Hi peoples, its time for me to say something too ;-)
Well, i understand Stefano.. I think his problem is that we change it in RC and not in beta.. Cause it's a "small" step from RC to final its some kind of "dangerous". Even if its just a small change. This kind of thinkin safe us to indreduce new bugs.. I think he whouldn't have problems with the change if it was made i beta stage. For me the jar removal was ok in beta stage ( like you can see from my vote about this), for others not.. Anyway we should try to not change anything in the code if not needed in RC stage. Anyhow, i don'T think the change will cause any problems. But we should try to not break the (unwritten) rule of not change any code in RC if no critical BUG will be fixed with the change. Maybe the best way to solve such "problems" whould be to try to test earlier the james releases.. So we can do such changes before we reach RC stage.. Don't get me wrong. Its not a personal critic .. its just my thinkin. bye Norman Am Sonntag, den 30.07.2006, 20:40 +0200 schrieb Stefano Bagnara: > As an example you added an addr.toString(): This could raise a > NullPointerException. > I reviewed the code that return the addr variable and I now know that it > never return null but this is *manual* review, *error* *prone*. > > Instead the jar have been removed in trunk since weeks and we run that > code without any problem. I'm still really convinced that removing an > unused jar is MUCH LESS dungerous that changing a single line of code > (even a comment). > > Btw this is personal opinion. I just wanted to put emphasys on the fact > that I don't like this and if I have to accept vetoes to removal of jars > I should be free to cast at least a -0 about this changes without too > much discussions. > > Again, and for the last time: I think that all of this stuff belong to > the RC process, but I simply aligned myself to the fact that other > committers seems to require a more strict approach to changes while in > rc. So this deserve my -0 and my dissappoint. About 2.4 I'll talk after > 2.3.0 final. Stop. > > Stefano > > Noel J. Bergman wrote: > > Stefano Bagnara wrote: > > > >> I still don't get how all of this changes can be less dungerous than > >> removing an unused jar (JAMES-515). > > > > OK, let's compare. I made the following kinds of changes: > > > > - added a few lines of clear code to generate an optional log entry, > > which was helpful in diagnosing a problem > > - changed an incorrect log level from .info() to .debug() > > - corrected the contents of a static text string > > > > Now, while these are changes that should be reviewed, they are also things > > that the compiler can do at least some sanity checking on, and are rather > > different from removing JAR files where even you had some questions to make > > sure that something wasn't being missed, and which someone else might be > > using. The level of visibility to the impact of the change is not the same. > > > > Although JAMES-515 does not really meet the LOW-RISK, HIGH-VALUE criteria, > > I would be OK to make that change for v2.4. > > > > --- Noel > > > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > !EXCUBATOR:1,44ccfd6443381146759681!
signature.asc
Description: Dies ist ein digital signierter Nachrichtenteil
