Distributions going smart & bugfixes?

[Jeff Johnson]

On Jan 10, 2009, at 5:32 AM, Axel Thimm wrote:

> On Fri, Jan 09, 2009 at 01:15:11PM -0800, Michael Jennings wrote:
>> On Friday, 09 January 2009, at 08:29:26 (+0200),
>>> Probably depends on what rpm branch you're talking about - since
>>> rpm.org is used in Fedora 10 and the latter is using more and more
>>> such dependency loops/conglomerations it would be quite funny if the
>>> Fedora rpm developers wouldn't notice killing their coworkers work.
>>
>> Well, there is only 1 "Fedora rpm developer," and I'm pretty sure you
>> were referring to the "Red Hat rpm.org developers."
>
> I'm not sure what you mean, but aren't these days "Red Hat developers"
> automatically "Fedora developers"? Or rephrased: Isn't Fedora rpm
> development in rpm.org just going to end up in RHEL6/7 etc? There are
> Fedora developers that are not Red Hat (or RHEL) developers, but only
> very few Red Hat developers not working on Fedora.
>
>> Regardless of the errors Fedora packagers make, there is really no
>> such thing as a dependency loop that is not inflicted by a packager.
>
> Even max-rpm discusses dependecy loops as a valid instrument, although
> it recommends against them. And it mentiones the PreReq vs Requires
> mechanism that was used to reorder packages within a loop.
>
> Furthermore the packages looping are packaged by the eldest of the Red
> Hat folks, so w/o seeking into archives I can just assume that the
> loops or conglomerations are known and accepted not to be
> errors.
>
> Therefore you can't consider this a repo problem and ask the packager
> to "fix" it.
>
>> (Otherwise, bootstrapping would be impossible.)
>
> There are "loops" in packages needed to bootstrap a minimal Fedora >=
> 10, so there doesn't seem to be an issue with bootstraping otherwise
> there wouldn't exist a Fedora 10.
>
> Consider a "loop" of the above kind like packages that are only
> accepted in pairs or tripples (or more) to be installed. The
> requirement for such loops is that the order they are installed is not
> important, e..g the dependencies are runtime, not installtime.
>
> One can start discussing whether this is a good thing or not (up to
> now this was only a discussion whether this exists or not), but this
> is probably the wrong list to do that, the various rpm lists are
> better for that. From smart's perspective it just needs to know that
> this scenario exists and that it needs to cope with it (like it
> already does with the bugfixes).

You know, arguing about dependency loops, and counting RPM developers,
and quoting from the Greek Bible of "Maximum RPM" , on a smart list,
in 2009, is even more of a waste of time than adding %{?dist}  
everywhere.

FYI: Remember your input onto EVR comparisons? There's a fair to  
middling
chance that EVR can be "unified" between dpkg <-> RPM and included in  
LSB 4.0.

hth

73 de Jeff

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 4664 bytes
Desc: not available
URL: <http://lists.labix.org/pipermail/smart-labix.org/attachments/20090110/22f6bfde/attachment-0003.bin>