Re: [gentoo-dev] Re: [2 GLEPs] metadata improvements

[Robin H. Johnson]

On Fri, Jul 04, 2008 at 09:58:21PM +0200, Tiziano M?ller wrote:
> > On Fri, Jul 04, 2008 at 04:22:02PM +0200, Tiziano M?ller wrote:
> >> One GLEP introduces new elements 'team', 'dev' and 'proxy':
> >> http://dev.gentoo.org/~dev-zero/glep/glep-new_metadata_elements.html
> > 1. With the addition of cpp, why
> > do we still need  elements?
> I'd see the herd element as kind of tag which could be used to search for
> packages (given the corresponding tools).
Clarify the GLEP to mention this please?

> > 2. "dev-zero". That's not possible to validate
> > sanely. The following would be better:
> > moodev
> > 
> > With the default value of the 'project' attribute being 'gentoo'.
> That's another good solution. The possibility I mentioned is rather suitable
> for others wanting to extend our metadata.xml format (and still be able to
> validate the content of gentoo:dev against our list of devs).
'gentoo:dev' isn't directly validable. It states element 'dev' in the
XML namespace of 'gentoo'. That's why I think that the the project
attribute variation would be more suitable, and also really easy for
other projects to handle.

When they are adding their own metadata.xml files, they can use an XSL
transform to add the explicit "project='gentoo'" for their own
metadata.xml files.

-- 
Robin Hugh Johnson
Gentoo Linux Developer & Infra Guy
E-Mail : [EMAIL PROTECTED]
GnuPG FP   : 11AC BA4F 4778 E3F6 E4ED  F38E B27B 944E 3488 4E85


pgpmEVxvsKQue.pgp
Description: PGP signature

Re: [gentoo-dev] Re: [2 GLEPs] metadata improvements

[Robin H. Johnson]

On Fri, Jul 04, 2008 at 10:03:10PM +0200, Tiziano M?ller wrote:
> > On Fri, Jul 04, 2008 at 04:22:02PM +0200, Tiziano M?ller wrote:
> >> And this is where the second GLEP comes in:
> >> http://dev.gentoo.org/~dev-zero/glep/glep-xsd.html
> > Don't remove the DTD DOCTYPE line. Leave it there.
> > Let simple tools be able to use the DTD to validate the well-formedness,
> > then upgrade to the XSD to check the actual semantics of the file.
> Very good point, thanks. But the changes of GLEP 46 can't be validated
> without allowing the "status" attribute for the toplevel maintainer. But I
> guess we could do it as long as it's documented (and commented) correctly.
Sure. In other places where DTD->XSD migrations have been done, there
ARE still changes made to the DTD for new elements/attributes etc.

-- 
Robin Hugh Johnson
Gentoo Linux Developer & Infra Guy
E-Mail : [EMAIL PROTECTED]
GnuPG FP   : 11AC BA4F 4778 E3F6 E4ED  F38E B27B 944E 3488 4E85


pgprIhTAdYQh0.pgp
Description: PGP signature

[gentoo-dev] Re: RFC: 0-day bump requests

[Tiziano Müller]

Robin H. Johnson wrote:

> On Fri, Jul 04, 2008 at 12:26:13AM +0100, Tony Chainsaw Vroon wrote:
>> Just an idea:
>> How about a metadata.xml tag that indicates whether early bump requests
>> are welcome? It's more of an individual developer preference, but that
>> seems the right place for it.
> If used, what about including and reviving the project that scraped
> Freshmeat and other spots looking for new releases automatically?
> 
-> GLEP 46, already approved. Need XSD to implement it properly (see my
other comment on).


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

[gentoo-dev] Re: [2 GLEPs] metadata improvements

[Tiziano Müller]

Robin H. Johnson wrote:

> On Fri, Jul 04, 2008 at 04:22:02PM +0200, Tiziano M?ller wrote:
>> And this is where the second GLEP comes in:
>> http://dev.gentoo.org/~dev-zero/glep/glep-xsd.html
> Don't remove the DTD DOCTYPE line. Leave it there.
> Let simple tools be able to use the DTD to validate the well-formedness,
> then upgrade to the XSD to check the actual semantics of the file.
> 

Very good point, thanks. But the changes of GLEP 46 can't be validated
without allowing the "status" attribute for the toplevel maintainer. But I
guess we could do it as long as it's documented (and commented) correctly.



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

[gentoo-dev] Kernel 2.6.25

[Christian Faulhammer]

Hi,

above series is going stable soon, so please check that the packages
you maintain have a stable version that is ok with .25.  If not, file a
stabilisation request and make it block bug 230285.  Thanks.

V-Li

-- 
Christian Faulhammer, Gentoo Lisp project
http://www.gentoo.org/proj/en/lisp/>, #gentoo-lisp on FreeNode

http://www.faulhammer.org/>


signature.asc
Description: PGP signature

[gentoo-dev] Re: [2 GLEPs] metadata improvements

[Tiziano Müller]

Robin H. Johnson wrote:

> On Fri, Jul 04, 2008 at 04:22:02PM +0200, Tiziano M?ller wrote:
>> One GLEP introduces new elements 'team', 'dev' and 'proxy':
>> http://dev.gentoo.org/~dev-zero/glep/glep-new_metadata_elements.html
> 1. With the addition of cpp, why
> do we still need  elements?

I'd see the herd element as kind of tag which could be used to search for
packages (given the corresponding tools).

> 
> 2. "dev-zero". That's not possible to validate
> sanely. The following would be better:
> moodev
> 
> With the default value of the 'project' attribute being 'gentoo'.
> 
That's another good solution. The possibility I mentioned is rather suitable
for others wanting to extend our metadata.xml format (and still be able to
validate the content of gentoo:dev against our list of devs).



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

Re: [gentoo-dev] [2 GLEPs] metadata improvements

[Robin H. Johnson]

On Fri, Jul 04, 2008 at 04:22:02PM +0200, Tiziano M?ller wrote:
> One GLEP introduces new elements 'team', 'dev' and 'proxy':
> http://dev.gentoo.org/~dev-zero/glep/glep-new_metadata_elements.html
1. With the addition of cpp, why
do we still need  elements?

2. "dev-zero". That's not possible to validate
sanely. The following would be better:
moodev

With the default value of the 'project' attribute being 'gentoo'.

-- 
Robin Hugh Johnson
Gentoo Linux Developer & Infra Guy
E-Mail : [EMAIL PROTECTED]
GnuPG FP   : 11AC BA4F 4778 E3F6 E4ED  F38E B27B 944E 3488 4E85


pgpribKEvUgIK.pgp
Description: PGP signature

Re: [gentoo-dev] [2 GLEPs] metadata improvements

[Robin H. Johnson]

On Fri, Jul 04, 2008 at 04:22:02PM +0200, Tiziano M?ller wrote:
> And this is where the second GLEP comes in:
> http://dev.gentoo.org/~dev-zero/glep/glep-xsd.html
Don't remove the DTD DOCTYPE line. Leave it there.
Let simple tools be able to use the DTD to validate the well-formedness,
then upgrade to the XSD to check the actual semantics of the file.

-- 
Robin Hugh Johnson
Gentoo Linux Developer & Infra Guy
E-Mail : [EMAIL PROTECTED]
GnuPG FP   : 11AC BA4F 4778 E3F6 E4ED  F38E B27B 944E 3488 4E85


pgpwgwF1vjSjK.pgp
Description: PGP signature

Re: [gentoo-dev] RFC: 0-day bump requests

[Robin H. Johnson]

On Fri, Jul 04, 2008 at 12:26:13AM +0100, Tony Chainsaw Vroon wrote:
> Just an idea:
> How about a metadata.xml tag that indicates whether early bump requests are 
> welcome?
> It's more of an individual developer preference, but that seems the right 
> place for it.
If used, what about including and reviving the project that scraped
Freshmeat and other spots looking for new releases automatically?

