Florian Boor | 2 Apr 11:03 2007
Picon

Re: [oe] [oe-commits] org.oe.dev prismstumbler: add 0.7.4pre1

Hi,

Koen Kooi schrieb:
> That is against our versioning policy, this way you can't upgrade to 0.7.4 when it gets
> released[1]. Use _0.7.3+0.7.4pre1 instead.

I remember we had these kind of trouble with ipkg before. Wouldn't it be better
to make it possible to handle this properly? I guess we will run into this again
and again in future.

Greetings

Florian

--

-- 
The dream of yesterday                  Florian Boor
is the hope of today                    Tel: +49 271-771091-15
and the reality of tomorrow.            Fax: +49 271-771091-19
[Robert Hutchings Goddard, 1904]        florian.boor <at> kernelconcepts.de

1D78 2D4D 6C53 1CA4 5588  D07B A8E7 940C 25B7 9A76
Paul Sokolovsky | 3 Apr 12:00 2007
Picon

Re: [oe] [oe-commits] org.oe.dev prismstumbler: add 0.7.4pre1

Hello Florian,

Monday, April 2, 2007, 12:03:34 PM, you wrote:

> Hi,

> Koen Kooi schrieb:
>> That is against our versioning policy, this way you can't upgrade to 0.7.4 when it gets
>> released[1]. Use _0.7.3+0.7.4pre1 instead.

> I remember we had these kind of trouble with ipkg before.

  Why we and why ipkg? Any packaging system which tries to do its work
well hits into such issues. In this regard, OE follows Debian's way to
handle this. (Even though I was unable to find exact best
practices to resolve such cases, they only mumble that it should be
done: http://www.debian.org/doc/debian-policy/ch-binary.html#s-versions ;
anyway, + notation is what actually being used).

  I'm not sure how other packaging systems solve this, for example,
look at FC6 RPM pile shows that they have *pre* stuff without clearly
distinguishable pattern of handling it syntactically. But it's known
where that leads to - to inability to upgrade to a new distro release.

> Wouldn't it be better
> to make it possible to handle this properly? I guess we will run into this again
> and again in future.

> Greetings

(Continue reading)

Koen Kooi | 3 Apr 12:14 2007
Picon
Picon

Re: [oe] [oe-commits] org.oe.dev prismstumbler: add 0.7.4pre1


Paul Sokolovsky schreef:

>>> That is against our versioning policy, this way you can't upgrade to 0.7.4 when it gets
>>> released[1]. Use _0.7.3+0.7.4pre1 instead.
> 
>> I remember we had these kind of trouble with ipkg before.
> 
>   Why we and why ipkg? Any packaging system which tries to do its work
> well hits into such issues. In this regard, OE follows Debian's way to
> handle this. (Even though I was unable to find exact best
> practices to resolve such cases, they only mumble that it should be
> done: http://www.debian.org/doc/debian-policy/ch-binary.html#s-versions ;
> anyway, + notation is what actually being used).
> 
>   I'm not sure how other packaging systems solve this, for example,
> look at FC6 RPM pile shows that they have *pre* stuff without clearly
> distinguishable pattern of handling it syntactically. But it's known
> where that leads to - to inability to upgrade to a new distro release.
> 
>> Wouldn't it be better
>> to make it possible to handle this properly? I guess we will run into this again
>> and again in future.

It would be best if upstream (prismstumbler in this case) would use proper versioning to
start with, but if that is not possible we can paper over it in OE.

regards,

Koen
(Continue reading)

Florian Boor | 3 Apr 15:02 2007
Picon

Re: [oe] [oe-commits] org.oe.dev prismstumbler: add 0.7.4pre1

Hi,

Koen Kooi schrieb:
> It would be best if upstream (prismstumbler in this case) would use proper versioning to
> start with, but if that is not possible we can paper over it in OE.

I do not agree... sure it might be better to have simpler versions but that's
not the point. I can change this in prismstumbler because I maintain it, but
this world is not perfect we will run into the same situation again with another
package. These "pre" versions are quite common and the documentation explains
how to handle these already so we should be able to teach our software to do the
same. That's way better then teaching users and even if it does not work in all
cases it is quite likely to be more reliable than having a guideline for it.

Greetings

Florian

--

-- 
The dream of yesterday                  Florian Boor
is the hope of today                    Tel: +49 271-771091-15
and the reality of tomorrow.            Fax: +49 271-771091-19
[Robert Hutchings Goddard, 1904]        florian.boor <at> kernelconcepts.de

1D78 2D4D 6C53 1CA4 5588  D07B A8E7 940C 25B7 9A76
Koen Kooi | 3 Apr 15:09 2007
Picon
Picon

Re: [oe] [oe-commits] org.oe.dev prismstumbler: add 0.7.4pre1


Florian Boor schreef:
> Hi,
> 
> Koen Kooi schrieb:
>> It would be best if upstream (prismstumbler in this case) would use proper versioning to
>> start with, but if that is not possible we can paper over it in OE.
> 
> I do not agree...

You don't agree with OE's versiong policy!?!?!

Gmane