87996[!!!][TASK] Move EXT:linkvalidator task TSconfig from string to file
Commit message
Review
-
Code review CR 0
-
Verified V 0
Issues 2
Files 7
- typo3/sysext/core/Documentation/Changelog/14.0/Breaking-106062-LinkValidatorSchedulerTaskNoLongerAllowsTypoScriptConfigurationAsString.rst2new+67 −0
- typo3/sysext/linkvalidator/Classes/Exception/FileConfigurationException.phpnew+25 −0
- typo3/sysext/linkvalidator/Classes/Exception/LegacyConfigurationException.phpnew+25 −0
- typo3/sysext/linkvalidator/Classes/Task/ValidatorTask.php+52 −5
- typo3/sysext/linkvalidator/Classes/Task/ValidatorTaskAdditionalFieldProvider.php+40 −11
- typo3/sysext/linkvalidator/Classes/Exception.php8new+25 −0
- typo3/sysext/linkvalidator/Resources/Private/Language/locallang.xlf+5 −2
Review log 2831
Show 46 earlier messages
-
Garvin Hicking uploaded patch set 1 02/03/25, 11:40
-
core-ci voted on patch set 1 V +1 02/03/25, 11:46 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/73386
-
Georg Ringer voted on patch set 1 CR +1 V +1 02/03/25, 14:41
-
Garvin Hicking rebased onto patch set 2
CR +1V +102/19/25, 10:47 -
Georg Ringer voted on patch set 2 CR +1 V +1 02/19/25, 10:47
-
core-ci voted on patch set 2 V +1 02/19/25, 10:57 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/74371
-
Simon Schaufelberger commented on patch set 2 02/28/25, 00:58
Exception.php:25can we define this one as abstract as it is not directly in use? -
Garvin Hicking commented on patch set 2 02/28/25, 07:18
Exception.php:25ReplyDo you have examples of this in the core? I copied it from other places. if there's "precedence" to have an abstract for it, I'd be ok with this, but I don't want to introduce different concepts from the rest. -
Garvin Hicking commented on patch set 3
CR +1V +102/28/25, 07:19Published edit on patch set 2. -
Garvin Hicking commented on patch set 3 02/28/25, 07:19
Breaking-106062-LinkValidatorSchedulerTaskNoLongerAllowsTypoScriptConfigurationAsString.rst:20ReplyDone -
core-ci voted on patch set 3 V +1 02/28/25, 07:25 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/74866
-
Georg Ringer voted on patch set 3 CR +1 V +1 02/28/25, 07:46
-
Simon Schaufelberger commented on patch set 3 02/28/25, 12:53
Exception.php:25ReplyI'm not aware, I just find it dumb to use a generic exception directly. -
Garvin Hicking commented on patch set 3 02/28/25, 13:38
Exception.php:25ReplyI would suggest then to address this in a follow-up across the places where we currently also do not do this? -
Simon Schaufelberger commented on patch set 3 02/28/25, 14:55
Exception.php:25Replywouldn't this be potentially breaking if some ext developer would use this generic exception? -
Garvin Hicking commented on patch set 3 02/28/25, 16:15
Exception.php:25ReplyMaybe I'm not understanding the goal here: If we define the exception as an abstract, we lose the ability to throw that (to my understanding). Thus, if we do not have it as an abstract, extensions could throw it. Since it is not caught specifically, it would show up as an exception in logs/output. That would be the/my expectation though, extension developers can throw any kind of exception already. It would not be covered by this current patch's caught exceptions. I think having a generic exception is more forward-compatible without requiring to change the exception or invent a generic one from the abstraction, if needed. With an abstract such a generic will not be possible again. Are you ok with keeping it like this with thag reasoning? The core has a total of 6 abstract exceptions and 290 non-abstract ones, 229 of them extending a generic or PHP Core exception... -
Simon Schaufelberger commented on patch set 3 CR +1 V +1 02/28/25, 21:20
Exception.php:25ReplyI'm ok with it. It was just an idea if it would make sense, not a request that it has to be done ;) -
Garvin Hicking commented on patch set 3 02/28/25, 21:32
Exception.php:25ReplyAll good, I appreciate your thoughts on it, and thinking this through helped me think through it. Let's see if someone else wants to chime in on the matter. thanks! -
