Community
Participate
Working Groups
The new warnings from bug 349326 can be suppressed in the code using @SuppressWarnings("resource") It would be good if these warnings can also be configured from the command line, thus I propose to support "resource" in: -warn:resource -warn:-resource -warn:+resource
Please also update the batch compiler documentation when the options are added. This is located in org.eclipse.jdt.doc.user/tasks/task-using_batch_compiler.htm (still in CVS). Thanks!
Created attachment 204459 [details] proposed fix Could someone please comment on the wording: ecj -help:warn: .... resource + (potential) leaks of resources of type Closeable
someone ate part of my comment, here's the rest: We could also say "(potentially) unsafe usage of resources of type Closeable" which would better capture also the warning "should be managed with t-w-r" Other suggestions? Another observation: message.properties does not mention "redundantSuperinterface", what happended there?
(In reply to comment #3) > someone ate part of my comment, here's the rest: > > We could also say "(potentially) unsafe usage of resources of type Closeable" > which would better capture also the warning "should be managed with t-w-r" How about "(potential) leakage/unsafe usage of resource of type Closeable" > > Another observation: message.properties does not mention > "redundantSuperinterface", what happended there? It seems like "intfRedundant" now handles that warning.
(In reply to comment #4) > (In reply to comment #3) > > someone ate part of my comment, here's the rest: > > > > We could also say "(potentially) unsafe usage of resources of type Closeable" > > which would better capture also the warning "should be managed with t-w-r" > > How about "(potential) leakage/unsafe usage of resource of type Closeable" that'd cover it all but it sounds a bit complicated to me :) What's the precedence? Is it ( potential (leakage/unsafe usage) ) or ((potential) leakage) / unsafe usage :-/ > > Another observation: message.properties does not mention > > "redundantSuperinterface", what happended there? > > It seems like "intfRedundant" now handles that warning. I see, so redundantSuperinterface is just handled for compatibility, no longer documented because it shouldn't be used anymore? Makes sense. (In reply to comment #1) > Please also update the batch compiler documentation when the options are added. > This is located in org.eclipse.jdt.doc.user/tasks/task-using_batch_compiler.htm > (still in CVS). Thanks! For the records: also "Excluding warnings" needs an update.
(In reply to comment #1) > Please also update the batch compiler documentation when the options are added. > This is located in org.eclipse.jdt.doc.user/tasks/task-using_batch_compiler.htm That page wants a "-" or "+" to indicate the default, however, the new token "resource" is an umbrella for 3 warnings. By default one of them is "warning" the other two are "ignore". Now, how do I document the default for "resource"?
Created attachment 205436 [details] proposed doc update Here's a proposed doc update which marks several option defaults as "+/-" and gives according explanation. Also, all-static-method was missing from the list. Fixed. What do you think?
Looks good. Thanks for fixing the other missing pieces in the doc as well. :) Btw, this patch looks like its coming from the CVS repo. Isn't this doc project now in o.e.platform.common.git?
Stephan please go ahead and release the change. Thanks!
Released for 3.8 M3: JDT/Core commit cb924aa1e344454d0d95de687a37524ed808452e jdt.doc.user commit b938d41e8feb948d7cb2afc5b3027f939688a119
Verified batch compiler and doc for 3.8M3 using build N20111022-2000.