Skip to the content
Documentation
Help Start
Reading in:
  • German: 9,090 of 9,090 pages
About the help in other languages
All help pages

Links, URL aliases, wildcards and search statistics

This guide teaches the four pages of the administration interface that look after addresses: the link list (which external and internal links your content uses and whether they work), the global URL aliases (addresses…

Download as PDF

Address: https://doc.exponential.earth/exphelp/viewer/guides/urls-and-aliases.md

On this page

Links, URL aliases, wildcards and search statistics

This guide teaches the four pages of the administration interface that look after addresses: the link list (which external and internal links your content uses and whether they work), the global URL aliases (addresses of your choosing for module views), the URL wildcards (one rule for a whole group of addresses) and the search statistics (what visitors search for and whether they find it).

It is written for administrators and editors who keep a site's links and addresses in order. Every page, field and rule below was checked against the code of this repository on 6 October 2026; the files are listed under References.

Guides · Related: Paging in the admin · Cronjobs · Operating a site · Migrating URLs

In short

  • Setup > Link management (/url/list) lists every address that published content links to, once, with its state (valid, invalid, never checked), when it was checked and changed, and which objects use it. Search by any part of the address, show one state, order by address, last check or last change, and mark ticked links valid or invalid by hand.
  • The link check (linkcheck.php, part infrequent) requests http and https links alike, judges links to content by their node or object, never requests addresses on a private network and waits between requests to one host. Its settings are in cronjob.ini [linkCheckSettings].
  • Setup > URL translator (/content/urltranslator) makes and removes global URL aliases: login for user/login, for example. Each alias shows its whole path, its destination and what that resolves to now; an alias to a module that no longer exists is marked.
  • Setup > URL wildcards (/content/urlwildcards) makes and removes wildcards: news/* to articles/{1}. The page refuses a destination that uses a {n} the pattern has no * for, and Try an address shows which wildcard an address matches and where it leads, without changing anything.
  • Setup > Search statistics (/search/stats) shows every recorded search phrase, how often it was searched for and the average number of results. It says when no siteaccess records searches (the default), filters the phrases that found nothing, orders by count, phrase or fewest results, and runs a phrase again with one click.
  • Every list is paged, every search and filter is kept in the address (so it survives paging and can be bookmarked), and every removal of more than the ticked rows asks in place first.

Contents

Open Setup > Link management. The page starts with five figures:

FigureWhat it countsClick to
Links in published contentaddresses that at least one published version links toshow all
Validof those, marked validshow the valid ones
Invalidmarked invalid (red when there are any)show the invalid ones
Never checkednever tried by the link checkshow them
Last link checkthe time of the most recent check of any link–

Below them, Find a link searches any part of the address (example.org, /about, mailto:), without regard to upper and lower case; % and _ are matched as themselves. Show picks the list (All, Valid, Invalid, Never checked) and Order sorts it: Address A to Z (the default), Last checked (most recent first) or Last modified.

Each link is a card with:

  • the address, which opens the link's own page (/url/view/<id>) with every object version that uses it;
  • badges: Valid or Invalid, Never checked, and how the link check treats the address (see section 2);
  • View, Edit (change the address in one place for every object that uses it) and Open, which opens the address in a new window. Open is offered only for web, mail and site addresses, never for javascript: or data:, so a bad link in content cannot run in the admin;
  • Checked and Modified times;
  • Used by: up to three published objects that link to it, each linked to its main node, and "and N more", which leads to the link's page.

To mark links by hand, tick them (or Select all on this page) and press Mark selected valid or Mark selected invalid. The page says how many were marked. Marking changes only the state; nothing is removed and no content changes. The next link check tests the links again and may change the state back.

The page size is chosen under the list (10, 25 or 50, remembered per user). The sizes on offer are the setting ItemsPerPageList_url_list[] in admininterface.ini [PaginationSettings], like every admin list (paging).

