Review main Merged

77997[FEATURE] Introduce Content Security Policy headers

Oliver Hader CR +2 V +2 Expert +2398 −57 updated 03/16/23, 00:32

Commit message

This change introduces various representations of the
Content-Security-Policy domain as PHP classes.

The PSR-15 middlewares `ContentSecurityPolicyHeaders` are applying
the corresponding HTTP headers to each response in the frontend and
backend scope - in case no other component already added a
`Content-Security-Policy` or `Content-Security-Policy-Report-Only`
header, and only if the corresponding feature flag is enabled:

+ `security.backend.enforceContentSecurityPolicy`
+ `security.frontend.enforceContentSecurityPolicy`

For new installations `security.backend.enforceContentSecurityPolicy`
is enabled via factory default settings.

Review

  • Code review CR +2 Done
    • Oliver Hader +2
    • Markus Klein +1
    • Benni Mack +1
  • Verified V +2 Done
    • Oliver Hader +2
    • Markus Klein +1
    • Benni Mack +1
    • core-ci +1 CI

Issues 3

Files 51

Show 31 more files

Review log 91124

Show 7299 earlier messages
  1. Oliver Hader uploaded patch set 1 03/01/23, 22:05

  2. core-ci voted on patch set 1 V +1 03/01/23, 22:12 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38276

  3. Simon Schaufelberger commented on patch set 1 03/01/23, 22:34

    corresponding... response...

    Important-99499-EnforceBackendContent-Security-Policy.rst:13

    too many empty lines
  4. Oliver Hader commented on patch set 1 03/01/23, 22:44

    Important-99499-EnforceBackendContent-Security-Policy.rst:13 Reply

    Ah, seems like I forgot to add some details here...
  5. Benni Mack commented on patch set 1 03/02/23, 07:00

    any chance we can make this non-breaking? e.g. by a) setting DefaultConfiguration to false, but to true in FactoryConfiguration b) provide an UpgradeWizard (or 2, one for BE and FE) to enable the flag with a note on what this change means?
  6. Oliver Hader commented on patch set 1 03/02/23, 07:57

    Reply

    Factory default sounds good to me
  7. Oliver Hader uploaded patch set 2 V +1 03/02/23, 08:07

  8. Oliver Hader commented on patch set 2 03/02/23, 08:08

    Reply

    Done, now only enabled via factory defaults for the backend scope

    Reply

    Done

    Important-99499-EnforceBackendContent-Security-Policy.rst:13 Reply

    Done, file has been removed
  9. Georg Ringer voted on patch set 2 CR +1 03/02/23, 08:13

  10. core-ci voted on patch set 2 V +1 03/02/23, 08:14 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38293

  11. Christian Kuhn commented on patch set 1 03/02/23, 08:17

    ContentSecurityPolicyHeaders.php:34

    final, probably. there shouldn't be a reason to extend middlewares.

    ContentSecurityPolicies.php:15

    a summary what this default be configuration does and achieves wouldn't hurt i suppose.

    RequestId.php:27

    i really prefer non-public properties paired with getters: those are much easier to deprecate in the getter in case we ever need to get rid of them. i think we should mainly try to avoid public properties throughout the framework for this reason.

    RequestId.php:32

    this is not identical to before: $this->short is now a raw binary string, while md5() by default returns character range.

    PageRenderer.php:844

    could we maybe protect this? or maybe at least mark @internal? feels as if it would be better if we don't directly make this API, or is there no way around that in PageRenderer? Also, a comment this thing is CSP related wouldn't hurt.

    PolicyMutatedEvent.php:36

    as with other classes: i'd really prefer protected / private properties to mitigate possible future deprecation problems. have getters for listeners instead.

    ModelService.php:25

    maybe final, maybe @internal for now?

    Mutation.php:26

    maybe final, maybe @internal for now?

    Mutation.php:31

    as with other classes: i'd really prefer protected / private properties to mitigate possible future deprecation problems.

    MutationCollection.php:20

    a class comment on where / how this class is used could be helpful. maybe final, maybe @internal for now?

    MutationCollection.php:25

    as with other classes: i'd really prefer protected / private properties to mitigate possible future deprecation problems.

    MutationMode.php:20

    a comment on where / how this is used and what this class is good for could be helpful. maybe final, maybe @internal for now?

    Policy.php:27

    final probably, @internal eventually (at least for now?)

    PolicyProvider.php:30

    final maybe, @internal eventually (maybe at least for now?)

