Re: [Mageia-dev] Mirror tree structure

2010-10-20 Thread Lombard Marianne
Le 20/10/2010 18:34, Olivier Thauvin a écrit :
> Hi,
> 
> You can find here:
> http://distrib-coffee.ipsl.jussieu.fr/pub/linux/Mageia/
> the current mirror tree proposal.
> We now have to discuss it, I think.
> 
> Here notes:
> 
> The 'mageia_timestamp' is a file updated on the main server every 5
> minutes. It allow to check is a mirror synced or not
> 
> Comparing to mandriva there is no more split between devel/stable, all
> distributions goes into distribs/ and all isos file into iso/.
> 
I've seen that the distribs directory has sub-directory called stable1
and cauldron, will this sub-directory will be present in iso/ too ?

> The updates/ tree disapear, avoiding some possible dependencies issues
> in updates rpms.
> 
> The peoples/ directory is dedicated to contributors and must allow
> anyone to share files related to the distribution (testing rpms,
> preworks) in an unofficial ways.
> How this can be setup in practice still have to be discuss.
> 
> Finally the software/ directory should allow you to distribute the
> tarball (not rpm) for software we do. As soon we do free software people
> must be able to distribute our code in the same way any projects does.
> Replies like "svn is readable" or "we have rpm" is not appropriated.
> 
> Now come the question: "what is a valid mirror ?", eg, what a mirror
> should have as file to be valid ?
> 
> I suggest to not give the choice and avoid mistake by saying except
> "peoples" a mirror must respect the whole tree to be valid.
> This mean everything must exists with this structure under the top level
> path.
> This way may avoid issues like mandriva on ibiblio (only 2005 and 2007.1
> seems to exists...)
> ftp://ibiblio.org/pub/linux/distributions/mandriva/Mandrivalinux/
> 
> Comment and idea welcome.
> 
> Regards.
> 

Nothing else to say

Regards
-- 
Marianne (Jehane) Lombard
Mandriva User - Mageia french translation team
Inside every fat girl, there is a thin girl waiting to get out (and a
lot of chocolate) - Terry Pratchett


Re: [Mageia-dev] Mirror tree structure

2010-10-20 Thread Wolfgang Bornath
While leaving possible changes to further discussion, as a potential
mirror maintainer my first question is:

How much harddisk space will the "valid" mirror need?


Re: [Mageia-dev] Mirror tree structure

2010-10-20 Thread Tux99
On Wed, 20 Oct 2010, Olivier Thauvin wrote:

> Hi,
> 
> You can find here:
> http://distrib-coffee.ipsl.jussieu.fr/pub/linux/Mageia/
> the current mirror tree proposal.
> We now have to discuss it, I think.

Looks great to me and I agree with all points you made in your email.
I especially like the provision of the people (no 's'!) and software 
directory, seem like a good idea to me.



Re: [Mageia-dev] Mirror tree structure

2010-10-20 Thread Tux99
On Wed, 20 Oct 2010, Wolfgang Bornath wrote:

> While leaving possible changes to further discussion, as a potential
> mirror maintainer my first question is:
> 
> How much harddisk space will the "valid" mirror need?
> 

RTFM :)
http://distrib-coffee.ipsl.jussieu.fr/pub/linux/Mageia/mirror.readme



Re: [Mageia-dev] Mirror tree structure

2010-10-20 Thread Wolfgang Bornath
2010/10/20 Tux99 :
>
> RTFM :)
> http://distrib-coffee.ipsl.jussieu.fr/pub/linux/Mageia/mirror.readme

I read the FM but I thought it may be outdated because it was written
almost 3 weeks before this new information about the tree structure
was posted. 3 weeks in this stage means a long time

-- 
wobo


Re: [Mageia-dev] Mirror tree structure

2010-10-20 Thread Wolfgang Bornath
2010/10/20 Wolfgang Bornath :
>
> I read the FM but I thought it may be outdated because it was written
> almost 3 weeks before this new information about the tree structure
> was posted. 3 weeks in this stage means a long time

If the readme is showing realistic numbers I am afraid our server
can't act as mirror because we only have 2x500 GB harddisks.