The link check is the cronjob script linkcheck.php in the infrequent part (cronjob.ini). It runs from the shell only, never in a web request (also not under Velocity). Run the whole part, or the link check alone:

php runcronjobs.php -s <your public siteaccess> infrequent
php runcronjobs.php -s <your public siteaccess> --script=linkcheck.php

Run it as the user the web server runs as, like the other cronjobs. It checks the links checked longest ago first and prints one line per link with the result and why, then a summary such as:

All links have been checked! 189 valid, 72 invalid, 89 not decided, 0 skipped (checked recently or over MaxURLsPerRun), 48 changed, 725 requests, 566.7 s

A link changes state only when the check decides; not decided keeps the state it had. The time of the check is recorded for every link it looked at. The badge on each card says how its kind is treated:

BadgeAddressWhat the check does
(none)http:, https:, ftp:requests it: HEAD, then GET when HEAD fails, following redirects up to MaxRedirects, with TLS certificates verified. A 2xx answer is valid; 404, 410, other errors, a bad certificate, an unknown host or too many redirects are invalid; 429 and 503 (try later) and 401, 403 (the site refuses automated requests) are not decided
E-mailmailto:valid when the domain has a mail server (MX record)
On this sitea path such as /about or /admin/content/dashboardwithout the query, the fragment and a leading siteaccess name: valid when it is the front page, a module view or a URL alias, else tried on the addresses of SiteURL[]
Link to contentezlocation://, eznode://, ezobject:// (links in rich text)valid while the node or object exists, is published and is not hidden; invalid otherwise. No request is sent
Not testedfile:names a file on a computer, not a page: never tested, the state is kept
Other addressany other scheme (tel:, javascript: ...)looked up as a path of this site, as it always was, so usually invalid

Private addresses are never requested. Before every request, also after a redirect, the host is resolved; when an address is on a private, loopback, link-local or otherwise reserved network (10.0.0.0/8, 127.0.0.0/8, 169.254.0.0/16, 192.168.0.0/16, ::1, fe80::/10 ...), the link is not decided and nothing is sent. Otherwise a link in content could make the server reach inside its own network. The request then goes to the address that was checked, so the name cannot point elsewhere a moment later. An intranet that should be tested is listed in AllowedPrivateHosts[]; the site's own SiteURL[] addresses are always allowed.

One host is not hammered. Requests to the same host wait HostDelay milliseconds, and a link checked less than RecheckInterval seconds ago is skipped, so running the part twice in a row costs nothing.

The settings, in cronjob.ini [linkCheckSettings] (override them in settings/override/cronjob.ini.append.php):

KeyDefaultMeaning
Timeout15seconds one request may take
ConnectTimeout5seconds to connect
MaxRedirects5redirects followed before the link counts as broken
HostDelay1000milliseconds between two requests to the same host
RecheckInterval72000seconds before a checked link is checked again; 0 checks every link on every run
MaxURLsPerRun0links checked in one run (the ones checked longest ago first); 0 checks all
MaxRunTime900seconds a run may spend checking links; the rest waits for the next run; 0 no limit
UserAgentExponential Link Validator (+https://exponential.earth)the User-Agent of the requests
AllowedPrivateHosts[]emptyhosts, addresses or IPv4 ranges (10.1.0.0/16) on a private network that may be tested
SiteURL[]emptythe site's own addresses, tried for a path that is no alias

On the 350 links of a demonstration site the first run took about nine and a half minutes and sent 725 requests: most of the time is the per-host delay between links to the same few hosts.

3. Global URL aliases

A global URL alias gives a module view, or another alias, an address of your choosing. Open Setup > URL translator. The aliases of content (the addresses of nodes) are not here: they are on the URL aliases tab of each node.

Create one

FieldWhat to enter
New URL aliasthe address visitors will use, without the host and without a leading slash: login, campaign/autumn. Characters that are not allowed are changed, and the page says to what
Destinationa module view (user/login, content/search) or an existing address of content (about-us). content/view/full/<node ID> makes an alias of that node instead, which is then listed on the node's URL aliases tab, not here
Languagethe alias works in siteaccesses that show this language
Include in other languagesmakes it work in every language
Alias should redirect to its destinationchecked (the default): visitors get a 301 redirect to the destination's address. Unchecked: the destination is shown under the alias

Press Create. When the alias cannot be created, the message says why, the field is marked, and what you typed stays in the form:

MessageMeans
Text is missing for the URL aliasthe alias field is empty
Text is missing for the URL alias destinationthe destination is empty
The specified destination URL does not exist in the systemthe destination is neither a module view nor an existing address
The URL alias already exists, and it points to ...the address is taken; both links lead to what it is now
The specified language code is not validthe language does not exist

Read the list

Each alias shows its whole path (each part links to what it is), its kind (Redirect or Direct), the Destination, what it Resolves to (the view and module, or a node), its language and whether it is always available. Module not found marks an alias whose module no longer exists, usually because an extension was switched off: visitors get an error page there.