    PolicyProvider.php:32

    as with other classes: i'd really prefer protected / private properties to mitigate possible future deprecation problems.

    RawValue.php:26

    finaly probably?

    Scope.php:29

    final, probably? @internal eventually?

    SourceCollection.php:22

    final eventually? maybe @internal (in general: see where we could / should @internal throughout the patch) also, a class comment wouldn't hurt, but may not be critical here.

    SourceCollection.php:27

    as with other classes: i'd really prefer protected / private properties to mitigate possible future deprecation problems.

    WeakMapUtility.php:21

    using the class name as explanation of the class does not add much value. could we add some more info here on what this class is for, where it's typically used and in which scope it's supposed to be used?

    WeakMapUtility.php:23

    final? it's static anyways ...

    ContentSecurityPolicies.php:14

    a summary comment on the goal of this backend related configuration could be helpful.

    NonceViewHelper.php:54

    do not rely on $GLOBALS['TYPO3_REQUEST'], please refactor so it's not needed.

    ContentSecurityPolicyHeaders.php:34

    could be declared final as well? there shouldn't be a good reason to extend the middleware.

    ContentSecurityPolicies.php:15

    a comment outlining goals of this default FE configuration could be helpful.
  12. Christian Kuhn commented on patch set 2 CR -1 03/02/23, 08:18

    -1 for now ... had a rough look on reading and left a series of comments.
  13. Oliver Hader uploaded patch set 3 CR +1 V +1 03/02/23, 10:06

  14. Oliver Hader commented on patch set 2 03/02/23, 10:09

    Concerning using `public readonly $property` we should find a general consensus with all devs. I see the point that deprecating properties gets a bit more complicated (still, it is still possible though). On the other hand a lot of stupid boiler plate code can be omitted, which is just used for the eventual case that something might(!) be deprecated in the future.

    ContentSecurityPolicyHeaders.php:34 Reply

    Fine, I don't have a preference here. Are there other middlewares that are `final` already?

    ContentSecurityPolicies.php:15 Reply

    Like reading all the comments below instead?

    RequestId.php:27 Reply

    I really prefer public readonly properties instead of using superfluous boilerplate code with getters... in this scenario `RequestId` is marked `@internal`. Why should that be deprecated instead of removed/adjusted?! Anyway, if it has to be getters, so be it... I don't really care...

    RequestId.php:32 Reply

    No true, see the `bin2hex`

    PageRenderer.php:844 Reply

    I'll drop it...

    ModelService.php:25 Reply

    Done

    Mutation.php:26 Reply

    People are supposed to use it... ´@internal` would not make much sense here

    MutationCollection.php:20 Reply

    Done

    MutationMode.php:20 Reply

    Enums can be `final`?

    Policy.php:27 Reply

    `@internal` is fine, but not `final`.

    PolicyProvider.php:30 Reply

    `final` since it's an injected provider

    RawValue.php:26 Reply

    No

    Scope.php:29 Reply

    Done

    SourceCollection.php:22 Reply

    Done

    WeakMapUtility.php:21 Reply

    I really have no clue what you're expecting here? Please provide a suggestion instead of ranting, thx!

    WeakMapUtility.php:23 Reply

    Done

    ContentSecurityPolicies.php:14 Reply

    Done

    NonceViewHelper.php:54 Reply

    How to pass that kind of information globally then? There are still zillions of usages for `site`, `normalizeParams`, `lang` and any other stuff that is relying on `$GLOBALS['TYPO3_REQUEST']`... I dropped it with being less flexible as a consequence...

    ContentSecurityPolicyHeaders.php:34 Reply

    Fine, I don't have a preference here. Are there other middlewares that are `final` already?

    ContentSecurityPolicies.php:15 Reply

    Done
  15. core-ci voted on patch set 3 V +1 03/02/23, 10:13 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38296

  16. Oliver Hader uploaded patch set 4 V +1 03/02/23, 11:28

  17. core-ci voted on patch set 4 V +1 03/02/23, 11:35 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38297

  18. Christian Kuhn commented on patch set 4 03/02/23, 11:54

    Reply

    I expected an answer like that. I think we should raise this in coredev channel to decide on this, there are reasons for both. I guess we agree that we have to be able to deprecate things, even if we currently don't know things have to be deprecated later again. You're right: access to public (readonly?!) typed properties *can* be deprecated, we have this public property trait for it. However, this trait is pretty ugly, it needs a class property with a list of properties, and it (afaik?) does not work with static stuff. It's easier with getters: add a trigger_error() to the method with some deprecation message and done. And it's obvious on reading. From a framework point of view, that's more elegant. Until now, we rather tried to avoid public API properties in new code, but I don't insist on this view. Let's get a decision via chat once and for all, and if it's fine, then maybe with an advise on when it's ok (maybe for readonly properties only) and when it's not ok?

