- Issue created by @tedbow
- π§πͺBelgium wim leers Ghent π§πͺπͺπΊ
π€
I don't think I agree with this proposal. π Create an API for base requirement validators which run before other validators and stop event propagation if they fail Fixed is already solving this at the root, without the need for every validator to do a
ComposerInspector::validate()
call? - πΊπΈUnited States tedbow Ithaca, NY, USA
will comment on π Create an API for base requirement validators which run before other validators and stop event propagation if they fail Fixed I think that issue will cause other problems
- πΊπΈUnited States phenaproxima Massachusetts
I don't think I agree with this proposal. π Create an API for base requirement validators which run before other validators and stop event propagation if they fail Fixed is already solving this at the root, without the need for every validator to do a ComposerInspector::validate() call?
The whole idea here is that validators won't need to call ComposerInspector::validate(). They will only want to do it if they need to invoke Composer directly, without going through ComposerInspector's domain methods.
- π§πͺBelgium wim leers Ghent π§πͺπͺπΊ
#6 confuses me thoroughly.
#6 says:
The whole idea here is that validators won't need to call ComposerInspector::validate(). They will only want to do it if they need to invoke Composer directly, without going through ComposerInspector's domain methods.
But the issue summary says:
Any other validators that use Composer inspector to get Composer info or make their own calls to it can do something like this beforehand:
try {
$this->composerInspector->validate($dir);
β¦Is that not a contradiction? π€
- πΊπΈUnited States phenaproxima Massachusetts
Right you are. I adjusted the IS to clarify.
- π§πͺBelgium wim leers Ghent π§πͺπͺπΊ
I β¦ don't know if that change is actually accurate.
But either way β I'm looking forward to feedback from you both at #3312960-25: Certain validators should be higher priority and stop further validation β thanks to @tedbow's excellent #3312960-24: Certain validators should be higher priority and stop further validation β he mentioned in #4.
I think this issue ought to be closed as a duplicate in favor of π Create an API for base requirement validators which run before other validators and stop event propagation if they fail Fixed for the reasons cited there.
- Assigned to tedbow
- πΊπΈUnited States tedbow Ithaca, NY, USA
Assigning this to myself because we need to figure out this one and π Create an API for base requirement validators which run before other validators and stop event propagation if they fail Fixed now both assigned to me so hopefully nobody will work on them till I can respond
- @tedbow opened merge request.
- Assigned to phenaproxima
- πΊπΈUnited States phenaproxima Massachusetts
I'm taking this on to keep @tedbow's plate slightly less crowded.
- π§πͺBelgium wim leers Ghent π§πͺπͺπΊ
To make this easier to review: the last commit in patch form with
diff -M1%
, to force it to detect renames. - Status changed to Needs review
over 1 year ago 11:30pm 24 February 2023 -
tedbow β
committed 24ce450a on 3.0.x
Issue #3344039 by phenaproxima, tedbow, Wim Leers: Add a validate()...
-
tedbow β
committed 24ce450a on 3.0.x
- Issue was unassigned.
- Status changed to Fixed
over 1 year ago 11:50pm 26 February 2023 - πΊπΈUnited States tedbow Ithaca, NY, USA
Looks good! Glad we scaled this back for now
- π§πͺBelgium wim leers Ghent π§πͺπͺπΊ
So the "scaling back" is why the
package_manager/src/Validator/ComposerDependentValidatorTrait.php
was removed, and π ComposerInspector::validate() should run `composer validate` Fixed is now supposed to do that, is that right? - π§πͺBelgium wim leers Ghent π§πͺπͺπΊ
Ah, no, it seems that those things were moved to π Create an API for base requirement validators which run before other validators and stop event propagation if they fail Fixed ? π€
A comment indicating in which way this was scaled back and where (which issues) to find where that work was shifted to, would've been very helpful π
- Status changed to Needs work
over 1 year ago 12:50pm 27 February 2023 - π§πͺBelgium wim leers Ghent π§πͺπͺπΊ
This MR in the way it was committed removed
ComposerJsonExistsValidator
, and thestopPropagation()
in it.But it did not remove
ComposerExecutableValidator
like the issue summary proposed.It also did not add a
ComposerValidator
like the issue summary stated.It also did not update
package_manager.api.php
, AGAIN, which is why it still says* be stopped to prevent further event subscribers from breaking. For example, * Package Manager stops event propagation if: β¦ * - No composer.json file exists in active directory. πππ
Regressions like this keep making π Define the Package Manager API (package_manager.api.php is outdated) Fixed impossible to finish. I'll fix this too there, but please still either add a change record or update the issue summary, so that everybody working on Automatic Updates is clear on the direction without having to re-read a lot of code (and the same for the folks working on Project Browser).
- @tedbow opened merge request.
- π§πͺBelgium wim leers Ghent π§πͺπͺπΊ
Also, why was
ComposerExecutableValidatorTest
removed if we still have that validator? π€ At minimum, theassertResultsWithHelp()
coverage is not included inComposerInspectorTest
? - πΊπΈUnited States tedbow Ithaca, NY, USA
I added new merge request to rename the class and the service which we should have done here
- π§πͺBelgium wim leers Ghent π§πͺπͺπΊ
RE #20: by reviewing π Create an API for base requirement validators which run before other validators and stop event propagation if they fail Fixed , I discovered that MR731 over there is restoring
ComposerJsonExistsValidator(Test)
? π³ - πΊπΈUnited States phenaproxima Massachusetts
- Status changed to Needs review
over 1 year ago 1:43pm 27 February 2023 - πΊπΈUnited States tedbow Ithaca, NY, USA
Needs review for the validator class name change. I think this is more correct because it is not simply just testing executable anymore.
- Status changed to Needs work
over 1 year ago 2:04pm 27 February 2023 - πΊπΈUnited States phenaproxima Massachusetts
There are now merge conflicts that need to be resolved.
-
phenaproxima β
committed 0d0911b8 on 3.0.x
Issue #3344039 by phenaproxima, tedbow: Rename...
-
phenaproxima β
committed 0d0911b8 on 3.0.x
- Status changed to Fixed
over 1 year ago 8:20pm 27 February 2023 Automatically closed - issue fixed for 2 weeks with no activity.