Find an alias searches the last segment of each alias (login in campaign/login) and its destination, Show picks All, Redirecting or Direct. Both are conditions of the database query, so a page costs one count and one page query however many aliases there are.

Remove

Tick aliases and press Remove selected; the browser asks once more. Remove all opens a confirmation in place that says how many aliases go; aliases of content nodes are kept. A removed alias stops working at once.

4. URL wildcards

A wildcard sends a group of addresses elsewhere with one rule. Each * in the pattern matches any text, and {1}, {2} ... in the destination put that text back. Wildcards are only consulted when no URL alias matches, and they are tried in the order they were created; the first that matches wins.

PatternDestinationnews/2026/october becomes
news/*articles/{1}articles/2026/october
news/*/*archive/{2}/{1}archive/october/2026
old-blog/*https://blog.example.org/{1} (redirecting)a redirect to the other site

The pattern matches from the start of the address, without regard to case; text after the matched part is kept.

Create one

Enter New URL wildcard (the pattern) and Destination, check Redirecting URL to answer with a 301 redirect (visitors see the destination address), or leave it unchecked to show the destination under the address asked for. The page refuses:

  • an empty pattern or destination;
  • a pattern that exists already (the message names where the old one leads);
  • a destination with a placeholder the pattern has no * for, such as {2} with one *. Without this check the wildcard would send visitors to an address with a part missing.

Try an address

Try an address takes an address such as news/2026/october and says which wildcard it matches, what it becomes, whether visitors are redirected, and what the result resolves to on this site (an alias, a module view, another site, or nothing, which means an error page). It changes nothing; the address is kept in the page's address, so it can be sent to a colleague.

Find and remove

Find a wildcard searches pattern and destination, Show picks All, Redirecting or Direct (both in the database query, in the order the wildcards are tried). Removing works as for aliases: Remove selected asks once, Remove all asks in place. Every change empties the wildcard cache.

5. Search statistics

Open Setup > Search statistics.

Searches are only recorded where switched on. A search counts when the siteaccess the visitor searches in has

# settings/siteaccess/<public siteaccess>/site.ini.append.php
[SearchSettings]
LogSearchStats=enabled

It is disabled by default. The page says which siteaccesses record searches, or warns that none does; what is listed was then recorded earlier.

The figures count the different phrases, all searches, the phrases that found nothing (click to list them) and the share of all searches that found nothing.

Find a phrase searches the phrases; Show picks all phrases or only those that found nothing; Order sorts by Most searched (the default), Phrase A to Z or Fewest results (average results, lowest first). The bar beside each count compares it with the most searched phrase on the page. Search now runs the phrase in the administration search, to see what it finds today.

A phrase searched for often that finds nothing points to content that is missing, or to words the content does not use: add the content, or the words (a synonym in the text or in the keywords).

Reset statistics opens a confirmation in place that says how many phrases go, and removes all of them for good.

Phrases are what visitors typed. They can contain names or other personal data; treat the page accordingly and reset it when the numbers are no longer needed.

6. Addresses and parameters of the pages

PageAddressParameters
Link list/url/list/<all|valid|invalid|unchecked>(sort)/checked, (sort)/modified, (offset)/<n>, (q)/<text>
Global aliases/content/urltranslator(kind)/redirect, (kind)/direct, (offset)/<n>, (q)/<text>
Wildcards/content/urlwildcards(kind)/..., (offset)/<n>, (q)/<text>; the Try form posts test (an address)
Search statistics/search/stats(sort)/phrase, (sort)/fewest, (show)/none, (offset)/<n>, (q)/<text>

Unknown values fall back to the defaults. The forms, their field names and their buttons (SetValid, SetInvalid, URLSelection[], NewAliasButton, RemoveAliasButton, RemoveAllAliasesButton, ElementList[], NewWildcardButton, RemoveWildcardButton, RemoveAllWildcardsButton, WildcardIDList[], ResetSearchStatsButton) are the ones the pages always had, and every POST form carries the form token.

