On Thu, 9 Mar 2023 20:06:53 GMT, Jonathan Gibbons <[email protected]> wrote:

>> Pavel Rappo has updated the pull request incrementally with one additional 
>> commit since the last revision:
>> 
>>   Respond to feedback
>
> src/jdk.javadoc/share/classes/jdk/javadoc/internal/doclets/toolkit/util/VisibleMemberTable.java
>  line 647:
> 
>> 645:     }
>> 646: 
>> 647:     private boolean allowInheritedMethod(ExecutableElement 
>> inheritedMethod,
> 
> what does `allow` mean in this context? if it's inherited, doesn't that mean 
> it's allowed? how can an inherited method not be allowed?

While I agree that _allow_ seems strange here, it was here before and it is 
consistent with other methods, such as `allowInheritedMembers` (note "Members", 
not "Methods"). 

Not being a native English speaker, I could imagine that here "allow" means "to 
let inside", "enter" ( see 
https://idioms.thefreedictionary.com/allow+somebody+into ). That method is used 
as a predicate for filtering. It allows some method elements inside a list. 
This would be my best guess.

Anyhow, I suggest that we deal with that name later; there's a bigger issue 
with that method: while it's a predicate, it has side effects, such as 
modifications to `overriddenMethodTable`.

> src/jdk.javadoc/share/classes/jdk/javadoc/internal/doclets/toolkit/util/VisibleMemberTable.java
>  line 699:
> 
>> 697:             ExecutableElement lMethod = (ExecutableElement) le;
>> 698:             // Ignore private methods or those methods marked with
>> 699:             // a "hidden" tag. // FIXME I cannot see where @hidden is 
>> ignored
> 
> Generally, there is `utils.hasHiddenTag`
> 
> It _may_ be a good design question as to when `@hidden` should be taken into 
> account.  For example, should VMT build an accurate model and only suppress 
> on output, or should `@hidden` be reflected in the VMT model.
> 
> If it helps (?) I note that at one time a precursor of `@hidden` was 
> `@treatAsPrivate`

Can we leave it for the next round?

-------------

PR: https://git.openjdk.org/jdk/pull/12887

Reply via email to