-- 
Robin Hugh Johnson
Gentoo Linux Developer & Infra Guy
E-Mail : [EMAIL PROTECTED]
GnuPG FP   : 11AC BA4F 4778 E3F6 E4ED  F38E B27B 944E 3488 4E85


pgp6uQHdkfwdU.pgp
Description: PGP signature

Re: [gentoo-dev] [2 GLEPs] metadata improvements

[Rémi Cardona]

Tiziano Müller wrote:

What do you think of them?


+1 on the first GLEP.

The second GLEP seems like a much better way of doing things, so +1 as 
well, but I am no xml expert :)


Cheers,

Rémi
--
gentoo-dev@lists.gentoo.org mailing list

Re: [gentoo-dev] RFC: 0-day bump requests

[Jeroen Roovers]

On Fri, 04 Jul 2008 00:26:13 +0100
"Tony \"Chainsaw\" Vroon" <[EMAIL PROTECTED]> wrote:

> On Fri, 2008-07-04 at 01:16 +0200, Jeroen Roovers wrote:
> > 1) How do you feel when you receive an early version bump request?
> 
> If it is for software where I am also upstream (Audacious for
> example), it does tend to annoy me when people try their utmost to
> file bug reports before I commit my ebuild. (I have yet to miss a
> release by more then 6 hours)
> 
> > 2) If you had your way, would you discourage users from filing early
> > version bump requests?
> 
> For things like the nVidia drivers I do welcome it. The time I can
> spend trawling upstream sites for new releases is limited.
> 
> Just an idea:
> How about a metadata.xml tag that indicates whether early bump
> requests are welcome? It's more of an individual developer
> preference, but that seems the right place for it.

Its' half an idea, in my opinion. We need a process, not just a tag in a
file. The tag in the file would tell us how a bug should perhaps be
treated, and metadata.xml is an excellent place to concentrate such
information, but to tell a bug wrangler (or anyone else) to "do nothing
for X units of time" isn't going to work. As for what the tag might
tell us, I think leaving bugs on hold for a few days is not the right
approach - users (as well as, say, fellow developers and upstreams)
shouldn't have to "artificially" wait to make their release
announcements and bug wranglers shouldn't be expected to keep these
bugs on their own lists in some artificial sense - it just means more
work for everyone and more delay in communications between users and
developers.

I am currently thinking of making a very broad division between
bump requests for more or less "independent" packages on the one hand,
and packages that (clearly) belong to a suite (KDE and GNOME are good
examples, although the latter team "owns" quite a few independantly
useable packages) or to wildly popular packages that announces releases
weeks to months ahead (Mozilla).

I personally think that bump requests of the "KDE 5 OMG" and "WHEREIS
FF4?" kind are to be RESOLVED as LATER forthwith. That saves a lot of
dupe checking as well! :)


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

Re: [gentoo-dev] [2 GLEPs] metadata improvements

[Luca Barbato]

Tiziano Müller wrote:

What do you think of them?


Both seem good, I'm looking forward to see them completed =)

lu

--

Luca Barbato
Gentoo Council Member
Gentoo/linux Gentoo/PPC
http://dev.gentoo.org/~lu_zero

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

[gentoo-dev] [2 GLEPs] metadata improvements

[Tiziano Müller]

Hi everyone

Since people are talking about metadata.xml again I'd like to present two
more GLEPs which will ease the process of auto-assigning bugs on one side
and make it a bit safer on the other side. Both of them are still in draft
state and if someone would like to contribute I'd be really happy.

One GLEP introduces new elements 'team', 'dev' and 'proxy':

http://dev.gentoo.org/~dev-zero/glep/glep-new_metadata_elements.html

The idea is to reduce some of the redundency we have in the tree and to
explicitly write who the maintainer is. It logically depends on my former
GLEP about Gentoo's herd/team metastructure, but doesn't depend on it in a
technical sense.
Furthermore it makes it possible to define maintainer-proxy relationships.

It would make the auto-assignment process based on metadata.xml safer since
it makes it possible to validate the maintainer.
And this is where the second GLEP comes in:

http://dev.gentoo.org/~dev-zero/glep/glep-xsd.html

In this GLEP I propose to switch from DTD to XSD for metadata.xml (and
herd.xml) to be able to validate content.
Even though some people like to argue that content should not be validated,
it makes sense to do it since the metadata.xml are not "prosa" but can be
interpreted as kind of database entries. Another reason is that
parent-child constraints can't be validated using DTD (especially GLEP 46
can't be properly implemented using DTD). But that's all written in the
GLEP...

What do you think of them?

Cheers,
Tiziano


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

Re: [gentoo-dev] RFC: 0-day bump requests

[Luca Barbato]

Jeroen Roovers wrote:

1) How do you feel when you receive an early version bump request?


They are useful as reminder.


2) If you had your way, would you discourage users from filing early
version bump requests?


I won't prevent anybody to send them, but I can understand why people 
would rather like to have less bugs in their mailbox.


lu

--

Luca Barbato
Gentoo Council Member
Gentoo/linux Gentoo/PPC
http://dev.gentoo.org/~lu_zero

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

Re: [gentoo-dev] RFC: 0-day bump requests

[Marijn Schouten (hkBst)]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

Jeroen Roovers wrote:
>  Hi fellow developers,
> 
> 
> it seems I've run into a minor issue with fellow bug wrangler carlo
> (who has been putting a lot of work into that, for which we should all
> be grateful).
> 
> Carsten has a cut-and-paste message that he posts in comments to
> version bump bug reports that he finds have been filed on the day the
> software version in question was released/announced. The gist of the
> message is that none of or most ebuild developers do not like these
> "0-day requests" and that users (and developers) should refrain from
> filing them on the same day. Waiting a week would be OK, the message
> seems to say.
> 
> Being an ebuild developer myself, I have to say that I do not hold that
> stance and that I welcome early version bump requests. Therefore, I
> refrain from adding such messages to the bugs that I wrangle and indeed
> welcome any bump requests[1].
> 
> Finding myself in conflict with someone I have come to share a certain
> workload with, notably someone who has a few more years of Gentoo
> experience, I wonder what the majority of our ebuild developers
> actually think. In that spirit, I hope the following questions are
> neutral enough for everyone to *not* start a flamewar over this. :)
> 
> 
> -
> 1) How do you feel when you receive an early version bump request?

Since current mores make sure there are not so many, I don't mind them.

> 2) If you had your way, would you discourage users from filing early
> version bump requests?

To prevent every package from getting a 0-day bump request, I'd say give it a
day or two at least, unless you have some info other than that there is a new
version. For example that the current ebuild still works with the new version or
that it doesn't. It helps with gauging which bumps are trivial and which aren't.

If someone only wants to tell me some new version is out, I prefer they ping me
on irc.

> -
> 
> I know, it's not a particularly good survey, but I hope the plenty and
> diversity of your answers will shed more light on the matter. :)
> 
> 
> Thank you and kind regards,
>  JeR
> 
> 
> [1] In fact I regularly use the opportunity to check on the HOMEPAGE
> whether the release was security related, and I assign directly to
> security@ when that is the case (CC'ing the package's maintainers) and
> perhaps pasting ChangeLog or advisory info in a comment.