7. When something does not behave as expected

You seeWhy, and what to do
Most links say Never checkedthe link check has not run on them: run the infrequent cronjob part
A link that works in the browser stays as it was after a runthe site answered 401, 403, 429 or 503 to the link check (refused or busy), which decides nothing; the line of the run says which
A link to an intranet page is never testedits host is on a private network; add it to AllowedPrivateHosts[]
A link to content is Invalidits node or object was removed, is hidden or is not published; the line of the run names the node or object
A link to /admin/... or an extension's module is Invalidthe module is only active in that siteaccess: run the part with -s of that siteaccess, or ignore the state
A run takes longmost of it is HostDelay between links to the same host; lower it carefully, or set MaxURLsPerRun; a run stops after MaxRunTime (15 minutes) and the next one goes on
The search statistics stay emptyno siteaccess has LogSearchStats=enabled; the page says so at the top
A new alias "was modified by the system"characters not allowed in an address were changed; the message shows the result
A new alias to a node does not appear in the listaliases of nodes are on the node's URL aliases tab
A wildcard does nothinga URL alias matches the address first, or an older wildcard matches it; use Try an address
The wildcard is refused with "The destination uses {2}"the pattern has fewer * than the destination needs: add a * or remove the placeholder

8. Decoding and cleaning addresses: two optional settings

Two settings of site.ini [URLTranslator] change how addresses are read and how new URL aliases are written. Both are off as shipped, on new installations and after an upgrade alike, so nothing changes until you switch them on. Both come from LoveStack, Mugo Web's distribution of the same kernel, with thanks (commits f435c2f91e and 7517c0e925).

# settings/override/site.ini.append.php
[URLTranslator]
# decode the address of a request once instead of twice
DecodeOnce=enabled
# IRI aliases: treat the double quote as a reserved character
TrailingCleanup=enabled

DecodeOnce. The address of a request is decoded when the request starts, and was decoded a second time when the kernel split it into module, view and parameters. An encoded + in a path therefore arrived as a space: /(tag)/bar%2Bbar gave the view parameter tag the value bar bar, and %2525 became %. With DecodeOnce=enabled the address is decoded once: bar+bar, and %2525 stays %25.

  • What an editor sees: nothing, unless a link was written with an encoded + or % in a view parameter or a path.
  • What an administrator checks before switching it on: links in templates or content that relied on the second decoding (a parameter written as a%2520b to mean a b). URL aliases themselves are not affected: their text never holds a + or % in the urlalias and urlalias_compat transformations, and on the demonstration installation every one of its 696 node addresses resolves the same with the setting off and on.
  • Set it in settings/override/site.ini.append.php, not in a siteaccess: the address is read before the siteaccess is known. It works the same under PHP-FPM and the Velocity engine, which hand the address to the kernel as the browser sent it; the response cache keys are made from that address too, so they do not change.

TrailingCleanup. With the IRI transformation (TransformationGroup=urlalias_iri, aliases that keep accents and other letters) the double quote was kept in an alias: an article named Lecture "Storm surges" got the alias Lecture-"Storm-surges". With TrailingCleanup=enabled the double quote is a reserved character like ? or +: it becomes a word separator, and separators, dots and ! are stripped at the beginning and the end, so the alias is Lecture-Storm-surges. The urlalias and urlalias_compat transformations already drop the double quote and already strip !, dots and separators at the end, so for them the setting changes nothing.

  • What changes on an upgrade: nothing. Existing aliases stay as they are; an object gets the new alias when it is published again (or when bin/php/updateniceurls.php rebuilds the aliases), and the old alias then stays as a redirect to the new one, as with any renamed object.
  • On a copy of the demonstration installation's database switched to urlalias_iri, republishing every node with the setting on changed 6 of 2942 aliases, all of them names with double quotes.

9. View parameters: how a value travels in an address

The options of the administration views are view parameters, /(name)/value (a search, a filter, a sort order). Some characters cannot travel in such a value as they are: Apache refuses an encoded slash (%2F), the kernel decodes the whole address (twice unless DecodeOnce is enabled) before it splits it at every /, and a segment (name) starts the next view parameter. Every view writes its values in one reversible encoding, so a search for example.com/a keeps its slash:

CharacterWritten as
/~2F
( and )~28 and ~29
~ (the escape itself)~7E
% and +~25 and ~2B (the second decoding would change them)
\, ?, #~5C, ~3F, ~23
control characters (a tab, a line break)~09, ~0A ...
a value that is only . or ..~2E, ~2E~2E (a browser would resolve it)

Everything else stays readable: letters, digits, , : @ = ! * are written as they are, and spaces and other text are percent-encoded as usual, which a browser shows as text. Examples:

/url/list/all/(q)/example.com~2Fa
/content/urltranslator/(q)/news~2F2026
/contentfinder/search/(text)/var~2Fsite~2Fstorage~2Fimages~2Fa~2Fb%20~281~29.jpg
/settings/view/admin/site.ini/(q)/RootNode/(scope)/all
/setup/logs/view/error/(level)/error,warning/(search)/not%20found/(from)/2026-10-17%2010:00
/visual/templatelist/(filter)/content~2Fview
/package/list/7x/(search)/demo%20content

Links written before keep working: the content finder and the settings page still read their ~ + base64url form, and the log viewer's and the template list's ~XX form (a space was ~20) is the same encoding. The package list, comparison and browser still read their search encoded twice (demo%2520content).

