, while a compiled template gave the empty list. The missing value
   contains everything, which put every link of the admin4 left menu and the Setup overview into the first group
   without GroupLinks_[] (pull request #357). The same form is used in about 45 places of the admin, admin4
   and standard designs; they now get the empty value in both modes.
   
   - The parity test renders cond with an empty array, empty hash, 0, false, an empty string and later clauses,
     processed and compiled, with literal and variable input (12 of its new cases failed processed before).
   - The template language page (English and German) describes cond's last parameter; the 6.0.15 changelog names
     the menu fix and the engine change.

   f845c2ca02
3. Merge pull request #357 from ekked/fix-ini-menu-group-links-interpreted (opens in a new window) · 11/05/2026 12:14 pm
   98d4a81d94
4. Updated: The translation wave pages count the German example descript… (opens in a new window) · 11/05/2026 11:28 am

   Message

   Updated: The translation wave pages count the German example descriptions added today
   
   The German description catalogue of the generated examples grew by 131
   entries, so the help tables now hold 117233 words instead of 112725; the
   English and German translation wave pages are regenerated with
   ./console exp:i18n waves, as their test asks.

   7666a06bd7
5. Renamed: The help pages' script lives in design/standard, so the doc … (opens in a new window) · 11/05/2026 11:20 am

   Message

   Renamed: The help pages' script lives in design/standard, so the doc site's filter and search work again
   
   The contents page's filter box, the dashboard's instant search, the lazy
   folders of the viewer's tree, the copy and print buttons and the reading
   position all come from exphelp.js, which the help templates load with
   ezscript_require. The script existed only in design/admin and
   design/admin4, so on a front-end siteaccess such as the doc site
   (designs doc/doc2025, then standard) the packer found nothing and left
   it out without a word: the filter stayed hidden and typing in a search
   field did nothing. Signed in to the admin, where the file was found,
   everything worked.
   
   The two identical copies become one, design/standard/javascript/exphelp.js,
   which every design finds through the standard fallback. The tests read
   that copy and now fail if a design or an extension gets its own again;
   the test map and the file manifest follow the move.

   2a7ee17b6e
6. Updated: German design pattern pages link to the German area pages (opens in a new window) · 11/05/2026 11:06 am

   Message

   Updated: German design pattern pages link to the German area pages
   
   The German pattern catalogue linked 104 times to area-*.md, chain.md and
   fetch-function.md, the English file names, which do not exist below
   doc/ger-de/standards/muster/ (the German pages are bereich-*.md, kette.md
   and fetch-funktion.md). Each link now names the German file the path map
   gives for that page, so the German link check no longer reports them.

   de911a1fd3
7. Updated: The German generated examples translate every source descrip… (opens in a new window) · 11/05/2026 11:06 am

   Message

   Updated: The German generated examples translate every source description again
   
   The German example pages quoted 131 descriptions from the sources in English:
   newer INI comments (the SQLite busy wait and PRAGMA settings, the search
   suggestions, Smarty security, preview links, the review hold, migrations,
   static publishing, Verify it's you groups), new command options and the
   code_file migration step. They now have German translations in the
   description catalogue; 23 older translations whose English source had
   changed are updated in place, and the pages are regenerated with
   exp:help:examples --language=ger-de (no text left in English).
   
   The German directory name map is written sorted, one entry per line, as
   exp:help maps expects (content unchanged).

   d0a4987476
8. Updated: The generated help examples show the help's reading_key and … (opens in a new window) · 11/05/2026 10:34 am

   Message

   Updated: The generated help examples show the help's reading_key and report_links fetch functions and its ReportSecurityAnchor setting
   
   The reference examples below doc/eng-us/examples/ and their German pages below
   doc/ger-de/beispiele/ are generated again from the current sources with
   exp:help:examples, so they match what has landed:
   
   - the fetch functions exphelp/reading_key and exphelp/report_links get their
     pages, and the exphelp index counts seven functions;
   - exphelp.ini [ContextHelp] lists ReportSecurityAnchor, and the path rules of
     [AudienceSettings] and [CategorySettings] and the [TreeSettings] and
     [TypeSettings] entries follow the shipped file;
   - the German site.ini [DatabaseSettings] page lists the SQLite busy-wait and
     transaction-wait settings;
   - count_chars, extract_right, html_entity_decode and simplify quote their first
     use in a shipped template, now the help menu and the text part of form mail.
   
   doc/ger-de/.paths.json maps the two new German pages. exp:help:examples --check
   reports no difference for either language.

   ff3d9701b2
9. Merge remote-tracking branch 'origin/main' into wip-xmltext-literal-l… (opens in a new window) · 11/05/2026 10:25 am

   Message

   Merge remote-tracking branch 'origin/main' into wip-xmltext-literal-links

   780411c42e
10. Updated: A subtree copy re-points every link and embed of its XML tex… (opens in a new window) · 11/05/2026 10:23 am

    Message

    Updated: A subtree copy re-points every link and embed of its XML text at the copies, also those before a literal block
    
    The string scan of step 5 jumped past the next literal block before it looked for links, so every link and embed
    that came before a literal (code) block in the same text kept pointing at the original; it also took the next
    link before an earlier embed or object. A copy of 2,714 nodes left 730 node ids in 189 XML texts pointing at the
    originals.
    
    The XML text is now parsed as XML by expContentJobCopySubtree::remapXMLText(): node_id and object_id of link,
    embed, embed-inline and a, id of object, and href values eznode:// and ezobject:// are re-pointed wherever
    they are. Nothing inside a literal block is changed, ids outside the copied subtree stay, and a text with nothing
    to change is kept byte for byte. The content job "copy", content/copysubtree,
    eZContentObjectTreeNodeOperations::copySubtree() and bin/php/ezsubtreecopy.php all use it.
    
    Tests with an XML fixture cover links before, between and after literal blocks, several literal blocks, embeds,
    eznode:// and ezobject:// hrefs and ids outside the copy (no database).

    3c24718c25