On 12/8/11 3:47 PM, Stuart Marks wrote: > On 12/8/11 9:24 AM, Sean Mullan wrote: >> Now that they are warning free, should there also be a change to the >> Makefile? > > Eventually yes. I'll coordinate with Kurchi on this, but we probably want to > add -Werror to various makefiles in the future, after we do more testing to > ensure that we don't start breaking builds. > >> In javax/script/ScriptException.java, why is the serialVersionUID field not >> marked private? > > No good reason. Well, there's a bad reason :-) which is that the "serialver" > output doesn't make it private. I see that the majority (> 80%) of > serialVersionUID fields in the JDK are private. I'll make this change. > > Aside: it seems that making svuid private is preferred. Is this true, and if > so, what's the rationale? Is this a bug that serialver doesn't make its > declaration private?
Not sure if there is an obvious security risk by making it package-private but I think it's good practice in general to make it private. >> All the other changes look good. The only other suggestion I would make is to >> add a note to the CR with more rationale as to why the SuppressWarnings, etc >> were added to the API so that future maintainers will understand that. In >> short, >> this API was specified as a standalone JSR (105) and any API changes would >> need >> to be first published via a maintenance review per the JCP guidelines. > > OK, I'll add a note to the CR that covers this. > > Anything else? No. --Sean