  • In a template, write a value with the operator view_parameter: {concat( 'settings/view/', $siteaccess, '/', $file, '/(q)/', $name|view_parameter )|ezurl}. Never |urlencode: it leaves / to Apache and + to the second decoding.
  • In a script of the administration designs, window.a4ViewParameter.encode( value ) (a4_url.js, loaded in the page head of admin, admin3 and admin4) writes the same; window.a4ViewParameter.decode() is its inverse.
  • In PHP, expViewParameters::suffix( array( 'q' => $text ) ) writes /(q)/<value>, expViewParameters::encode( $value ) one value, and expViewParameters::get() and text() read a value back decoded (raw() gives it as the kernel handed it over). Tested in tests/tests/kernel/classes/expViewParametersTest.php (round trips through the kernel's decoding, once and twice) and tests/tests/design/AdminViewParametersOnlyTest.php (VP-05, VP-06: the views, templates and scripts use it, and the script writes exactly what PHP writes).

10. Titles in scripts without a transliteration table

The shipped transformation urlalias writes aliases in ASCII. Latin, Cyrillic, Greek and Hebrew letters are transliterated (Руководство пользователя becomes rukovodstvo-pol-zovatelya, מדריך למשתמש becomes mdryk-lmstms), but a script without a transliteration table lost every letter: an Arabic, Persian, Pashto, Chinese, Japanese, Korean, Thai, Devanagari, Bengali, Georgian or Armenian title gave an empty alias, and the node was then called node_<node id> (a custom alias noname1, any other element _1).

Such a title is now converted with the IRI transformation instead, the one TransformationGroup=urlalias_iri uses for every alias: دليل المستخدم becomes دليل-المستخدم, 用户手册 stays 用户手册. The setting is site.ini [URLTranslator] ScriptFallbackTransformationGroup, shipped as urlalias_iri; leave it empty to turn it off.

# settings/override/site.ini.append.php
[URLTranslator]
# keep the old node_<id> aliases for titles in such scripts
ScriptFallbackTransformationGroup=

The rule is deliberately narrow: the fallback applies only when the configured transformation leaves no letter of a title that has letters.

  • A title that keeps a letter gets exactly the alias it got before, so no Latin, Cyrillic, Greek or Hebrew alias changes, and a mixed title keeps what urlalias leaves: Exponential دليل is still exponential. Falling back as soon as some letters are lost would have changed the alias of every such mixed title on its next publication.
  • Digits are not letters: دليل 2024 gave 2024 and now gives دليل-2024.
  • A title without any letter (punctuation, emoji) still gets node_<node id>, noname1 or _1.
  • The fallback is not used when it names the configured transformation itself (urlalias_iri). With TransformationGroup=urlalias_compat it applies in the same way. The path_identification_string of a node (urlalias_compat, always) and identifiers (identifier) are not affected.
  • It applies wherever the kernel makes an alias element from a name: node aliases, custom and global URL aliases entered in the administration, bin/php/updateniceurls.php, and the file names of uploaded images, which are made from the object name the same way. Wildcard patterns are stored as they are typed and are not converted.
  • Existing aliases stay as they are. A node gets the new alias when it is published again or the aliases are rebuilt, and the old node_<node id> alias then stays as a redirect to the new one, as with any renamed object.
  • Before the fallback, control and format characters (NUL, the direction override U+202E, the zero-width space U+200B, U+FEFF) are removed and < > " { } | \ ^ ` become a word separator, because the IRI transformation would let them through; the zero-width non-joiner U+200C and joiner U+200D stay, Persian needs them.
  • Image file names made from such a title now carry its script (دليل-المستخدم.jpg instead of _1.jpg). A site whose storage or tools expect ASCII file names sets ScriptFallbackTransformationGroup= empty.

Tested without a database in tests/tests/kernel/classes/urlalias/eZURLAliasMLScriptFallbackTest.php.