    ContentSecurityPolicyHeaders.php:34 Reply

    existing middlewares may not be all final currently, but new ones definately should be, imho.

    RequestId.php:27 Reply

    let's continue on the patch set comment and mark this comment done.

    Mutation.php:26 Reply

    ack.

    MutationMode.php:20 Reply

    actually unsure. keep as is, sorry.

    WeakMapUtility.php:21 Reply

    ok, ignore. I've read up on \WeakMap again and understand it better now. It's a rather new and not so well known construct, so I thought some explanation on it wouldn't hurt.

    ContentSecurityPolicyHeaders.php:34 Reply

    existing ones are probably not (all?) final, but new ones should be.
  19. Oliver Hader commented on patch set 4 03/02/23, 12:40

    Reply

    I'm just pasting an example here, to get a feeling what it would take to deprecate a `public readonly $property`. Using `PublicPropertyDeprecationTrait` is a bit different, but from my point of view not complicated. Base class: ``` class DemoSubject { public readonly string $value; public function __construct() { $this->value = 'demo'; } } ``` Deprecating `$value` requires to add the trait, change `public → protected`, and finally add `@deprecated` and a corresponding message (last aspect would be the same for `getValue()` as well): ``` class DemoSubject { use PublicPropertyDeprecationTrait; /** @deprecated */ protected readonly string $value; /** @var array<string, string> */ protected array $deprecatedPublicProperties = [ 'value' => self::class . '::$value is deprecated', ]; public function __construct() { $this->value = 'demo'; } } ``` And the same deprecation when using `getValue()` instead of direct `$subject->value` access: ``` class DemoSubject { /** @deprecated */ protected readonly string $value; public function __construct() { $this->value = 'demo'; } /** @deprecated */ public function getValue(): string { trigger_error(self::class . '::$value is deprecated', E_USER_DEPRECATED); return $this->value; } } ```

    ContentSecurityPolicyHeaders.php:34 Reply

    Ack

    MutationMode.php:20 Reply

    Enums cannot be extended, they do not support the concept of inheritance.

    ContentSecurityPolicyHeaders.php:34 Reply

    Ack
  20. Oliver Hader commented on patch set 4 03/02/23, 12:52

    Reply

    Finally, I hope that PHP folks would pick up https://wiki.php.net/rfc/property_accessors again for future versions.
  21. Christian Kuhn commented on patch set 4 03/03/23, 10:01

    NonceViewHelper.php:54 Reply

    request is incoming in arguments with fluid typo3 context. see other VH's
  22. Oliver Hader updated the commit message in patch set 5 V +1 03/07/23, 10:33

  23. Georg Ringer voted on patch set 5 V +1 CR +1 03/07/23, 11:46

  24. core-ci voted on patch set 5 V -1 03/07/23, 11:46 Core CI is not happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38588

  25. core-ci commented on patch set 5 03/07/23, 12:28 Core CI is not happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38588

  26. Oliver Hader rebased onto patch set 6 CR +1 V +1 03/07/23, 15:14

  27. core-ci voted on patch set 6 V +1 03/07/23, 15:21 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38609

  28. Georg Ringer voted on patch set 6 V +1 CR +1 03/07/23, 15:31

  29. Oliver Hader uploaded patch set 7 03/07/23, 17:37

    Copied Votes: * Code-Review+1 (copy condition: "changekind:NO_CHANGE OR is:MIN") * Verified+1 (copy condition: "changekind:NO_CHANGE")
  30. Benjamin Franzke voted on patch set 7 CR +1 03/07/23, 17:39

  31. core-ci voted on patch set 7 V +1 03/07/23, 17:55 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38618

  32. Markus Klein commented on patch set 7 03/08/23, 14:30

    JavaScriptRenderer.php:110

    ?string
  33. Markus Klein commented on patch set 7 03/08/23, 14:36

    Feature-99499-IntroduceContent-Security-PolicyHandling.rst:16

    can

    Feature-99499-IntroduceContent-Security-PolicyHandling.rst:27

    flags
  34. Markus Klein commented on patch set 8 CR +1 V +1 03/08/23, 14:41

    Published edit on patch set 7.
  35. Markus Klein commented on patch set 8 03/08/23, 14:41

    JavaScriptRenderer.php:110 Reply

    Done

    Feature-99499-IntroduceContent-Security-PolicyHandling.rst:16 Reply

    Done

    Feature-99499-IntroduceContent-Security-PolicyHandling.rst:27 Reply

    Done
  36. core-ci voted on patch set 8 V +1 03/08/23, 14:48 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38657

  37. Oliver Hader uploaded patch set 9 V +1 03/08/23, 18:10

  38. core-ci voted on patch set 9 V -1 03/08/23, 18:40 Core CI is not happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38665

  39. Oliver Hader uploaded patch set 10 V -1 03/09/23, 08:53

  40. core-ci voted on patch set 10 V -1 03/09/23, 09:08 Core CI is not happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38680

  41. Torben Hansen commented on patch set 10 03/09/23, 09:13

    I tried to verify, that a custom CSP will have a higher priority that a CSP defined in configuration trough this feature. In frontend scope, this works fine and a CSP header defined in `config.additionalHeaders` TypoScript has priority and is used instead of one configured in site settings. For backend scope, this does not seem to work in my test environment. I have configured a custom CSP for the backend using the following settings in `additional.php`: ``` $GLOBALS['TYPO3_CONF_VARS']['BE']['HTTP']['Response']['Headers'] = [ 'Content-Security-Policy: default-src \'self\'', ]; ``` I would now expect, that my custom CSP will be used, since the `ContentSecurityPolicyHeaders` middleware is called after the `AdditionalResponseHeaders` middleware. This is however not the case and the response object has now 2 CSP headers (my own and the one from the `ContentSecurityPolicyHeaders`) resulting in a CSP header containing a mix of my custom and the `ContentSecurityPolicyHeaders` CSP.. When `$response = $handler->handle($request);` is called in `AdditionalResponseHeaders` middleware, the `ContentSecurityPolicyHeaders` middleware is executed and at this point, no configured BE response headers have been added.
  42. Oliver Hader commented on patch set 10 03/09/23, 10:49

    Reply

    Good catch! Both middlewares are focussing on the response (not on the request). Thus, their order had to be switched.
  43. Oliver Hader uploaded patch set 11 V -1 03/09/23, 10:49

  44. Oliver Hader uploaded patch set 12 03/09/23, 11:09

  45. core-ci voted on patch set 11 V -1 03/09/23, 11:11 Core CI is not happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38691

  46. core-ci voted on patch set 12 V +1 03/09/23, 11:16 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38692

  47. Benjamin Kott uploaded patch set 13 V +1 03/09/23, 16:32

  48. core-ci voted on patch set 13 V -1 03/09/23, 16:41 Core CI is not happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38705

  49. Oliver Hader uploaded patch set 14 V -1 03/09/23, 16:44

  50. core-ci voted on patch set 14 V +1 03/09/23, 16:51 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38707

  51. Oliver Hader commented on patch set 14 03/09/23, 17:09

    One thing is missing for `lit` elements defining their own styles, see https://github.com/lit/lit/pull/2134
  52. Oliver Hader uploaded patch set 15 V +1 03/10/23, 10:26

  53. core-ci voted on patch set 15 V -1 03/10/23, 11:21 Core CI is not happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38763

  54. Oliver Hader uploaded patch set 16 V -1 03/10/23, 15:04

  55. core-ci voted on patch set 16 V +1 03/10/23, 15:13 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38775

  56. Markus Klein commented on patch set 16 03/10/23, 21:21

    RequestMiddlewares.php:101

    Why is this adjustment needed?
  57. Oliver Hader commented on patch set 16 03/10/23, 21:28

    RequestMiddlewares.php:101 Reply

    `typo3/cms-backend/response-headers` might extend the response headers with `Content-Security-Policy`, `typo3/cms-backend/csp-headers` is checking that previous response. Middleware order is focussed on the request(!), however when focussing on the response(!) the order is in reverse order. The onion chain looks like this: ``` invoke typo3/cms-backend/csp-headers → handle request(!) → invoke typo3/cms-backend/response-headers → handle request(!) → invoke additional middlewares/handlers → ... → extend response(!) → check and extend response(!) ```
  58. Oliver Hader commented on patch set 16 03/10/23, 21:31

