Re: Criticism of "version numbering"-document
| From: | Stefan Neufeind | Date: | Mon, 08 Mar 2004 17:54:55 +0000 |
| Subject: | Re: Criticism of "version numbering"-document | ||
| References: | 1 | Groups: | php.pear.qa |
| Request: | Send a blank email to pear-qa+get-873@lists.php.net to get a copy of this message | ||
On 8 Mar 2004 at 18:37, Lukas Smith wrote:
> Stefan Neufeind wrote:
>
> > Right between beta and stable you mean? That's also fine with me,
> > yeah. Then we should have a hierarchy of
> >
> > cvs -> alpha -> beta -> rc -> stable
>
> s/cvs/devel
Follwing http://pear.php.net/manual/en/developers.packagedef.php:
stable, dev, beta, alpha, devel, or snapshot
Oh, while I read that: The "dev" between stable and beta should be
"rc" instead - shouldn't it? But in that document?
So are we using "dev" (as in the PEAR-group document) or "devel" (as
in the package.xml) for a development release? Or doesn't this have
to be consistent?
And how about snapshots? Do we also need a state "snap", so I can
have a "2.0.0snap2004-03-08"?
Note: I'm *not* trying to make fun of this topic - it's just that
it's currently not that consistent in my view ...
[...]
> > But imho there are only two "good" options:
> >
> > a) Going for a "2.0.1alpha"-name instead of "2.0.0alpha2".
> > b) Implement the missing indicators for the state where possible and
> > avoid having the state in the package version completely. (imho
> > *really* the best and cleanest solution).
>
> Exactly. It is confusing as MDB2 1.0.0 might get confused with MDB
> 1.0.0 and vice versa. At the same time I find it "ugly" if the first
> stable release of a major version is not "x.0.0", because then it
> becomes very untransparent to all outsiders that a given package is no
> finalizing its API. This is why we ended up with the current proposal.
Although I understand your objections why 2.0.0 should be a stable
release, I think this is not a clean solution. For all packages
<1.0.0 we're in the lucky situation of being able to call them "not
stable below 1.0.0". However for 2.x this does not work anymore.
Think about how many releases (and in which states) you usually do
below 1.0:
0.1.0dev, 0.2.0dev, 0.3.0alpha, 0.8.0alpha, 0.8.0beta, 1.0.0RC1 and
then 1.0.0? Or better hav dev1, dev2, alpha1, alpha2 etc.?
This would actually mean that we have, according to the imaginary
example above, 7 releases with all 2.0.0 and major feature additions
inside each dev/alpha-step. Although this prevents us from having any
2.0.0 which is not "stable" I don't think this is a good idea. But
let's hear what others think ...
> Who knows that 2.1.6 stable is the first stable release of major
> version 2? Not unless you take a look at all previous releases and/or
> the changelog. Just as much nobody knows that 2.1.5 RC is the release
> candidate just before that first stable release.
You can ensure that 2.1.6 is a stable release simply by looking at
the state. And if we indicate the state (e.g. in form "2.1.6
(stable)") everywhere this is surely no problem. What's so special
about a "first stable" release? Normally you're looking at least for
a stable or maybe for the "latest stable" imho.
Regards,
Stefan