From: Prashant Srinivasan <Prashant.Srinivasan@...>
Date: 2007-10-03T15:28:26+09:00
Subject: Re: Release compatibility/train



David Flanagan wrote:
> Yukihiro Matsumoto wrote:
>
>> In message "Re: Release compatibility/train"
>>     on Wed, 3 Oct 2007 12:33:33 +0900, Prashant Srinivasan 
>> <Prashant.Srinivasan@Sun.COM> writes:
>>
>>
>> |2-> Ruby 1.9(and other odd minor versions) will available only from 
>> |CVS[2], and cannot, in general, be expected to be compatible across 
>> |point releases.
>> |
>> |That is:
>> |<Version>.<OddNumber-minor-version>.<TeenyVersion> and 
>> |<Version>.<OddNumber-minor-version>.<TeenyVersion + x> where x > 0.
>> |
>> |are not expected to be compatible.
>>
>> Now it's available from Subversion.  It's not supposed to be
>> compatible with 1.8, and once it's released (planned to be December),
>> it should follow rule-1 above.
>
> Also, if I understand it correctly, Ruby 1.9 will remain an odd 
> version number when released in December. It will not suddenly turn 
> into 2.0 This is a change from the way previous releases were done.  
> During development, there are no point releases, just snapshots of 
> 1.9.0.  When it is released, it will be 1.9.1.  So the statement of 
> the rule above is wrong.  It is not the odd or evenness of the second 
> digit that indicates a non-stable development branch; it is the 
> absence of a third digit.
>

Matz and David,

 Can this be elaborated on please . . .

Are you saying that 1.9 (aka 1.9.0) will be a development version, and 
that 1.9.1 will be a stable release?

(or to put it another way, that 1.9.0 will be a development version, and 
1.9.x where x > 0 will be stable releases?).

Thanks,
 -ps


> I hope Matz will correct me if I'm wrong about this.
>
>     David
>
>

-- 
Prashant Srinivasan
F/OSS Enthusiast
Sun Microsystems, Inc.
http://blogs.sun.com/prashant
GnuPG key: http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x82FBDE5A