Mailing List Archive

On Wed, Nov 22, 2006 at 02:08:12PM -0700, Steve Dibb wrote:
> Hi guys,
>
> There are more than a few packages with missing metadata.xml in the
> portage tree. I've setup my funky little QA website to report on which
> ones fall in that category, and here is the list right here:
>
> http://spaceparanoids.org/gentoo/gpnl/qa.php?q=metadata
>
> I've spent the morning fixing up most of them, adding blank metadata.xml
> to them and assigning maintainer-needed@gentoo.org as the main
> maintainer, which in hindsight was probably not the best approach (my
> apologies).
>
> Anyway, either way, it would be nice to get the few remaining packages
> cleaned up, and if one of your packages is on that list, please update
> or create the metadata.
>
> I'll still be going through the rest of them and sorting out which ones
> were last maintained by a dev that is now retired and continue assigning
> them to maintainer-needed.
>
I think the most important thing about adding "empty" metadata.xml files
with maintainer-needed as maintainer is that it _changes_ the package to
be unmaintained by definition (that's what maintainer-needed means after
all) and that we can't be sure that's actually true unless we spend a
lot of time examining each package and asking potential maintainers
if it's unmaintained.

So while I enjoy getting metadata cleaned up etc. I think it's important
to think about exactly what we're doing before "fixing" up a lot of
packages - in this case 300+ packages. You (all devs!) might even want
to ask on -dev ML if it's a good idea before touching up a huge number
of packages to make sure you don't change things in subtle,
unintentional ways.

Anyway, I appreciate you spending time on cleaning up the metadata.xml
files even if it might not have been the best idea in hindsight.

Regards,
Bryan Østergaard
--
gentoo-dev@gentoo.org mailing list

Tach Steve, 0x2B859DE3 (PGP-PK-ID)

Steve Dibb schrieb:
> There are more than a few packages with missing metadata.xml in the
> portage tree. I've setup my funky little QA website to report on which
> ones fall in that category, and here is the list right here:

Nice idea, but you should really add <herd>no-herd</herd> as this is
required if there is only a "maintainer".

By the way, I tagged some metadata files with appropriate herd (where I am
allowed):

app-emacs/cperl-mode emacs
dev-lisp/cl-asdf-binary-locations common-lisp


I suggest the following changes:

app-misc/baobab vapier and/or Gnome (as this is part of Gnome
2.16)
dev-scheme/guile-pg scheme
dev-scheme/kawa scheme
dev-scheme/mzscheme scheme
dev-util/cweb text-markup and/or lang-misc

V-Li

--
Fingerprint: 68C5 D381 B69A A777 6A91 E999 350A AD7C 2B85 9DE3
http://www.gnupg.org/
--
gentoo-dev@gentoo.org mailing list

Bryan Østergaard napsal(a):
> I think the most important thing about adding "empty" metadata.xml files
> with maintainer-needed as maintainer is that it _changes_ the package to
> be unmaintained by definition (that's what maintainer-needed means after
> all) and that we can't be sure that's actually true unless we spend a
> lot of time examining each package and asking potential maintainers
> if it's unmaintained.

Actually, I don't mind much. There's a developers or two who keep on
adding packages without metadata.xml all the time (won't name anyone,
I'm pretty sure they'll find themselves here :P). This will either force
them to reclaim their packages via fixing the metadata.xml thing or will
leave the ebuilds orphaned to m-needed - and then they shouldn't have
been added in the first place.

Above, I'm not talking about legacy stuff maintained in an ad-hoc manner
for ages, but about fairly recent additions to the tree (~1 year or even
less). However, even for legacy stuff, nothing is preventing the people
from claiming their ebuilds the right way and adding themselves to
metadata.xml - will make assigning bugs much easier for me. ;)


--
Best regards,

Jakub Moc
mailto:jakub@gentoo.org
GPG signature:
http://subkeys.pgp.net:11371/pks/lookup?op=get&search=0xCEBA3D9E
Primary key fingerprint: D2D7 933C 9BA1 C95B 2C95 B30F 8717 D5FD CEBA 3D9E

... still no signature ;)

