finding teams or people who are willing to use the new parser will definitely not be a problem... at my job we're always using the trunk in about 4 projects concurrently and once the new parser is introduced i'd be more than willing to have those projects use the new parser. I'm sure there will be plenty more who feel the same way.
just to avoid any confusion: my only objection to already setting it as the default is that i think we should give it a bit of time to prove itself and not rush this into a release. i have no doubt that this parser will solve a lot of problems in the future though. oh and welcome aboard :) On Thu, Apr 16, 2009 at 5:38 PM, Steve Strong <[email protected]> wrote: > I'll give my thoughts on this. I think the new parser is in pretty good > shape. Baring a couple of areas which I'll clean up over the next 3 or 4 > weeks (when I can drag myself away from the LINQ code!), it's fairly well > structured, and fixing any bugs that exist should be relatively > straightforward. Note this is no reflection on my abilities, more on the > excellent work done by the original Hb team. > That said, there is a lot of code, and I'm quite sure that bugs exist that > haven't been found by the Nh test suite. Clearly shipping a new parser that > causes previously working code to break is going to cause some issues in the > community, so I think it's only right and proper to put this to a vote on > the NhUsers list and get their views. If they vote for the new parser to > become the default, and then issues are found, we can work swiftly to > resolve them knowing that the community understand why the bugs have > occurred, and knowing that we have their support. > > If they vote to keep the existing parser for now, then we need to find as > many "friendly" teams as possible to switch from the default so that we can > get the new parser battle-hardened, and then switch it to be the default in > the future. > > Just my 2 cents. I'm new here, so feel free to hit "delete" :) > > Cheers, > > Steve > > On 16 Apr 2009, at 17:14, Fabio Maulo wrote: > > 2009/4/16 Ken Egozi <[email protected]> > >> >> Stephen's argument is important. However as afaik from what Steve has >> written here - the vast majority of tests pass with flying colours. the only >> things failing are either "tests should be changed", or "this is such an >> edge case that users who are in these realms are usually of the supporting >> kind, and will have no problem with setting the backward-compat config >> option" >> > > > Yes, "tests should be changed" for that I ask the patch to analyze each > breaking-change. > About users... > NH 2.1.0 Alpha2 has 2764 downloads and I'm sure many people are using the > trunk so we have a lot of users using the "in work" version and running > their tests. > This matter, probably, is a big part of the success of each NH official > release. > > -- > Fabio Maulo > > >