Benni Mack voted on patch set 3 CR -1 03/21/25, 08:41
-
Garvin Hicking commented on patch set 3 03/22/25, 10:09
Blockenld due to https://review.typo3.org/c/Packages/TYPO3.CMS/+/88675 (I guess?) -
Benni Mack commented on patch set 3 03/27/25, 13:30
I think it is safe to say, that we get https://review.typo3.org/c/Packages/TYPO3.CMS/+/88675 in, the current solution might become much more useful? Because, in the final end, we might have a TCA field with codemirror to edit TSconfig in there... Much better than a file that changes config for a task "from the outside" WDYT? -
Garvin Hicking commented on patch set 3 03/27/25, 14:52
Reply
The problem this patch wants to solve is not about persistence. It is about to unblock migrating tasks to symfony commands so https://forge.typo3.org/issues/98453 can go on. Symfony commands cannot take multiline arguments as input. With the proposed patch of yours (which I fully embrace!) we would still not be able to migrate the linkvalidator task then. The only other way I see is: - keep one core task as a task and not a symfony command (then this linkvalidator task can't be ran ad-hoc wothout a scheduler task) - somehow contribute to Symfony so that multiline arguments are somehow setable (I don't think I could manage this) -
Benni Mack commented on patch set 3 03/27/25, 14:53
Reply
I think we need to fix Symfony commands then to support multline arguments! -
Garvin Hicking commented on patch set 3 03/27/25, 15:27
Reply
OK... specifyinf multines is a bit hard I guess, due tocsifferent shell handlings. Do you have contacts to the symfony team to ask about the possibilities to get this off the ground? -
Garvin Hicking commented on patch set 3 03/27/25, 15:34
Reply
(If symfony would only do this in a v8 or v7 that would mean we have to raise that as a minimum version then, something maybe worth considering - so would be cool if they see it as a bug for patchlevel -
Benni Mack commented on patch set 3 03/27/25, 15:44
Quoted multiline arguments work on my machine with symfony 7 and TYPO3 v13. -
Benni Mack commented on patch set 3 03/27/25, 15:44
./vendor/bin/typo3 benni:example "What is this With more than one line And much more" -
Garvin Hicking commented on patch set 3 03/27/25, 16:16
Reply
I need to revisit the patch. part of the problem was that symfony commands were lacking a "type" to receive a textarea input I think. So currently you cannot pass multilines from a scheduler symfony command task to the symfony command cliReply Unresolved
I might have an idea -
Benni Mack commented on patch set 3 03/27/25, 16:30
Reply
Here is my main point, why I think we shouldn't do this: a) we're creating a breaking change with a downside for site admins that do not want to deal with files for TSconfig (probably within fileadmin/ as they do not have access to Git usually) - if we go with Git, we combine creating a task (in the GUI) b) we want to migrate to Tasks to Commands (migration path -> also unclear still) to enhance DX c) we have no really good solution yet to have the flexibility when editing task fields (which will become CLI commands) via GUI yet as we know with AdditionalFieldProviders. all of that gives me the shivers to move forward with this patch (right now). I'd rather a) change the "let's serialize the objects" b) add TCA for scheduler tasks -> then change the rendering, add TCA types for it c) migrate additional field providers as additional fields of a task in the DB (or type=json) with type=user rendering. d) see what should be migrated to a CLI command (where it makes sense) and what should just stay a scheduler task (in a simple way) -> this is the point where we should talk about multiline arguments... :) -
Garvin Hicking commented on patch set 3 CR -2 03/27/25, 16:36
Reply Unresolved
agreed. sounds like a plan! I will copy that to the task migration patch and then abandon this here. -
Garvin Hicking abandoned the change 03/27/25, 18:48
Abandon to proceed in https://review.typo3.org/c/Packages/TYPO3.CMS/+/74053
Breaking-106062-LinkValidatorSchedulerTaskNoLongerAllowsTypoScriptConfigurationAsString.rst:20