[gentoo-dev] irregular metdata.xml check

[gentoo-dev] irregular metdata.xml check

[Thilo Bangert]

[-- Attachment #1: Type: text/plain, Size: 1688 bytes --]

Hi all,

welcome to the IMC - irregular metdata.xml check, issue 1.
As can be seen from the statistcis below, the herd tags have been fixed. 
Thanks to those who have helped with the cleanup.

Whats next?
==========
The current policy in the devmanual demands a maintainer tag - 
nevertheless half the tree is without it.
The policy should perhaps be changed to reflect the reality.

Also, many ebuilds put the herds email address as an additional 
<maintainer>. This is simply redundant and unless complaints are raised, 
all herd <maintainer> tags will be removed and replaced by the 
appropriate <herd> tag instead. Work on this will start over the weekend.

Statistics
=================================

Total number of packages:                    11701

metadata.xml missing                             0
<herd> missing                                   0
<herd> empty                                     0
<herd> unknown                                   0
<herd>=no-herd                                1880

<maintainer> missing                          6616
<maintainer> retired                             0
<maintainer> is a herd                        1306
<maintainer> unknown                           193
<maintainer>=maintainer-needed                 438

Proxy maintainer without gentoo association        11
Unmaintained packages                          596

The full metadata-check.log is available from:
http://dev.gentoo.org/~bangert/metadata-check.log

The script used to generate the log can be found here:
http://overlays.gentoo.org/dev/bangert/browser/scripts/check-metadata.rb

Feedback welcome.
Kind regards
Thilo

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] irregular metdata.xml check

[Chris Gianelloni]

[-- Attachment #1: Type: text/plain, Size: 1016 bytes --]

On Wed, 2007-05-23 at 19:49 +0200, Thilo Bangert wrote:
> Also, many ebuilds put the herds email address as an additional 
> <maintainer>. This is simply redundant and unless complaints are raised, 
> all herd <maintainer> tags will be removed and replaced by the 
> appropriate <herd> tag instead. Work on this will start over the weekend.

Remember that some do this legitimately.  If the herd is, for example,
"portage" but the email/bugzilla address is dev-portage@gentoo.org then
I think it is a good idea to leave the maintainer tag alone.  Remember
that this is an example and isn't what is actually in the tree.  I just
used it as a simple example.  The point is that in cases where the herd
name doesn't equal the email address used for that herd, the maintainer
tag should probably be filled in with the proper email address.

-- 
Chris Gianelloni
Release Engineering Strategic Lead
Alpha/AMD64/x86 Architecture Teams
Games Developer/Council Member/Foundation Trustee
Gentoo Foundation

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

[gentoo-dev] Re: irregular metdata.xml check

[Markus Ullmann]

Chris Gianelloni schrieb:
> The point is that in cases where the herd
> name doesn't equal the email address used for that herd, the maintainer
> tag should probably be filled in with the proper email address.

I think we have herds.xml for a good reason ;)

-Jokey

-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] irregular metdata.xml check

[Timothy Redaelli]

Alle mercoledì 23 maggio 2007, Chris Gianelloni ha scritto:
> On Wed, 2007-05-23 at 19:49 +0200, Thilo Bangert wrote:
> > Also, many ebuilds put the herds email address as an additional
> > <maintainer>. This is simply redundant and unless complaints are raised,
> > all herd <maintainer> tags will be removed and replaced by the
> > appropriate <herd> tag instead. Work on this will start over the weekend.
>
> Remember that some do this legitimately.  If the herd is, for example,
> "portage" but the email/bugzilla address is dev-portage@gentoo.org then
> I think it is a good idea to leave the maintainer tag alone.  Remember
> that this is an example and isn't what is actually in the tree.  I just
> used it as a simple example.  The point is that in cases where the herd
> name doesn't equal the email address used for that herd, the maintainer
> tag should probably be filled in with the proper email address.

