作者: Keiichi Takahashi
日時: 2002/8/11(15:16)
ビットウォークの高橋です。

最近のTCLCOREのMLで次のような提案が出ました。いろいろ議論gされるだろうと
思ったのですが、今のところ大きな議論はありません。
そのため、まだ最終方針が未確定の情報ですが、内容を引用の形で以下に紹介し
ます。ご参考まで。

----- ここから -----
From: "Jeff Hobbs" <JeffH@...>
To: <tcl-core@...>
Subject: [TCLCORE] 8.4b2 and beyond (even/odd stable/dev versioning)
Date: Thu, 8 Aug 2002 00:10:51 -0700
Sender: tcl-core-admin@...
X-Mailer: Microsoft Outlook CWS, Build 9.0.2416 (9.0.2911.0)

8.4b2 is firming up and should be final once I'm done tinkering tonight.
A day or two on ice and it should be ready for the world.  The number of
high prio bugs is low, but there are others that should be fixed before
we can stamp 8.4.0 as ready for prime time.

I'm considering making an 8.3.5 release because 8.3.4 is 10 months old
and we have made plenty of patches on the head.  8.3.5 would be an EOL
for that branch as far as main dev is concerned though, to focus on 8.4
and beyond.

Following 8.4, and in discussion with others, I believe we should move
to the even/odd stable/dev version system that Linux made popular and
is used by many other projects.  This has the advantage of never
including "a" or "b" in a release version (which, BTW, never did work
in the package mechanism anyway), which real world customers had an
aversion to, even if the code was solid (for us, "a" == feature creep,
"b" == testing/release candidates).

That means that the next version of Tcl would be 8.5.0, and new
features would be added in throughout 8.5.x.  Once we are happy that
it has reached a "stable" state, 8.6.0 would be born.

The only disadvantage is the minor confusion for the 8.3 release,
which is a very solid line (8.1 on the other hand ... well, that fits
into the even/odd line :/ ).  However, it should be easy for users to
transition to 8.4.

Jeff
----- ここまで -----
-- 
Keiichi Takahashi, bitWalk Co.,Ltd.
mailto:bitwalk@...
http://members10.tsukaeru.net/bitwalk/