I assume there is no regression in the classpath scanning, just a limitation in the indexer? This still worth addressing but not a regression in the usual sense.
I do not want (cannot) know all the modules (they might be hundreds) that will be on my classpath so supporting wildcards for jpa entity and repositories is a must.
Using a PathMatcher instead of just startsWith in the indexer introduced with #16509 should not hurt performance much and is very handy when you rely on a naming convention in a project with a lot of modules.
Petar Tahchiev opened SPR-16152 and commented
Looking at the latest 5.0.1
CandidateComponentIndexer
code:https://github.com/spring-projects/spring-framework/blob/master/spring-context/src/main/java/org/springframework/context/index/CandidateComponentsIndex.java#L68
candidates are filtered based on the
startsWith
criteria. This prevents packages containing wildcards to be scanned:Please note that this is a regression, as this used to work fine with 4.3.x
Affects: 5.0.1
Issue Links:
Referenced from: commits 9511d29, 1838ddb
1 votes, 4 watchers
The text was updated successfully, but these errors were encountered: