Let's take NH-3417 as example. This is yet another bug related to LINQ and 
OData support. Let's imagine there are issues already fixed in 3.3.3.CR1 
someone eagerly waits for and these issues have no connection with LINQ or 
OData. Now we have two options: either wait a bit more trying to fix 
NH-3417 and hope no more issues arise or, taking into account there are 
other important fixed issues, release 3.3.3 as is.

Let's look around on trends. Even Microsoft spin all the ASP.NET stuff out 
of VS release cycle. They want to release twice a year with pre-official 
releases available even more frequently. "Time to developer" for code is 
the same important metric as "time to market" for product.

I do agree there are concerns about multiple repo's, maintenance burden and 
so on. But imho we should think ahead and address these issues before, 
instead of simply dismiss the possibility of separating monolithic 
NHibernate into "core + components" with separate release cycle an so on.

P.S. As open source example we can look at Castle, which split it's core 
from other components.

понедельник, 4 марта 2013 г., 11:57:40 UTC+4 пользователь Oskar Berggren 
написал:
>
> 2013/3/4 Alexander I. Zaytsev <[email protected] <javascript:>>: 
> > Hi Patrick 
> > 
> > 1. I think that ability to make own small package instead of building 
> own 
> > NHibernate is an advantage. As for now many dialects/providers are 
> outdate 
> > or have minor or even trivial errors, and people often wait years to 
> have 
> > these issues to be fixed. Also there are number of independent providers 
>
> From time to time there are small but importand fixes everywhere. As 
> an alternative to this, I'm thinking we should try to increase the 
> pace of minor releases. With the stable assembly version we know have 
> I think this can be done, if only we decide that's what we want (the 
> 3.3.3 cycle is way longer than I would have preferred). If we agree 
> that we want shorter cycles, to make it happen we should probably aim 
> for a release every two months. With such a time frame we should be 
> able to maintain each release to a manageable set of changes 
> (acceptable risk). 
>
> As someone mentioned, I too fear the risk of increased maintenance 
> burden of having to handle multiple repo's. The situation with the 
> existing contrib modules isn't encouraging. 
>
> /Oskar 
>

-- 

--- 
You received this message because you are subscribed to the Google Groups 
"nhibernate-development" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.


Reply via email to