xorg-x11- packaging prefix
Mike A. Harris
mharris at mharris.ca
Tue May 9 02:07:41 UTC 2006
Nicolas Mailhot wrote:
> Le mercredi 03 mai 2006 à 14:10 +0200, Axel Thimm a écrit :
>
>> Which brings us back to the question, does the prefix really imply "by
>> <prefix>" or "for <prefix>". Usually in packaging practice
>> "<prefix>-foo" means foo built for <prefix>, e.g. the miriads of
>> perl-XXX packages, now python-XXX, too, java-XXX,
>
> There are no miriads of java-XXX precisely because "java" is not a code
> provider.
>
> You *do* have miriads of jakarta-XXX packages because jakarta represents
> the org which produces the code in question (likewise you have
> classpathx-XXX packages)
>
> I believe the perl-XXXX are close to being a cpan-XXXX synonym.
>
> IMHO the prefix means "code closely associated to project foo", either
> because foo is the upstream project or it's a well-known extension of
> foo which can easily be found from foo website. But in the end it's a
> packager preference.
Which would be an incorrect interpretation of what I chose the
prefix for. It is true that I (or anyone) can not prevent someone
from abusing this namespace in any enforceable manner, however I
hope people have the decency to understand and respect the choice
and not step all over the namespace due to differing opinions.
We do not want to see a tonne of bug reports hitting X.Org nor
Red Hat bugzilla from people using packages named "xorg-x11-anything"
which is software not actually produced or provided by X.Org.
Labelling things that did not come from X.Org with the xorg-x11
prefix, will suggest to people that said software did come from
X.Org, as all X packaging historically had the upstream X implementation
provider's name as the prefix. ie: XFree86-* xorg-x11-*
Please keep the xorg-x11 namespace clean and reserved for software
produced and distributed by the X.Org Foundation officially, to
which the trademark is owned.
--
Mike A. Harris * Open Source Advocate * http://mharris.ca
Proud Canadian.
More information about the fedora-devel-list
mailing list