http://thread.gmane.org/gmane.linux.gentoo.devel/48485

-- 
Timothy `Drizzt` Redaelli - http://dev.gentoo.org/~drizzt/
FreeSBIE Developer, Gentoo Developer, GUFI Staff
There are two major products that come out of Berkeley: LSD and UNIX.
We don't believe this to be a coincidence.      -- Jeremy S. Anderson
--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] irregular metdata.xml check

[Robin H. Johnson]

[-- Attachment #1: Type: text/plain, Size: 754 bytes --]

On Wed, May 23, 2007 at 07:49:43PM +0200, Thilo Bangert wrote:
> Also, many ebuilds put the herds email address as an additional 
> <maintainer>. This is simply redundant and unless complaints are raised, 
> all herd <maintainer> tags will be removed and replaced by the 
> appropriate <herd> tag instead. Work on this will start over the weekend.
No.

See the thread about automatic assignment for more about this.
More importantly, once the automatic stuff goes into play, the existence
of the herd tag will only matter on metadata that does not have any
other maintainer.


-- 
Robin Hugh Johnson
Gentoo Linux Developer & Council Member
E-Mail     : robbat2@gentoo.org
GnuPG FP   : 11AC BA4F 4778 E3F6 E4ED  F38E B27B 944E 3488 4E85

[-- Attachment #2: Type: application/pgp-signature, Size: 321 bytes --]

Re: [gentoo-dev] irregular metdata.xml check

[Thilo Bangert]

[-- Attachment #1: Type: text/plain, Size: 1709 bytes --]

"Robin H. Johnson" <robbat2@gentoo.org> said:
> On Wed, May 23, 2007 at 07:49:43PM +0200, Thilo Bangert wrote:
> > Also, many ebuilds put the herds email address as an additional
> > <maintainer>. This is simply redundant and unless complaints are
> > raised, all herd <maintainer> tags will be removed and replaced by
> > the appropriate <herd> tag instead. Work on this will start over the
> > weekend.
>
> No.
>
> See the thread about automatic assignment for more about this.
> More importantly, once the automatic stuff goes into play, the
> existence of the herd tag will only matter on metadata that does not
> have any other maintainer.

sorry - to have missed this earlier.
from your proposal:
>Case 2 - Metadata contains a single maintainer
>----------------------------------------------
> The herd field is not used.

so, you want to ignore the herd tag, as soon as there is a single 
maintainer tag? why?

we have <herd> on every single package in the tree (well ~1900 packages 
with <herd>no-herd</herd>). my guess is that most of the roughly 4500 
packages that currently have a <herd> and a <maintainer> which is not a 
<herd>, will need to adjust their metadata to reflect the situation where 
the maintainer should get the bug asssigned and the herd gets CC'd...

IMHO the herd should always get an email on bugs with packages belonging 
to the herd... if this is not the case, what is the purpose of the herd?

or asked differently: what can the herd in <maintainer> give you that the 
<herd> can't?

other than that i (still) agree with the overall proposal. lets just make 
sure to codify the policy which has been agreed upon...

regards
Thilo


[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] irregular metdata.xml check

[Radoslaw Stachowiak]

On 27/05/07, Thilo Bangert <bangert@gentoo.org> wrote:
[...and many others...]

What I deeply miss from gentoo-dev mailing list is lack of  summary at
end of discussions. Current situation is that at the end of such
disputes very often (of course not always :) is still unclear if:
* decision has been made (and what it is)
* previous approach has been changed (or it hasn't, with pointer (URL) to it)
* documents were updated (to address change or to simply provide
better wording) w/ URLs
* nothing has been changed due to X, and next approach will be in X months.

It's even more difficult to grasp outcome for non native english speakers.

Solution I propose is, that author of the first post in a thread, who
started the discussion (if there was some discusion), should at end of
it (let's assume it's 3 days after last mail) write summary with those
4 points I outlined above. Simple and will help greatly to track later
decisions/changes, or postponed items.