Currently we are mirroring Mandriva 2009.1, 2010.0 and 2010.1 -
altogether 360G (without Cooker). We were ready to reduce this to the
current Mandriva version to make place for Mageia. But 700 is too
much, sorry.

-- 
wobo


Re: [Mageia-dev] Mirror tree structure

2010-10-20 Thread Remy CLOUARD
On Wed, Oct 20, 2010 at 06:34:24PM +0200, Olivier Thauvin wrote:
> Hi,
> 
> The peoples/ directory is dedicated to contributors and must allow
> anyone to share files related to the distribution (testing rpms,
> preworks) in an unofficial ways.
> How this can be setup in practice still have to be discuss.
Hi, I would like to ask some details about this:
As some of you might know, I’m using awesome for a window manager and
have been providing RPMS for quite some time for it on my website.

May I use the peoples repository to provide it as well as a customized
cairo (with xcb enabled) ?

Will users be able to add this as another repository ?

I must admit the process was quite painful for me up till now, this
sounds to me like an easy way to provide and maintain this RPM.

If we switch to git for the VCS, I would just make my own branch on top
of the regular cairo/awesome branch, and it would definitely ease this
process. WDYT ?

Regards,
-- 
Rémy CLOUARD
() ascii ribbon campaign - against html e-mail
/\ www.asciiribbon.org - against proprietary attachments


pgpRL5WKB09rk.pgp
Description: PGP signature


Re: [Mageia-dev] Mirror tree structure

2010-10-20 Thread Olivier Thauvin
* Lombard Marianne (maria...@tuxette.fr) wrote:
> Le 20/10/2010 18:34, Olivier Thauvin a écrit :
> > Hi,
> > 
> > You can find here:
> > http://distrib-coffee.ipsl.jussieu.fr/pub/linux/Mageia/
> > the current mirror tree proposal.
> > We now have to discuss it, I think.
> > 
> > Here notes:
> > 
> > The 'mageia_timestamp' is a file updated on the main server every 5
> > minutes. It allow to check is a mirror synced or not
> > 
> > Comparing to mandriva there is no more split between devel/stable, all
> > distributions goes into distribs/ and all isos file into iso/.
> > 
> I've seen that the distribs directory has sub-directory called stable1
> and cauldron, will this sub-directory will be present in iso/ too ?

As soon stable1 means "our first distrib" yes, but I don't think
cauldron will be needed under iso/.

-- 

Olivier Thauvin
CNRS  -  LATMOS
♖ ♘ ♗ ♕ ♔ ♗ ♘ ♖


pgpDyV1EYwtIb.pgp
Description: PGP signature


Re: [Mageia-dev] Mirror tree structure

2010-10-20 Thread Olivier Thauvin
* Wolfgang Bornath (molc...@googlemail.com) wrote:
> 2010/10/20 Wolfgang Bornath :
> >
> > I read the FM but I thought it may be outdated because it was written
> > almost 3 weeks before this new information about the tree structure
> > was posted. 3 weeks in this stage means a long time
> 
> If the readme is showing realistic numbers I am afraid our server
> can't act as mirror because we only have 2x500 GB harddisks.
> 
> Currently we are mirroring Mandriva 2009.1, 2010.0 and 2010.1 -
> altogether 360G (without Cooker). We were ready to reduce this to the
> current Mandriva version to make place for Mageia. But 700 is too
> much, sorry.

I am not sure our mirror will reach 700MB but sice the current cooker
tree is near of 80Go (w/o iso), and if I take as base 2 release per
year, then I obtain 700GB in three years.

-- 

Olivier Thauvin
CNRS  -  LATMOS
♖ ♘ ♗ ♕ ♔ ♗ ♘ ♖


pgpczlZDB30ha.pgp
Description: PGP signature


Re: [Mageia-dev] Mirror tree structure

2010-10-20 Thread Wolfgang Bornath
2010/10/21 Olivier Thauvin :
>
> I am not sure our mirror will reach 700MB but sice the current cooker
> tree is near of 80Go (w/o iso), and if I take as base 2 release per
> year, then I obtain 700GB in three years.