Marijn

- --
Marijn Schouten (hkBst), Gentoo Lisp project, Gentoo ML
, #gentoo-{lisp,ml} on FreeNode
-BEGIN PGP SIGNATURE-
Version: GnuPG v2.0.9 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAkht4WkACgkQp/VmCx0OL2xJ6QCfbX/IvrzARx3AY2FzAHW4sD2P
TasAn2NTD0c+HE0ehaG3wd9bFdk+yzSh
=pj1H
-END PGP SIGNATURE-
-- 
gentoo-dev@lists.gentoo.org mailing list

Re: [gentoo-dev] RFC: 0-day bump requests

[Ferris McCormick]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On Fri, 4 Jul 2008 01:16:09 +0200
Jeroen Roovers <[EMAIL PROTECTED]> wrote:

>  Hi fellow developers,
> 
> 
> it seems I've run into a minor issue with fellow bug wrangler carlo
> (who has been putting a lot of work into that, for which we should all
> be grateful).
> 
> Carsten has a cut-and-paste message that he posts in comments to
> version bump bug reports that he finds have been filed on the day the
> software version in question was released/announced. The gist of the
> message is that none of or most ebuild developers do not like these
> "0-day requests" and that users (and developers) should refrain from
> filing them on the same day. Waiting a week would be OK, the message
> seems to say.
> 
> Being an ebuild developer myself, I have to say that I do not hold that
> stance and that I welcome early version bump requests. Therefore, I
> refrain from adding such messages to the bugs that I wrangle and indeed
> welcome any bump requests[1].
> 
> Finding myself in conflict with someone I have come to share a certain
> workload with, notably someone who has a few more years of Gentoo
> experience, I wonder what the majority of our ebuild developers
> actually think. In that spirit, I hope the following questions are
> neutral enough for everyone to *not* start a flamewar over this. :)
> 
> 
> -
> 1) How do you feel when you receive an early version bump request?
> 
Speaking only for myself as an arch developer.

It depends on the reason.  For example, recently there was a day 0
request for a freetype (I believe) stable request because current
stable didn't work is some such.  That sort of thing is OK.  Obviously
security bugs require quick processing.  New keyword/re-keyword
requests are OK (but then of course we don't go stable).

Otherwise, we will put the package into the normal cycle whenever it
enters the tree.
 
> 
> 2) If you had your way, would you discourage users from filing early
> version bump requests?
> 
> 
Makes no difference to me, but I am not a package maintainer.  I am
speaking from an arch point of view.  We only ask that the package
maintainer make sure it at least seems to work before they bump the
version.

(It's different when the new version is not compatible with the current
version, but that's off topic for this thread, I think.  I don't ever
want to see that sort of thing.)

> -
> 
> I know, it's not a particularly good survey, but I hope the plenty and
> diversity of your answers will shed more light on the matter. :)
> 
> 
> Thank you and kind regards,
>  JeR
> 
> 
> [1] In fact I regularly use the opportunity to check on the HOMEPAGE
> whether the release was security related, and I assign directly to
> security@ when that is the case (CC'ing the package's maintainers) and
> perhaps pasting ChangeLog or advisory info in a comment.
> -- 
> gentoo-dev@lists.gentoo.org mailing list
> 

I doubt that this addresses what you are asking, but in case it is
useful,
Regards,
Ferris
- --
Ferris McCormick (P44646, MI) <[EMAIL PROTECTED]>
Developer, Gentoo Linux (Sparc, Devrel, Userrel, Trustees)
-BEGIN PGP SIGNATURE-
Version: GnuPG v2.0.9 (GNU/Linux)

iEYEARECAAYFAkht270ACgkQQa6M3+I///edXwCfTPiTZ56Aw9ViJRs8hJTm8DrQ
7g4An1NdsU/hLteSFLmxT47eeWDEGehm
=62NW
-END PGP SIGNATURE-

Re: [gentoo-dev] RFC: 0-day bump requests

[Petteri Räty]

Jeroen Roovers kirjoitti:


-
1) How do you feel when you receive an early version bump request?



I don't mind as I don't have the time to follow upstreams that closely 
for the hundreds of Java packages out there.




2) If you had your way, would you discourage users from filing early
version bump requests?



I think the best approach is to discourage them from filing the bump 
requests in before a week or so has passed but assign the bug to the 
maintainer. Closing it has the problem that I doubt most users remember 
to check for closed bugs when filing bump requests so this will lead to 
more bump requests bugs being filed on the same version.


Regards,
Petter



signature.asc
Description: OpenPGP digital signature