Bryan Østergaard wrote:
> On Wed, Nov 22, 2006 at 02:08:12PM -0700, Steve Dibb wrote:
>> Hi guys,
>>
>> There are more than a few packages with missing metadata.xml in the
>> portage tree. I've setup my funky little QA website to report on which
>> ones fall in that category, and here is the list right here:
>>
>> http://spaceparanoids.org/gentoo/gpnl/qa.php?q=metadata
>>
>> I've spent the morning fixing up most of them, adding blank metadata.xml
>> to them and assigning maintainer-needed@gentoo.org as the main
>> maintainer, which in hindsight was probably not the best approach (my
>> apologies).
>>
>> Anyway, either way, it would be nice to get the few remaining packages
>> cleaned up, and if one of your packages is on that list, please update
>> or create the metadata.
>>
>> I'll still be going through the rest of them and sorting out which ones
>> were last maintained by a dev that is now retired and continue assigning
>> them to maintainer-needed.
>>
> I think the most important thing about adding "empty" metadata.xml files
> with maintainer-needed as maintainer is that it _changes_ the package to
> be unmaintained by definition (that's what maintainer-needed means after
> all) and that we can't be sure that's actually true unless we spend a
> lot of time examining each package and asking potential maintainers
> if it's unmaintained.

I see what you mean here, but asking potential maintainers doesn't seem
like too much of a solution, as it would take a lot of time and energy.
In my opinion, if the package is actually maintained, then it shouldn't
be hard for the maintainer to fix the metadata, adding himself as the
maintainer or at least assigning it to a herd.

--
David Shakaryan
GnuPG Public Key: 0x4B8FE14B
--
gentoo-dev@gentoo.org mailing list

On Thu, Nov 23, 2006 at 02:56:49AM -0800, David Shakaryan wrote:
> Bryan Østergaard wrote:
> >On Wed, Nov 22, 2006 at 02:08:12PM -0700, Steve Dibb wrote:
> >>Hi guys,
> >>
> >>There are more than a few packages with missing metadata.xml in the
> >>portage tree. I've setup my funky little QA website to report on which
> >>ones fall in that category, and here is the list right here:
> >>
> >>http://spaceparanoids.org/gentoo/gpnl/qa.php?q=metadata
> >>
> >>I've spent the morning fixing up most of them, adding blank metadata.xml
> >>to them and assigning maintainer-needed@gentoo.org as the main
> >>maintainer, which in hindsight was probably not the best approach (my
> >>apologies).
> >>
> >>Anyway, either way, it would be nice to get the few remaining packages
> >>cleaned up, and if one of your packages is on that list, please update
> >>or create the metadata.
> >>
> >>I'll still be going through the rest of them and sorting out which ones
> >>were last maintained by a dev that is now retired and continue assigning
> >>them to maintainer-needed.
> >>
> >I think the most important thing about adding "empty" metadata.xml files
> >with maintainer-needed as maintainer is that it _changes_ the package to
> >be unmaintained by definition (that's what maintainer-needed means after
> >all) and that we can't be sure that's actually true unless we spend a
> >lot of time examining each package and asking potential maintainers
> >if it's unmaintained.
>
> I see what you mean here, but asking potential maintainers doesn't seem
> like too much of a solution, as it would take a lot of time and energy.
> In my opinion, if the package is actually maintained, then it shouldn't
> be hard for the maintainer to fix the metadata, adding himself as the
> maintainer or at least assigning it to a herd.
>
I completely agree that adding metadata.xml files is easy for the
maintainers and should be done. What I'm objecting to is randomly adding
metadata.xml files to packages without any idea if the added files are
actually correct. If you can't solve the problem properly you should
probably stop to think about a proper solution instead of just taking
the easy (but quite possible wrong) solution.

Regards,
Bryan Østergaard
--
gentoo-dev@gentoo.org mailing list

On Thu, Nov 23, 2006 at 11:20:16AM +0100, Jakub Moc wrote:
> Bryan Østergaard napsal(a):
> > I think the most important thing about adding "empty" metadata.xml files
> > with maintainer-needed as maintainer is that it _changes_ the package to
> > be unmaintained by definition (that's what maintainer-needed means after
> > all) and that we can't be sure that's actually true unless we spend a
> > lot of time examining each package and asking potential maintainers
> > if it's unmaintained.
>
> Actually, I don't mind much. There's a developers or two who keep on
> adding packages without metadata.xml all the time (won't name anyone,
> I'm pretty sure they'll find themselves here :P). This will either force
> them to reclaim their packages via fixing the metadata.xml thing or will
> leave the ebuilds orphaned to m-needed - and then they shouldn't have
> been added in the first place.
>
> Above, I'm not talking about legacy stuff maintained in an ad-hoc manner
> for ages, but about fairly recent additions to the tree (~1 year or even
> less). However, even for legacy stuff, nothing is preventing the people
> from claiming their ebuilds the right way and adding themselves to
> metadata.xml - will make assigning bugs much easier for me. ;)
>
I not quite as concerned whether your job is "easy" or not as I am that
we don't lie about maintainers in metadata.xml. Wrong metadata.xml files
affects a lot more people (devs as well as users) than just
bug-wranglers.