So 350Go would be enough for the current version plus the previous 2
versions (including ISOs). Would that be a "valid" mirror?


Re: [Mageia-dev] Mirror tree structure

2010-10-20 Thread Olivier Thauvin
* Remy CLOUARD (shikam...@mandriva.org) wrote:
> On Wed, Oct 20, 2010 at 06:34:24PM +0200, Olivier Thauvin wrote:
> > Hi,
> > 
> > The peoples/ directory is dedicated to contributors and must allow
> > anyone to share files related to the distribution (testing rpms,
> > preworks) in an unofficial ways.
> > How this can be setup in practice still have to be discuss.
> Hi, I would like to ask some details about this:
> As some of you might know, I’m using awesome for a window manager and
> have been providing RPMS for quite some time for it on my website.
> 
> May I use the peoples repository to provide it as well as a customized
> cairo (with xcb enabled) ?

As a mirror manager I say yes: I'll discuss with sys-admin team about
how we can setup this, with a per user disk space limit (to avoid
issue).

From a project point of view it is veey different. The mirror tree is
what some people will see about Mageia and I do think:
- people content must be related to mageia
- everything that can be included in the distrib must provide as RPMs
  using BS
- people content must respect our policies.

In a nutshell: this must be discuss. Having a people is at time just an
idea, but it solve a need I had in mandriva to provide some rpms for
testing purpose.

> 
> Will users be able to add this as another repository ?

I don't know any process to deny someone to push hdlists and rpms.
However I don't think it is a good idea to use this space as an
alternative to oficial mageia. The risk is to confuse people.

Also, I'd like to notice this space is not a way to provide your own
project, there is platform for that (sourceforge, zarb, gnu, etc...)

> 
> I must admit the process was quite painful for me up till now, this
> sounds to me like an easy way to provide and maintain this RPM.
> 
> If we switch to git for the VCS, I would just make my own branch on top
> of the regular cairo/awesome branch, and it would definitely ease this
> process. WDYT ?
> 
> Regards,
> -- 
> Rémy CLOUARD
> () ascii ribbon campaign - against html e-mail
> /\ www.asciiribbon.org - against proprietary attachments


-- 

Olivier Thauvin
CNRS  -  LATMOS
♖ ♘ ♗ ♕ ♔ ♗ ♘ ♖


pgp6gMaSod0Xx.pgp
Description: PGP signature


Re: [Mageia-dev] Mirror tree structure

2010-10-20 Thread Olivier Thauvin
* Wolfgang Bornath (molc...@googlemail.com) wrote:
> 2010/10/21 Olivier Thauvin :
> >
> > I am not sure our mirror will reach 700MB but sice the current cooker
> > tree is near of 80Go (w/o iso), and if I take as base 2 release per
> > year, then I obtain 700GB in three years.
> 
> So 350Go would be enough for the current version plus the previous 2
> versions (including ISOs). Would that be a "valid" mirror?

If the rules I suggested are applied and at time, yes.
But be aware the tree will grow release after release.

I am waiting the release cycle to be defined to know how much space
we'll need. If we choose to do only one release per year, the 700GB size
will occur only in 6 years. And if we don't import all rpms for mandriva
the size need will be less again.

Regards.

-- 

Olivier Thauvin
CNRS  -  LATMOS
♖ ♘ ♗ ♕ ♔ ♗ ♘ ♖


pgpMO3HyAaxWy.pgp
Description: PGP signature


Re: [Mageia-dev] Mirror tree structure

2010-10-20 Thread Wolfgang Bornath
2010/10/21 Olivier Thauvin :
>
> I am waiting the release cycle to be defined to know how much space
> we'll need. If we choose to do only one release per year, the 700GB size
> will occur only in 6 years. And if we don't import all rpms for mandriva
> the size need will be less again.

