HV-1831 Unfinished experiments #1157
Conversation
gsmet
force-pushed
the
HV-1831-experiments-6.2
branch
2 times, most recently
from
0107720 to
51b94c5
Compare
gsmet
force-pushed
the
HV-1831-experiments-6.2
branch
from
51b94c5 to
d4acd34
Compare
| if ( !executableMetaData.isPresent() ) { | ||
| // the method is unconstrained so there's no need to worry about the tracking | ||
| return false; | ||
| } | ||
|
|
||
| return !executableMetaData.get().getValidatableParametersMetaData().hasCascadables(); | ||
| return processedBeansTrackingStrategy.isEnabledForReturnValue( executable, |
There was a problem hiding this comment.
Should the above line be:
return processedBeansTrackingStrategy.isEnabledForParameters( executable,
?
There was a problem hiding this comment.
You're absolutely right, I copy/pasted things a bit too fast.
| processedPaths = new HashSet<>(); | ||
| processedPathsPerBean.put( bean, processedPaths ); | ||
| processedPaths = new ArrayList<>(); | ||
| processedPathsPerBean.put( processedBean, processedPaths ); | ||
| } | ||
|
|
||
| processedPaths.add( PathImpl.createCopy( path ) ); |
There was a problem hiding this comment.
While debugging ProcessedBeansTrackingCycles1Test, I see that ValidatorImpl#validateConstraintsForDefaultGroup executes this code (in #markCurrentBeanAsProcessedForCurrentPath) twice for the same Parent bean ; first time for Parent.class, and second time for Object.class. Both times #markCurrentBeanAsProcessedForCurrentPath is called with the same Parent bean and the same PathImpl object as arguments.
Since processedPaths is an ArrayList<PathImpl>, it contains 2 copies of the same PathImpl. Was this intended?
I see that processedPaths was changed from a Set<PathImpl> to an ArrayList<PathImpl>, so before this change, (I think) there would have been just one PathImpl in Set.
I don't know if this is a real problem; it's just something I noticed.
There was a problem hiding this comment.
Hmmm. That's unfortunate. It was part of another experiment and shouldn't have been committed with this one.
How do you want us to proceed? Should I fix things and force push? Or we fix it later? Basically, everything in this class should be reverted, except for the change of meaning of the boolean.
This is a very rough experiment based on the 6.2 branch. It's not finished at all and I have no idea if we can make it work in the end.