User tests: Successful: Unsuccessful:
Pull Request for Issue # .
In Joomla 6, we use a single instance of ExtensionAdapter to find updates for all extensions/update sites. The latest update found for previous extension is not reset before finding update for the next extension and it causes the strange bug described in the issue #46066. Basically, if the previous extension has higher version number than the next extension, the update for the next extension could not be found.
ExtensionAdapter to find update for each extension)Note that you need to install the two extensions in exactly order above (Akeeba Backup needs to be installed first because it has higher version number than Admin Tools)
Please select:
Documentation link for docs.joomla.org:
No documentation changes for docs.joomla.org needed
Pull Request link for manual.joomla.org:
No documentation changes for manual.joomla.org needed
| Status | New | ⇒ | Pending |
| Category | ⇒ | Repository Administration com_admin SQL |
My worry is that whatever broke the code here may also have broken code elsewhere and we just havent come across it yet. I wouldnt want us to need to keep using sticking plasters
I have tested this item ✅ successfully on e10a351
Yes it seems to work correctly by resetting the variable.
@joomdonation Could you update the phpstan-baseline.neon file? It seems PHPStan fails for your PR. It fails for parts of code not changed by your PR, no idea why. You can fix that by running ./libraries/vendor/bin/phpstan -b after having run composer install. IT will not require redo of human tests. Can you do that? Or shall I help?
@richard67 It would be great if you could help with this PR. I am not familiar with the process yet.
@joomdonation Funny: Just restarting the PHPStan GitHub action has helped. No idea why. Anyway, no need for any changes on this PR.
Thanks Richard
| Labels |
Added:
bug
PR-6.0-dev
|
||
| Category | Repository Administration com_admin SQL | ⇒ | Libraries |
I've restored the previous human test result in the issue tracker as the commit which has invalidated the test count was just an improvement of a code comment.
I have tested this item ✅ successfully on eeb3106
| Status | Pending | ⇒ | Ready to Commit |
| Build | 5.3-dev | ⇒ | 6.0-dev |
RTC
| Status | Ready to Commit | ⇒ | Fixed in Code Base |
| Closed_Date | 0000-00-00 00:00:00 | ⇒ | 2025-09-26 17:23:49 |
| Closed_By | ⇒ | Bodge-IT | |
| Labels |
Added:
RTC
|
||
Thanks for the fix @joomdonation and tester @joeforjoomla and tests at last minute @richard67
appears to work correctly in my limited test
I wonder why the original code was changed