Community
Participate
Working Groups
Setting 'Missing Javadoc comments' (under the Javadoc tab of compiler preferences, either globally in the preferences or as specific settings for a project) is a bit overzealous in what it does; most notably the following two situations are undesired: 1. Any method of a higher visibility than the containing class (ie: a public void run() in a inner private class MyThread extends Thread) is marked as warning or error for a missing javadoc according to the higher visibility. This is undesired; in the previous example, if javadocs are generated only for protected and public elements, the public void run() in this inner private class will not show up anywhere in the generated javadocs. This error also occurs with private enums; the valid values for the enum are all marked as warnings even if you've set the 'only consider members as visible as' selector to Default or higher. 2. There is no option to specify if enum values should count as public declarations; Right now all enumeration values in an enum are considered public fields and hence always generate a warning/error if the 'Missing javadoc comments' option is set to cause warnings or errors. I'd like to see a new checkbox under 'Missing Javadoc comments': 'Consider enum values'. Regardless of this feature, enums suffer the same bug/'misfeature' as listed in #1; even for private enums, the contained values are marked as public for purposes of determining if it needs a javadoc comment. Last checked in 3.1M5a.
I'll fix problem 1) using this bug but not sure for point 2) (only if time permit). To make it easier, please reopen a separated bug for this enhancement: 'Consider enum values' for 'Missing Javadoc Comments'... Thx
(In reply to comment #1) > I'll fix problem 1) using this bug but not sure for point 2) (only if time permit). > To make it easier, please reopen a separated bug for this enhancement: 'Consider > enum values' for 'Missing Javadoc Comments'... > Thx Right. Thanks for fixing #1. I'll make a new feature request report for #2 now.
For new javadoc missing comment option for enum see bug 87500
Current (M6) effort is done on bugs which need API changes => move to M7
Unfortunately, it was too late to finalize fix for point 1). This kind of change is too risky at this late stage... So, set as LATER, sorry about that. Please reopen after 3.1 delivery
*** Bug 101829 has been marked as a duplicate of this bug. ***
*** Bug 104419 has been marked as a duplicate of this bug. ***
Please re-open this bug since it is really annoying!
As per earlier instructions, bug has been re-opened as 3.1 final is out.
Reinier, Bug status is set by team member and should not modified by users. As they are now several duplicate for this bug, there would be high chance that we reopened it, but please let us making our plan for next release and change it by ourself... Instead, you'd better increase severity to let us know that this bug seems to be more important for users than it was when you opened it. Thanks
Reiner, You did the right thing reopening this bug as I asked you to do it in comment 5... My fault, I thought you did it after comment 8 request. Forget my previous comment and sorry for confusion.
Increase severity as this issue is really annoying for some users (see bug 104419).
Created attachment 36881 [details] Patch to fix this issue This patch can be applied on top of JDT/Core v_648 version... Need to verify that there's no performance impact before releasing this change...
Philippe, what's your mind about this change?
Fix is not valid: it would be better to browse scopes until top level and rely on binding instead of looking at modifiers and add a new field... Unfortunately 3.2 endgame plan too short to continue and implement this new version, so I add to defer this fix again to next version.
Reopen as LATER is deprecated...
*** Bug 87500 has been marked as a duplicate of this bug. ***
Created attachment 77448 [details] patch + test case Jerome, Frederic: may you please review proposed patch - thanks
(In reply to comment #19) > Created an attachment (id=77448) [details] > patch + test case > > Jerome, Frederic: may you please review proposed patch - thanks > ProblemReporter.javadocMissing(...) is unnecessarily called when severity == ProblemSeverities.Ignore. Otherwise, it's OK for me
Created attachment 77452 [details] patch + test case (In reply to comment #20) improved version + code factorization as suggested by Jerome.
(In reply to comment #21) Patch looks good to me, so +1 for this patch
(In reply to comment #22) > (In reply to comment #21) > Patch looks good to me, so +1 for this patch > Also +1 for this patch
Released for 3.4M3 in HEAD stream
Verified for 3.4M3 using I20071029-0010 build.