Closed jarmoniuk closed 7 months ago
fullTreeMatch, which should by default be false as I don't suppose it to be used very often.
maybe not add new parameter ... when we add we also should add a test which show what for and what for cases it is .used ..
1042 is a pretty particular case which exposes a bug in the set logic.
The new version is initially set "to be tried out". Then, it is being set for the second time when processing changes. However, there's also some logic which is supposed to see if there are modules using the module that is being changed. This logic was responsible for the reversed change in the case of this particular issue.
When investigating, I've noticed that
processAllModules
has two meanings: one is to process two modules regardless whether they match the current module; the second makes the matching logic traverse the whole reactor instead of the current module's subtree in search of modules which use the current module as parent. Because of that, I have introduced a new option,fullTreeMatch
, which should by default befalse
as I don't suppose it to be used very often.@slawekjaranowski please review.