[Ecls-list] ECL 0.9l release candidate
Vadim Konovalov
vadrer at gmail.com
Sun Aug 3 22:06:08 UTC 2008
在 Sunday 03 August 2008 21:09:20,Juan Jose Garcia-Ripoll 写道:
> On Sun, Aug 3, 2008 at 10:42 PM, Vadim Konovalov <vadrer at gmail.com> wrote:
> > I'm very glad to see ECL improving,
> > but is it possible to move to normal versioning system? (of course I am
> > speaking about future releases, not this one)
> > letter versions are - so 1995... (*)
>
> What do you mean by normal? In my life I have seen almost anything,
> from those that use names instead of numbers, dates instead of
> revision numbers, arbitrary jumps in revision numbers like those by
> Acrobat and others...
>
> Current scheme stays below 1.0 because I do not want to make a 1.0
> until the compiler is fully stable. Then again, all releases are
> binary incompatible among each other, which precludes having three
> numbers systems as other linux libraries.
>
> So what would be your suggestion?
I meant having versions like 0.9.2.15, just like SBCL do, and many other SW
nowadays, because that's easier for users to eye-catch.
Perl also had such system, but it then came to numbers-only versions.
Here is an excerpt from "perldoc perlhist" to illustrate my point:
5.001k 1995-Jun-06
5.001l 1995-Jun-06 Stable.
5.001m 1995-Jul-02 Very stable.
5.001n 1995-Oct-31 Very unstable.
5.002beta1 1995-Nov-21
5.002b1a 1995-Dec-04
5.002b1b 1995-Dec-04
5.002b1c 1995-Dec-04
5.002b1d 1995-Dec-04
5.002b1e 1995-Dec-08
5.002b1f 1995-Dec-08
Tom 5.002b1g 1995-Dec-21 Doc release.
Andy 5.002b1h 1996-Jan-05
5.002b2 1996-Jan-14
Larry 5.002b3 1996-Feb-02
Andy 5.002gamma 1996-Feb-11
Larry 5.002delta 1996-Feb-27
Larry 5.002 1996-Feb-29 Prototypes.
Charles 5.002_01 1996-Mar-25
5.003 1996-Jun-25 Security release.
5.003_01 1996-Jul-31
Nick 5.003_02 1996-Aug-10
Andy 5.003_03 1996-Aug-28
5.003_04 1996-Sep-02
But this is of course up to you to decide...
BR,
Vadim.
More information about the ecl-devel
mailing list