On Jan 17, 5:22 am, Marc Guillemot <[EMAIL PROTECTED]> wrote: > Issue created with patch:https://bugzilla.mozilla.org/show_bug.cgi?id=412755 > > Additional question concerning the tests: > my patch contains a JUnit test case for the new error messages but it > would probably make more sense for such situation to have within Rhino a > folder of tests like those from mozilla/js/tests. What do you think? > > Cheers, > Marc. > -- > Blog:http://mguillem.wordpress.comAttila Szegedi wrote: > > Keep the old signature, and add an overload with the new signature. > > > Attila. > > > On 2008.01.17., at 9:02, Marc Guillemot wrote: > > >> yyy would be for instance the string representation on which the method > >> was called. > > >> Adding this can only be done in changing the signature of > >> ScriptRuntime#notFunctionError to accept an additional parameter but you > >> seem to say that it shouldn't be done to allow script compiled with > >> earlier versions of Rhino should continue to work. Is this position > >> definitive? I'm quite skeptic concerning such things that a project > >> wants to maintain but that are not covered by tests. Are you really sure > >> that the changes made between previous releases respect that? > > >> Yes, I want to propose a patch but I first need to know what kind of > >> changes are allowed. > > >> Cheers, > >> Marc. > >> -- > >> Blog:http://mguillem.wordpress.com > >> _______________________________________________
Thank you, Marc. I'm trying to finish up 1.7R1 release and this is a significant enough change I'd like to put it in 1.7R2. We do need thread safety in the default implementation as the previous implementation was used in multithreaded environments. I can add that in and do some additional performance tests. --N _______________________________________________ dev-tech-js-engine-rhino mailing list [email protected] https://lists.mozilla.org/listinfo/dev-tech-js-engine-rhino
