Review main Merged

84312[DOCS] Improve bin/typo3 cache:flush description

Lina Wolf CR +2 V +2 Easy +1 −1 updated 05/25/24, 12:27

Commit message

There is no word 'postparation' according to native speaker Sarah McCarthy.

And also the sentence was hard to understand for non-native speakers.

The command is not only useful during deployment so we mention other usecases.

Review

  • Code review CR +2 Done
    • Benjamin Franzke +2
    • Christian Kuhn +2
    • Chris Müller +1
    • Oliver Klee +1
  • Verified V +2 Done
    • Benjamin Franzke +2
    • Christian Kuhn +2
    • Chris Müller +1
    • Oliver Klee +1
    • core-ci +1 CI

Issues 1

Files 1

Review log 2933

Show 68 earlier messages
  1. Lina Wolf uploaded patch set 1 05/18/24, 08:07

  2. Lina Wolf updated the commit message in patch set 2 05/18/24, 08:08

  3. core-ci voted on patch set 1 V +1 05/18/24, 08:13 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/57887

  4. core-ci voted on patch set 2 V +1 05/18/24, 08:14 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/57888

  5. Chris Müller voted on patch set 2 CR +1 V +1 05/18/24, 08:18

  6. Sarah McCarthy commented on patch set 2 V +1 05/18/24, 08:22

    Looks good to me:-)
  7. Oliver Klee voted on patch set 2 CR +1 V +1 05/18/24, 08:28

  8. Garvin Hicking commented on patch set 2 CR -1 05/18/24, 15:39

    IMO we neec to not put the focus so much on deployment. It should just be a variant. Cache clearing can be used: - local development - after composer updates - to perform it without needing to login My proposal: -- This command can reset all TYPO3 caches. For example, this can be helpful for (local) development, after extension/code updates and after deployments on a remote server. -- Not perfectly happy with that, but I'd prefer this direction, what do you think?
  1. Garvin Hicking commented on patch set 2 05/18/24, 15:40

    Reply

    Oh, another thing: And when the backend/GUI is broken or unusable to changes!
  2. Lina Wolf commented on patch set 2 05/18/24, 16:07

    Reply

    Should we also mention the Opt-Cache and DI?
  3. Chris Müller commented on patch set 2 CR 0 V 0 05/18/24, 16:35

    Reply

    The opcache is not cleared by the command (it can't because it has to connect to nginx/apache/fpm which is not possible by this command). I would also be cautious with "all TYPO3 caches" - as "all" does not mean "all" in TYPO3 context: https://docs.typo3.org/m/typo3/reference-coreapi/12.4/en-us/ApiOverview/CachingFramework/Architecture/Index.html#caching-architecture-core
  4. Lina Wolf uploaded patch set 3 CR +1 V +1 05/18/24, 17:08

  5. Lina Wolf commented on patch set 3 05/18/24, 17:08

    I gave it another try with a different formulation
  6. core-ci voted on patch set 3 V +1 05/18/24, 17:13 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/57893

  7. Oliver Klee voted on patch set 3 CR +1 V +1 05/18/24, 17:17

  8. Garvin Hicking commented on patch set 3 05/18/24, 17:53

    Reply

    Thank you! That sounds to cover much more of this functionality! Chris is right with opcache and the "all caches" situation. Actually I wonder if we should mention something like: Note that the PHP opcache for the webserver process is not something that can be cleared by this command, and requires the Install tool or custom web requests (Again, not too happy about this exact wording and just thinking out loud, maybe you or Chris or someone else has an idea about phrasing this?)
  9. Chris Müller voted on patch set 3 CR +1 V +1 05/19/24, 07:41

  10. Benni Mack voted on patch set 3 CR +1 V +1 05/20/24, 20:38

  11. Benjamin Franzke commented on patch set 3 CR -1 05/21/24, 07:29

    CacheFlushCommand.php:47

    That's a promise that we cannot fully make. It *could* work if it is a web-related exception, but not if it is a system exception. I mean I know what you probably mean, `/typo3/` doesn't work but `bin/typo3` still works – but that also applied to `typo3/install.php`, as both use the same failsafe-boot-method. Maybe we can drop this second senteance, would that be fine with you?

    CacheFlushCommand.php:47

    Trailing whitespace.
  12. Lina Wolf uploaded patch set 4 CR +1 V +1 05/23/24, 13:31

  13. Lina Wolf commented on patch set 4 05/23/24, 13:32

    CacheFlushCommand.php:47 Reply

    Done

    CacheFlushCommand.php:47 Reply

    Done
  14. core-ci voted on patch set 4 V +1 05/23/24, 13:49 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/58104

  15. Chris Müller voted on patch set 4 CR +1 V +1 05/23/24, 18:01

  16. Christian Kuhn voted on patch set 4 CR +1 V +1 05/24/24, 07:17

  17. Benjamin Franzke voted on patch set 4 CR +2 V +2 05/24/24, 08:24

  18. Garvin Hicking commented on patch set 4 CR -1 05/24/24, 08:27

    Before this gets merged, could we address whether to mention that the PHP OpCache can NOT be cleared with this command? This comes up quite often.
  19. Benjamin Franzke commented on patch set 4 05/24/24, 08:57

    Reply

    Sorry, already merged the v12 backport. I personally think this is not the right place. Where does this come up often? Can we have a separate issue for that?
  20. Garvin Hicking commented on patch set 4 CR 0 05/24/24, 09:12

    Reply

    Deployment is specifically noted (already before the patch) so somehow people expect "all caches" imply "opcode", at least in my experience in supporting colleagues, slack and stackoverflow. I also feel this is documentation, so I'd just like to have a short addition like "excluding web-based PHP OPcache" somewhere so that people get hinted at this. googling "clear opcode cache" can reveal a lot of info then, independently from TYPO3. I'm good with addressing it in a follow-up, if we would agree that this is worthwile. I think this is a chance now to address this as a "classical problem" real-life users face.
  21. Oliver Klee voted on patch set 4 CR +1 V +1 05/24/24, 09:31

  22. Christian Kuhn commented on patch set 4 05/25/24, 12:12

    Reply

    Hey. I hope it's ok to merge here now, since we're currently in the unfortunate situation that v12 has been merged, while main hasn't. To the topic: opcache is usually transparent when configured well (default), and when admins did not mess around with opcache settings: opcaches should usually auto-invalidate when the according php source files has been modified in web context, and are disabled in cli context. i'm aware there are server-configs that mess around in this area aiming for (imho: broken) opcache improvements, and instances based on such platforms then may become victims of old opcaches. i'd argue this is a server environment thing, not an application topic. imho, the best thing we should do is to 'document-away' such situations and tell platforms with such settings are risky and may lead to additional efforts when php applications are based on this, which they should then communicate accordingly to users. i think there is no need to hint on this in a general 'help' sentence in the cache-flush command.
  23. Christian Kuhn voted on patch set 4 CR +2 V +2 05/25/24, 12:15

  24. Christian Kuhn merged patch set 5 05/25/24, 12:15

  25. Garvin Hicking commented on patch set 5 05/25/24, 12:27

    Reply

    Acknowledged