For developers

  • A link within the site in one call: eZURI::build( '/content/view/full/2' ) gives the path the ezurl template operator gives (index file and, for a siteaccess in the URL, its name in front), eZURI::build( $url, true ) the absolute URL with scheme and host. Like ezurl the result is escaped for HTML; pass a third argument false for a URL that goes into a header or a JSON answer. A link with a scheme (https:, mailto:) comes back as it is. Tested in tests/tests/lib/ezutils/eZURIBuildTest.php. With thanks to Mugo Web (LoveStack, commit f2284944b2).
  • new eZURI( $uri, $decode = null ) and setURIString( $uri, $fullInitialize = true, $decode = null ): true for a string that is URL encoded, false for decoded text, null (the default) to follow DecodeOnce (decoded unless it is enabled). eZURI::decodeOnce() says whether it is enabled; eZURI::decodeIRI() and decodeURL() decode as before. eZCharTransform::trailingCleanup() says whether TrailingCleanup is enabled. Both settings are read on every call, so a Velocity worker sees a change after --clear-tag=ini. Tested in tests/tests/lib/ezutils/eZSysRequestURIDecodeOnceTest.php and tests/tests/lib/ezi18n/eZCharTransformTrailingCleanupTest.php.
  • The views are \Exponential\View\Kernel\Url\ListView, \Exponential\View\Kernel\Search\Stats, \Exponential\View\Kernel\Content\UrlaliasGlobal and \Exponential\View\Kernel\Content\UrlaliasWildcard (kernel/private/classes/views/...). Their static helpers (modes, orders, the search text, usage, the link kind, alias destinations, wildcard matching and placeholders) are tested without a database in tests/tests/kernel/classes/expUrlAdminPagesTest.php.
  • eZURL::fetchList() / fetchListCount() take last_checked (never, checked), search (with only_published) and sort (address, checked, modified, id), on every database including MongoDB.
  • eZURLAliasQuery has the properties search (the last segment or the action contains the text) and redirects (alias_redirects); eZURLWildcard::fetchFilteredList() and fetchFilteredListCount() search pattern and destination and filter by type. Both are conditions of the SQL query, or a $match on MongoDB.
  • expLinkCheck (kernel/classes/explinkcheck.php) decides whether a link works: check( $url ) answers valid, invalid or unknown with a reason. The request (curlFetch()), the DNS lookup, the content, alias and MX lookups, the clock and sleeping are callables given to the constructor, which is how tests/tests/kernel/classes/expLinkCheckTest.php tests every decision without a network or a database.
  • Every variable the templates had is still set; new ones: url_summary, url_usage, url_checks, url_search, url_sort, url_feedback, limit, limit_choices (url/list); search_stats_*, search_total_count (search/stats); alias_list, alias_count, alias_total_count, alias_info, alias_search, alias_kind, alias_form (urltranslator); wildcards_total_count, wildcard_search, wildcard_kind, wildcard_test, wildcard_test_result (urlwildcards). The templates fall back to the old variables when a view does not set them.
  • The look is design/admin{,4}/templates/url/exp_style.tpl and search/exp_style.tpl, scoped to the page; no shared stylesheet changed.

References

  • kernel/private/classes/views/url/list.php, kernel/private/classes/views/search/stats.php, kernel/private/classes/views/content/urlalias_global.php, kernel/private/classes/views/content/urlalias_wildcard.php
  • kernel/classes/datatypes/ezurl/ezurl.php (handleList(), listOrderSQL(), searchLikePattern())
  • kernel/private/classes/cronjobs/linkcheck.php (the link check part), kernel/classes/explinkcheck.php (its decisions), settings/cronjob.ini [linkCheckSettings], kernel/classes/ezurlwildcard.php (matching, fetchFilteredList()), kernel/classes/ezurlaliasquery.php (search, redirects)
  • design/admin4/templates/url/list.tpl, search/stats.tpl, content/urlalias_global.tpl, content/urlalias_wildcard.tpl and their copies in design/admin
  • kernel/classes/ezurlaliasml.php (convertToAlias()), settings/site.ini [URLTranslator] ScriptFallbackTransformationGroup
  • settings/site.ini [SearchSettings] LogSearchStats, settings/admininterface.ini [PaginationSettings]
  • Changelog 6.0.15

Become an Exponential Partner

Partners share their information, provide funding, contribute back and develop the products with us, and that is how Exponential grows. A partner website and dashboard are planned for mid-2027: please become a partner today to speed up our growth rate and land new features faster.

Something wrong on this page? Report an error

Documentation license

This documentation is free documentation, licensed under the GNU Free Documentation License 1.3 or later.

Exponential Copyright © 1998-2026 Exponential Foundation and 7x