txs 4 the review! > But the hotfix also rejects numbers like > 2.22507385850720120e-10 which is not so abnormal. not abnormal but still moderately unlikely.
I agree for a long term scenario. Basically the default should be to disable this workaround and to make it available via configuration. Btw, it seems that Oracle finally reacted and will hopefully ship a fixed JVM 1.6 soon (no help for Java5 users of course). > The fix should also be done for 1.2, because many > productive systems using it. +1 LieGrue, strub --- On Thu, 2/10/11, Udo Schnurpfeil <u...@schnurpfeil.de> wrote: > From: Udo Schnurpfeil <u...@schnurpfeil.de> > Subject: About the JVM bug with 2.2250738585072012e-00308 > To: "MyFaces Development" <dev@myfaces.apache.org> > Date: Thursday, February 10, 2011, 10:59 AM > Hi, > > I've some comments to the JVM bug for the bad number > 2.2250738585072012e-00308 (https://issues.apache.org/jira/browse/MYFACES-3024) > > The problem occures for values which are "very very low". > But the hotfix also rejects numbers like > 2.22507385850720120e-10 which is not so abnormal. > > Would it not be better, when the hotfix is configurable (be > default turned on), so that the admin can switch it off, > when the JVM bugfix is applied? > > The fix should also be done for 1.2, because many > productive systems using it. > > What do you think? > > Regards > > Udo > >