-- 
radoslaw.
-- 
gentoo-dev@gentoo.org mailing list

[gentoo-dev] Re: irregular metdata.xml check

[Ryan Hill]

Thilo Bangert wrote:
> "Robin H. Johnson" <robbat2@gentoo.org> said:

>> See the thread about automatic assignment for more about this.
>> More importantly, once the automatic stuff goes into play, the
>> existence of the herd tag will only matter on metadata that does not
>> have any other maintainer.
> 
> sorry - to have missed this earlier.
> from your proposal:
>> Case 2 - Metadata contains a single maintainer
>> ----------------------------------------------
>> The herd field is not used.
> 
> so, you want to ignore the herd tag, as soon as there is a single 
> maintainer tag? why?
> 
> we have <herd> on every single package in the tree (well ~1900 packages 
> with <herd>no-herd</herd>). my guess is that most of the roughly 4500 
> packages that currently have a <herd> and a <maintainer> which is not a 
> <herd>, will need to adjust their metadata to reflect the situation where 
> the maintainer should get the bug asssigned and the herd gets CC'd...
> 
> IMHO the herd should always get an email on bugs with packages belonging 
> to the herd... if this is not the case, what is the purpose of the herd?
> 
> or asked differently: what can the herd in <maintainer> give you that the 
> <herd> can't?

I agree here, we have many packages that are assigned to a herd, with a
certain member of that herd listed as maintainer.  For example I'll use
freetype since I've been working on it today:

dirtyepic@tycho /usr/portage/media-libs/freetype $ cat metadata.xml
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE pkgmetadata SYSTEM "http://www.gentoo.org/dtd/metadata.dtd">
<pkgmetadata>
<herd>fonts</herd>
<maintainer>
  <email>foser@gentoo.org</email>
</maintainer>
</pkgmetadata>

In this case, a bug should be assigned to foser and fonts@gentoo.org
should be CCed.  Freetype has frequent security bugs and foser is
usually not present to deal with them.  If the herd wasn't CC'd there'd
be trouble.

Another case is where we have a maintainer who isn't a member of the
herd, like dejavu for example.

dirtyepic@tycho /usr/portage/media-fonts/dejavu $ cat metadata.xml
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE pkgmetadata SYSTEM "http://www.gentoo.org/dtd/metadata.dtd">
<pkgmetadata>
<herd>fonts</herd>
<maintainer>
 <email>pva@gentoo.org</email>
 <name>Peter Volkov</name>
</maintainer>
<longdescription>
The DejaVu fonts are a font family based on the Vera Fonts release 1.10. Its
purpose is to provide a wider range of characters while maintaining the
original
look and feel through the process of collaborative development. Available
families (Sans = sans serif, Mono = monospaced): DejaVu Sans, DejaVu
Sans Mono,
DejaVu Sans Condensed (experimental), DejaVu Serif, DejaVu Serif Condensed
</longdescription>
</pkgmetadata>

We would want bugs filed against dejavu to be CC'd to the fonts herd, in
case pva isn't present or it's something more generally font-related.

No matter what, if the herd tag is not "no-herd", it should be CC'd on
all bugs.


-- 
dirtyepic                 salesman said this vacuum's guaranteed
 gentoo org          it could suck an ancient virus from the sea
  9B81 6C9F E791 83BB 3AB3  5B2D E625 A073 8379 37E8 (0x837937E8)

-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: irregular metdata.xml check

[Robin H. Johnson]

[-- Attachment #1: Type: text/plain, Size: 672 bytes --]

On Sun, Jun 10, 2007 at 02:58:27PM -0600, Ryan Hill wrote:
> I agree here, we have many packages that are assigned to a herd, with a
> certain member of that herd listed as maintainer.  For example I'll use
> freetype since I've been working on it today:
....
> No matter what, if the herd tag is not "no-herd", it should be CC'd on
> all bugs.
Please see my latest email about automatic processing.

I did agree with bangert's points, but I just hadn't gotten to revising
my original proposal.

-- 
Robin Hugh Johnson
Gentoo Linux Developer & Council Member
E-Mail     : robbat2@gentoo.org
GnuPG FP   : 11AC BA4F 4778 E3F6 E4ED  F38E B27B 944E 3488 4E85

[-- Attachment #2: Type: application/pgp-signature, Size: 321 bytes --]