Hi, In the past, javadoc problems were fixed. I'm glad Mark switched validation to checkstyle since I no longer trust the javadoc developers for doing the right thing after the Java 18 changes. Also it allows easy configuration of what is important and what should not be fixed.
Looking at the list, I would propose: - Remove javadoc validation for tests. This would mean doing validation twice (once as usual, another one for the javadoc). Although it is better to document everything, realistically we won't be able to do everything. - checkFirstSentence: After testing, this means that the javadoc first (and quite often only) sentence should end with a period. This seems pointless to me, so I plan to add a comment that this should be set to "false". - checkEmptyJavadoc: Actually this means the description is empty. It happens often for certain obvious methods where @return is documented instead. I'm slightly wavering on that one, in the generated HTML it does indeed look better if both are set (to the same thing as the content of @return - but with a period at the end, see the first property). - MissingJavadocMethod: This one is justified. - MissingJavadocType: Totally justified. - RequireEmptyLineBeforeBlockTagGroup: Should be ignored, another pedantic arbitrary syntax rule. Note: The error counts given include tests, so they're lower than that actually. So: can I proceed with the separation between tests and non tests, and then document why some settings will remain disabled ? This gives a reasonable baseline where it can be assumed the rest is a useful improvement. Rémy --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