    RequestMiddlewares.php:101 Reply

    for response-propagation it also could stay like that, but then csp-headers needs to have an additional order setting concerning response-propagation - which seems to be better → I'll adjust it there
  59. Markus Klein commented on patch set 16 03/10/23, 21:37

    RequestMiddlewares.php:101 Reply

    I'm fully aware of the order principle here. But currently we have a strict order relationship between response-headers -> response-propagation. If this relation is relaxed, the order of these two might switch inadvertently. That's why I was asking. Just to avoid side effects
  60. Oliver Hader commented on patch set 16 03/10/23, 21:39

    RequestMiddlewares.php:101 Reply

    I see. I'm changing this to use only `after` and omitting confusion with `before`.
  61. Oliver Hader uploaded patch set 17 V +1 03/10/23, 21:40

  62. Oliver Hader commented on patch set 17 03/10/23, 21:43

    RequestMiddlewares.php:101 Reply

    Done
  63. Markus Klein commented on patch set 17 CR +1 03/10/23, 21:43

    RequestMiddlewares.php:101 Reply

    Ack
  64. core-ci voted on patch set 17 V +1 03/10/23, 21:48 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38779

  65. Oliver Hader commented on patch set 17 03/10/23, 22:09

    Services.yaml:528

    `shared` can be removed
  66. Torben Hansen commented on patch set 17 V +1 CR +1 03/11/23, 06:16

    Feature-99499-IntroduceContent-Security-PolicyHandling.rst:93

    Unused import
  67. Oliver Hader uploaded patch set 18 CR +1 V +1 03/11/23, 11:06

  68. Oliver Hader commented on patch set 18 03/11/23, 11:10

    Services.yaml:528 Reply

    Done

    Feature-99499-IntroduceContent-Security-PolicyHandling.rst:93 Reply

    Done
  69. core-ci voted on patch set 18 V -1 03/11/23, 12:14 Core CI is not happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38790

  70. Oliver Hader uploaded patch set 19 V -1 03/11/23, 12:51

  71. core-ci voted on patch set 19 V +1 03/11/23, 12:58 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38792

  72. Markus Klein voted on patch set 19 CR +1 03/11/23, 13:24

  73. Torben Hansen voted on patch set 19 V +1 CR +1 03/11/23, 13:41

  74. Markus Klein voted on patch set 19 V +1 03/13/23, 11:56

  75. Georg Ringer voted on patch set 19 V +1 CR +1 03/14/23, 04:06

  76. Oliver Hader commented on patch set 19 CR -1 03/14/23, 19:54

    Okay, \WeakMap is probably not what I thought it is... PHP emphasized "weak" over "map": → https://3v4l.org/TSO37#v8.1.16 It works for enums as keys, but real objects will disappear when leaving the call-hierarchy scope.
  77. Oliver Hader commented on patch set 19 V -1 03/14/23, 21:20

    → `\WeakMap` replacement, please review https://review.typo3.org/c/Packages/TYPO3.CMS/+/78127 first
  78. Oliver Hader uploaded patch set 20 CR +1 V +1 03/14/23, 21:37

  79. Oliver Hader uploaded patch set 21 03/14/23, 21:40

  80. core-ci voted on patch set 21 V +1 03/14/23, 21:50 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38951

  81. Oliver Hader rebased onto patch set 22 V +1 03/14/23, 22:41

  82. core-ci voted on patch set 22 V +1 03/14/23, 22:49 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38954

  83. Oliver Hader rebased onto patch set 23 V +1 03/15/23, 09:15

  84. Benni Mack commented on patch set 23 03/15/23, 09:21

    ApplicationScope.php:23

    we already have something called "ApplicationType" - should re-use that? Or at least bring them closer together? I mean, we use the abbrevation extensively everywhere. If you want to keep the applicationScope enum, we should at least move it to TYPO3\CMS\Core\Core as this thing does not have anything to do with Http.
  85. core-ci voted on patch set 23 V +1 03/15/23, 09:22 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38961

  86. Oliver Hader uploaded patch set 24 03/15/23, 09:25

    Copied Votes: * Verified+1 (copy condition: "changekind:NO_CHANGE")
  87. core-ci voted on patch set 24 V +1 03/15/23, 09:36 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38962

  88. Oliver Hader commented on patch set 24 03/15/23, 09:39

    ApplicationScope.php:23 Reply

    `ApplicationType->$type` is not specific enough, it has a bit-mask that in theory would allow all values. Thus, current `ApplicationType` could be frontend, backend, CLI and install-tool at the same time - I want to avoid that ;-) In general I like the idea of having just one `ApplicationType` or `ApplicationScope` from a semantic point of view.
  89. Oliver Hader uploaded patch set 25 V +1 03/15/23, 09:48

  90. core-ci voted on patch set 25 V +1 03/15/23, 09:57 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38966

  91. Oliver Hader commented on patch set 25 03/15/23, 10:21

    ApplicationScope.php:23 Reply

    I've created a draft at https://review.typo3.org/c/Packages/TYPO3.CMS/+/78129 changing `ApplicationType` from `class` to `enum`. Which seems doable, since the constructor was private and the `$type` property was not exposed.
  92. Oliver Hader commented on patch set 25 03/15/23, 10:54

    ApplicationScope.php:23 Reply

    Done
  93. Oliver Hader uploaded patch set 26 V +1 03/15/23, 10:55

  94. core-ci voted on patch set 26 V +1 03/15/23, 11:03 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38973

  95. Oliver Hader commented on patch set 27 V +1 03/15/23, 11:17

    Published edit on patch set 26.
  96. core-ci voted on patch set 27 V +1 03/15/23, 11:24 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38975

  97. Markus Klein commented on patch set 27 03/15/23, 11:59

    AbstractServiceProvider.php:168

    ?string
  98. Markus Klein commented on patch set 27 03/15/23, 12:00

    my comment is just one example, we have multiple places where we have optional params where the type is not nullable. I guess we should have nullable, right?
  99. Oliver Hader commented on patch set 27 03/15/23, 12:09

    Reply

    What file/class are you referring to?
  1. Oliver Hader commented on patch set 27 03/15/23, 12:17

    Reply

    Ah got it, did not see the other comment. `function(string $value = null)` is okay in current PHP versions. I guess that might change in future PHP version to be strict and use only `?string $value = null`. That also would be something for php-cs-fixer.
  2. Oliver Hader uploaded patch set 28 V +1 03/15/23, 12:17

  3. Oliver Hader commented on patch set 28 03/15/23, 12:18

    Extracted site configuration from `config.yaml` to new `csp.yaml` file.
  4. Markus Klein commented on patch set 28 03/15/23, 13:03

    Reply

    Ack
  5. Markus Klein commented on patch set 28 03/15/23, 13:04

    AbstractServiceProvider.php:168 Reply

    Ack
  6. core-ci voted on patch set 28 V -1 03/15/23, 13:09 Core CI is not happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38979

  7. Oliver Hader uploaded patch set 29 V -1 03/15/23, 13:22

  8. Oliver Hader commented on patch set 29 03/15/23, 13:23

    Classic! I forgot to amend the the new class...
  9. core-ci voted on patch set 29 V +1 03/15/23, 13:30 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/38983

  10. Benni Mack uploaded patch set 30 V +1 03/15/23, 19:46

  11. Benni Mack uploaded patch set 31 03/15/23, 19:54

  12. core-ci voted on patch set 30 V -1 03/15/23, 20:01 Core CI is not happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/39004

  13. Benni Mack uploaded patch set 32 03/15/23, 20:17

  14. Benni Mack commented on patch set 32 03/15/23, 20:19

    PolicyProvider.php:93

    I wonder if we should inheritDefault as "true" by default. WDYT
  15. Benni Mack commented on patch set 32 03/15/23, 20:19

    Reply

    Not relevant now.
  16. core-ci voted on patch set 31 V -1 03/15/23, 20:23 Core CI is not happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/39005

  17. core-ci voted on patch set 32 V +1 03/15/23, 20:24 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/39008

  18. Oliver Hader uploaded patch set 33 V +1 03/15/23, 20:56

  19. Oliver Hader commented on patch set 32 03/15/23, 20:57

    PolicyProvider.php:93 Reply

    👍
  20. core-ci voted on patch set 33 V +1 03/15/23, 21:04 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/39011

  21. Benni Mack voted on patch set 33 V +1 CR +1 03/15/23, 21:35

  22. Markus Klein voted on patch set 33 CR +1 03/15/23, 21:56

  23. Markus Klein commented on patch set 33 V +1 03/15/23, 22:47

    I tested as much as I could, lets get this in and tweak things afterwards if anything should pop up. Thanks for all the work!
  24. Oliver Hader commented on patch set 33 V +2 CR +2 03/16/23, 00:31

    Thanks for all of your reviews, tests and feedback! 👌
  25. Oliver Hader merged patch set 34 03/16/23, 00:32