It seems to me that either: a) the new parser is ready-for-prime-time, tested, reliable, trustworthy, preferred going forward, etc.
-or- b) a work-in-progress that isn't reliable for general adoption and people can (and should) experiment with it If a) is true by the 2.1GA release, then it seems fine to me for this to be default and if there are edge-cases that affect a small number of users then the option to revert to the present parser is a perfect fall-back position. But if b) is true and its not ready by the 2.1GA release + reliable for most people to adopt, the the present 'older' parser should be the default so as to break the fewest people's existing work. Bleeding edge or no, this seems to me to be largely about 'ready' or not. -Steve B. On Thu, Apr 16, 2009 at 10:50 AM, Davy Brion <[email protected]> wrote: > using a released version shouldn't be considered as 'living on the edge'... > OSS or not > > > On Thu, Apr 16, 2009 at 4:46 PM, Ken Egozi <[email protected]> wrote: > >> I think that users who are living on the edge in OSS, will have no problem >> with setting a line in the config to get the old parser should they need it. >> >> The users that have trust issues, are using 1.2 anyway. >> >> >> >> >> On Thu, Apr 16, 2009 at 5:44 PM, Davy Brion <[email protected]> wrote: >> >>> i know why i write tests, but i have no idea why people in the past >>> didn't always do so... our tests simply won't catch everything >>> >>> personally, i think the safest choice should be the default... the new >>> parser should IMO become the default after the 2.1 release >>> >>> if this new parser turns out to cause problems in the 2.1 release which >>> could've been avoided by staying with the old parser, our users probably >>> won't have much trust left in the NH team and it won't be easy to win it >>> back either. >>> >>> >>> On Thu, Apr 16, 2009 at 4:33 PM, Fabio Maulo <[email protected]>wrote: >>> >>>> 2009/4/16 Davy Brion <[email protected]> >>>> >>>>> wait, are we really going to have the new AST parser as the default one >>>>> for the 2.1 release? We might have 1700 tests, but our test suite isn't >>>>> really all that great and having the new AST parser as the default at this >>>>> point would be a bit risky i think. >>>> >>>> >>>> If you are afraid you can configure NH to use the "Classic" >>>> queryTranslator (is only one line in the session-factory configuration). >>>> >>>> Now think about this: >>>> Why you are writing tests ? >>>> >>>> Don't be afraid Davy, trust in your self and in the NH's team. >>>> >>>> -- >>>> Fabio Maulo >>>> >>>> P.S. check the log of SVN to understand when 2.0 born and than check >>>> which is the most important issue fixed between 2.0.0 and 2.0.1 (note 2.0.1 >>>> has the record of NH downloads; today is 56412) >>>> >>> >>> >> >> >> -- >> Ken Egozi. >> http://www.kenegozi.com/blog >> http://www.delver.com >> http://www.musicglue.com >> http://www.castleproject.org >> http://www.gotfriends.co.il >> > > -- Steve Bohlen [email protected] http://blog.unhandled-exceptions.com http://twitter.com/sbohlen