Regards,
Bryan Østergaard
--
gentoo-dev@gentoo.org mailing list

Dňa Thu, 23 Nov 2006 11:20:16 +0100
Jakub Moc <jakub@gentoo.org> napísal:

> Actually, I don't mind much. There's a developers or two who keep on
> adding packages without metadata.xml all the time (won't name anyone,
> I'm pretty sure they'll find themselves here :P).

Why? If someone does this, they need to be spanked - and for that, we
need to know who they are.

Kind regards,
--
Andrej Kacian <ticho@gentoo.org>

Christian Faulhammer wrote:
> Tach Steve, 0x2B859DE3 (PGP-PK-ID)
>
> Steve Dibb schrieb:
>> There are more than a few packages with missing metadata.xml in the
>> portage tree. I've setup my funky little QA website to report on which
>> ones fall in that category, and here is the list right here:
>
> Nice idea, but you should really add <herd>no-herd</herd> as this is
> required if there is only a "maintainer".

Really? <herd/> isn't valid? I'd rather see that than adding a "fake" herd.

Thanks,
Donnie
--
gentoo-dev@gentoo.org mailing list

Donnie Berkholz wrote:
> Christian Faulhammer wrote:
>> Nice idea, but you should really add <herd>no-herd</herd> as this
>> is required if there is only a "maintainer".
>
> Really? <herd/> isn't valid? I'd rather see that than adding a "fake" herd.

Neither are valid from what I understand. IIRC, kloeri said that every
package should belong to a herd.

--
David Shakaryan
GnuPG Public Key: 0x4B8FE14B

On Thu, 2006-11-23 at 11:14 +0100, Christian Faulhammer wrote:
> I suggest the following changes:
>
> app-misc/baobab vapier and/or Gnome (as this is part of Gnome
> 2.16)

Not gnome. baobab is part of gnome-utils as of 2.16, and gets installed
from gnome-extra/gnome-utils which, I realize now, is begging for a
app-misc/baobab blocker to avoid collisions *adds to his growing TODO list*

Earlier I put app-admin/sabayon under gnome, which seemed to have been
forgotten when moving over to portage from gnome-experimental overlay.

I intend to personally take over maintenance of dev-util/sysprof and
dev-util/memprof later on if no-one else is willing. /me pokes dev-tools
herd for co-maintenance

--
Mart Raudsepp
Gentoo Developer
Mail: leio@gentoo.org
Weblog: http://planet.gentoo.org/developers/leio

Tach Donnie, 0x2B859DE3 (PGP-PK-ID)

Donnie Berkholz schrieb:
> Christian Faulhammer wrote:
>> Steve Dibb schrieb:
>>> There are more than a few packages with missing metadata.xml in the
>>> portage tree. I've setup my funky little QA website to report on
>>> which ones fall in that category, and here is the list right here:
>> Nice idea, but you should really add <herd>no-herd</herd> as this is
>> required if there is only a "maintainer".
> Really? <herd/> isn't valid? I'd rather see that than adding a "fake"
> herd.

I though I read about it being mandatory with "no-herd"...yes Alec Warner
wrote that (implicitly). But I am not really sure. At least, herd tag is
mandatory, in whatever way.

