Page MenuHomePhabricator
Create Task
Maniphest T401252

Configure title-case consistency mapping for PHP 8.1 -> 8.3 transition
Closed, ResolvedPublic
Actions

Description

Process:

  • Use maintenance/language/generateUpperCharTable.php to generate title-case tables for both 8.1 and 8.3. -- These are identical (T401252#11066941).
  • [SKIPPED] Use maintenance/language/generateUcfirstOverrides.php to generate mappings that override 8.3 title-casing (--override) with 8.1 title-casing (--with).
  • [SKIPPED] Merge the resulting overrides into wmf-config/UcfirstOverrides.php while leaving the permanent override for Eszett intact.
NOTE: The mbstring extension in both 8.1 and 8.3 supports Unicode 14, resulting in identical title-casing behavior. Thus, for the purposes of the 8.1 -> 8.3 migration, no consistency overrides should be necessary.

See T372603: Regenerate UcfirstOverrides.php for PHP 7.4 -> 8.1 transition for prior art during the last PHP migration.

Related Objects
Search...

Event Timeline

Scott_French created this task.Aug 5 2025, 7:50 PM
Restricted Application added a subscriber: Aklapper. · View Herald TranscriptAug 5 2025, 7:50 PM
Scott_French added a subtask: T399884: Configure production MediaWiki image builds for PHP 8.3.Aug 5 2025, 7:50 PM
Scott_French mentioned this in T401254: Upgrade mw-debug/next to PHP 8.3.Aug 5 2025, 8:04 PM
Scott_French added a parent task: T401254: Upgrade mw-debug/next to PHP 8.3.
Krinkle added a parent task: T360995: Migrate Wikimedia production from PHP 8.1 to PHP 8.3.Aug 5 2025, 8:35 PM
Krinkle mentioned this in T360995: Migrate Wikimedia production from PHP 8.1 to PHP 8.3.Aug 6 2025, 1:29 PM
Scott_French changed the task status from Open to In Progress.Aug 6 2025, 7:32 PM
Scott_French closed subtask T399884: Configure production MediaWiki image builds for PHP 8.3 as Resolved.

PHP 8.3 images now exist, with the first-available cli image being docker-registry.discovery.wmnet/restricted/mediawiki-multiversion-cli:2025-08-06-191212-publish-83.

Scott_French added a subscriber: Krinkle.Aug 6 2025, 11:09 PM

With 8.3 images now built (but otherwise unused), I was able to pick this up today.

Since durable storage for output files is not supported under mwscript-k8s, I've used the workaround referenced here in Wikitech in what follows - i.e., opening a shell.php job in --attach mode and executing scripts within that container, then using kubectl cp to retrieve output files.

Not pictured: Creation of two shell.php jobs, one targeting the then-latest PHP 8.1 image (i.e., the default behavior) and another passing --mediawiki_image restricted/mediawiki-multiversion-cli:2025-08-06-191212-publish-83 to target the new 8.3 image. These correspond to the mw-script.eqiad.hc0dqjj3-gljr7 and mw-script.eqiad.y7cwin33-2fgk pods, respectively.

swfrench@deploy1003:~/uctables_8.1_8.3$ kube_env mw-script-deploy eqiad  # Note: -deploy credentials are needed for exec and cp
swfrench@deploy1003:~/uctables_8.1_8.3$ kubectl exec pod/mw-script.eqiad.hc0dqjj3-gljr7 -c mediawiki-hc0dqjj3-app -it -- /bin/bash
www-data@mw-script:/data$ php -v
PHP 8.1.33 (cli) (built: Jul 24 2025 21:07:29) (NTS)
Copyright (c) The PHP Group
Zend Engine v4.1.33, Copyright (c) Zend Technologies
    with Zend OPcache v8.1.33, Copyright (c), by Zend Technologies
www-data@mw-script:/data$ mwscript maintenance/language/generateUpperCharTable.php testwiki --titlecase --outfile /tmp/uctable_8.1.json
www-data@mw-script:/data$ 
exit
swfrench@deploy1003:~/uctables_8.1_8.3$ kubectl cp mw-script.eqiad.hc0dqjj3-gljr7:/tmp/uctable_8.1.json uctable_8.1.json
swfrench@deploy1003:~/uctables_8.1_8.3$ kubectl exec pod/mw-script.eqiad.y7cwin33-2fgkl -c mediawiki-y7cwin33-app -it -- /bin/bash
www-data@mw-script:/data$ php -v
PHP 8.3.24 (cli) (built: Aug  4 2025 18:25:53) (NTS)
Copyright (c) The PHP Group
Zend Engine v4.3.24, Copyright (c) Zend Technologies
    with Zend OPcache v8.3.24, Copyright (c), by Zend Technologies
www-data@mw-script:/data$ mwscript maintenance/language/generateUpperCharTable.php testwiki --titlecase --outfile /tmp/uctable_8.3.json
www-data@mw-script:/data$ 
exit
swfrench@deploy1003:~/uctables_8.1_8.3$ kubectl cp mw-script.eqiad.y7cwin33-2fgkl:/tmp/uctable_8.3.json uctable_8.3.json
swfrench@deploy1003:~/uctables_8.1_8.3$ ls -l
total 63672
-rw-rw-r-- 1 swfrench wikidev 32599320 Aug  6 20:31 uctable_8.1.json
-rw-rw-r-- 1 swfrench wikidev 32599320 Aug  6 20:38 uctable_8.3.json
swfrench@deploy1003:~/uctables_8.1_8.3$ md5sum *
ce210aaf11c6d32695f6478a19e2fd2f  uctable_8.1.json
ce210aaf11c6d32695f6478a19e2fd2f  uctable_8.3.json
swfrench@deploy1003:~/uctables_8.1_8.3$

Indeed, this is consistent with inspection of the mbstring Unicode tables in the PHP source code as noted in the task description: both PHP versions support Unicode 14 and thus produce identical title-casing behavior.

Which is to say, for the purposes of the 8.1 -> 8.3 upgrade, it seems we will not need to introduce title-case consistency overrides.

Scott_French updated the task description. (Show Details)Aug 6 2025, 11:14 PM
Scott_French closed this task as Resolved.Aug 7 2025, 4:45 PM
Content licensed under Creative Commons Attribution-ShareAlike (CC BY-SA) 4.0 unless otherwise noted; code licensed under GNU General Public License (GPL) 2.0 or later and other open source licenses. By using this site, you agree to the Terms of Use, Privacy Policy, and Code of Conduct. · Wikimedia Foundation · Privacy Policy · Code of Conduct · Terms of Use · Disclaimer · CC-BY-SA · GPL · Credits