V-Li

--
Fingerprint: 68C5 D381 B69A A777 6A91 E999 350A AD7C 2B85 9DE3
http://www.gnupg.org/
--
gentoo-dev@gentoo.org mailing list

On Thursday 23 November 2006 11:20, Jakub Moc wrote:
> Bryan Østergaard napsal(a):
> > I think the most important thing about adding "empty" metadata.xml files
> > with maintainer-needed as maintainer is that it _changes_ the package to
> > be unmaintained by definition (that's what maintainer-needed means after
> > all) and that we can't be sure that's actually true unless we spend a
> > lot of time examining each package and asking potential maintainers
> > if it's unmaintained.
>
> Actually, I don't mind much. There's a developers or two who keep on
> adding packages without metadata.xml all the time (won't name anyone,
> I'm pretty sure they'll find themselves here :P). This will either force
> them to reclaim their packages via fixing the metadata.xml thing or will
> leave the ebuilds orphaned to m-needed - and then they shouldn't have
> been added in the first place.
>
> Above, I'm not talking about legacy stuff maintained in an ad-hoc manner
> for ages, but about fairly recent additions to the tree (~1 year or even
> less). However, even for legacy stuff, nothing is preventing the people
> from claiming their ebuilds the right way and adding themselves to
> metadata.xml - will make assigning bugs much easier for me. ;)

Repoman should check for missing metadata. The only packages that are allowed
not to have metadata.xml would be those that have not been changed for over 3
years (since the introduction of metadata.xml). Developers who violate the
repoman checks by omitting a metadata.xml brought mayhem over themselves.

Paul

--
Paul de Vrieze
Gentoo Developer
Mail: pauldv@gentoo.org
Homepage: http://www.devrieze.net

Andrej Kacian napsal(a):
> Dòa Thu, 23 Nov 2006 11:20:16 +0100
> Jakub Moc <jakub@gentoo.org> napísal:
>
>> Actually, I don't mind much. There's a developers or two who keep on
>> adding packages without metadata.xml all the time (won't name anyone,
>> I'm pretty sure they'll find themselves here :P).
>
> Why? If someone does this, they need to be spanked

Meh, even you got it right on the first try ^^^^^^^^ - no need to name
anyone *g*


--
Best regards,

Jakub Moc
mailto:jakub@gentoo.org
GPG signature:
http://subkeys.pgp.net:11371/pks/lookup?op=get&search=0xCEBA3D9E
Primary key fingerprint: D2D7 933C 9BA1 C95B 2C95 B30F 8717 D5FD CEBA 3D9E

... still no signature ;)

On Fri, 24 Nov 2006 10:59:32 +0100
Jakub Moc <jakub@gentoo.org> wrote:

> Andrej Kacian napsal(a):
> > Why? If someone does this, they need to be spanked
>
> Meh, even you got it right on the first try ^^^^^^^^ - no need to name
> anyone *g*

Well, in secret everybody knows that the "no-herd" herd has existed for
ages and has two very active members...


Kind regards,
JeR
--
gentoo-dev@gentoo.org mailing list

On 25/11/06, Jeroen Roovers <jer@gentoo.org> wrote:
> On Fri, 24 Nov 2006 10:59:32 +0100
> Jakub Moc <jakub@gentoo.org> wrote:
>
> > Andrej Kacian napsal(a):
> > > Why? If someone does this, they need to be spanked
> >
> > Meh, even you got it right on the first try ^^^^^^^^ - no need to name
> > anyone *g*
>
> Well, in secret everybody knows that the "no-herd" herd has existed for
> ages and has two very active members...

'no' and 'herd'. *nod* --beu
--
gentoo-dev@gentoo.org mailing list

Tach, 0x2B859DE3 (PGP-PK-ID)

> dev-scheme/guile-pg scheme
> dev-scheme/kawa scheme
> dev-scheme/mzscheme scheme

Done.

V-Li

--
Fingerprint: 68C5 D381 B69A A777 6A91 E999 350A AD7C 2B85 9DE3
http://www.gnupg.org/
--
gentoo-dev@gentoo.org mailing list

opfer@gentoo.org (Christian Faulhammer) wrote:
> Donnie Berkholz schrieb:
> > Christian Faulhammer wrote:

> >> Nice idea, but you should really add <herd>no-herd</herd> as this
> >> is required if there is only a "maintainer".
> > Really? <herd/> isn't valid? I'd rather see that than adding a
> > "fake" herd.
>
> I though I read about it being mandatory with "no-herd"...yes Alec
> Warner wrote that (implicitly). But I am not really sure. At least,
> herd tag is mandatory, in whatever way.

http://devmanual.gentoo.org/ebuild-writing/misc-files/metadata/index.html
http://www.gentoo.org/proj/en/metastructure/herds/index.xml