Yes, there are some variables not yet defined but at least for the
first 3-4 releases our space will do :)

Thx for explaining. Suggestion: I think it will avoid other people
having the same question if you add a short sentence to the readme
explaining that the 700Go is the cumulated size for 3 years :)

-- 
wobo


Re: [Mageia-dev] Mirror tree structure

2010-10-20 Thread J.A. Magallón
On Wed, 20 Oct 2010 18:34:24 +0200, Olivier Thauvin 
 wrote:

> Hi,
...
> 
> Now come the question: "what is a valid mirror ?", eg, what a mirror
> should have as file to be valid ?
> 

Some ideas (probably silly ;)).

Why don't you split mirrors in 2 categories:
- Software (RPM) repo mirrors, available for network install or urpmi.
  (distrib+people+software).
- ISO mirrors. Avaliable for torrenting or ftp. MDV2010.1 release isos are
  about 15Gb. With 2 releases per year, thats 30Gb per year, 200 Gb on 6
  years. Some places can just mirror this...you lower the 700Gb to 500,
  not bad.

so mirroring can be done by version (or was this what you wanted to avoid?).

BTW

- could you kill the final 's' on all the names ? 
  distribs -> distrib
  iso (you didnt named it 'isos')
  peoples -> people
  software (not softwares)
- could you not mix case in names (SRPMS <-> x86_64), just srpms...
- could be the arch names more uniform ? in my personal scripts/setups
  I use x86-32 and x86-64. Moreover, perhaps in a not so near future some
  adventurous soul builds Mageia on ARM or Sparc, so why not sort things
  like
distrib/cauldron/srpm
distrib/cauldron/x86/32/iso
   /rpm
/64
/arm/32
/64
/sparc/32
  /64
  The drawback is that ISOs are in the same tree so no separate mirroring
  is possible, but I think you didn't want this anyways.

All this renaming in case it is not some kind of standard for tree layout
or the like...


Just my 2€cent, you will judge...

-- 
J.A. Magallon  \   Software is like sex:
 \ It's better when it's free


Re: [Mageia-dev] Mirror tree structure

2010-10-20 Thread Olivier Thauvin
* J.A. Magallón (jamagal...@ono.com) wrote:
> On Wed, 20 Oct 2010 18:34:24 +0200, Olivier Thauvin 
>  wrote:
> 
> > Hi,
> ...
> > 
> > Now come the question: "what is a valid mirror ?", eg, what a mirror
> > should have as file to be valid ?
> > 
> 
> Some ideas (probably silly ;)).
> 
> Why don't you split mirrors in 2 categories:
> - Software (RPM) repo mirrors, available for network install or urpmi.
>   (distrib+people+software).
> - ISO mirrors. Avaliable for torrenting or ftp. MDV2010.1 release isos are
>   about 15Gb. With 2 releases per year, thats 30Gb per year, 200 Gb on 6
>   years. Some places can just mirror this...you lower the 700Gb to 500,
>   not bad.
> 
> so mirroring can be done by version (or was this what you wanted to avoid?).

Yes, I tried to avoid it because I perfectly know mirrors will take care
to mirror new version every 6 months (ibiblio example again, nobody had
a look for several years, at least since 2008).

> 
> BTW
> 
> - could you kill the final 's' on all the names ? 
>   distribs -> distrib

Done

>   iso (you didnt named it 'isos')
>   peoples -> people

Done

>   software (not softwares)

Is already 'software'

> - could you not mix case in names (SRPMS <-> x86_64), just srpms...

This part depend on BS and how this team working on it will organise the
distrib. I'll suggest them.

> - could be the arch names more uniform ? in my personal scripts/setups
>   I use x86-32 and x86-64. Moreover, perhaps in a not so near future some
>   adventurous soul builds Mageia on ARM or Sparc, so why not sort things
>   like
> distrib/cauldron/srpm
> distrib/cauldron/x86/32/iso
>/rpm
> /64
> /arm/32
> /64
> /sparc/32
>   /64

This depend again on BS part.

>   The drawback is that ISOs are in the same tree so no separate mirroring
>   is possible, but I think you didn't want this anyways.

I'll take care to this.

> 
> All this renaming in case it is not some kind of standard for tree layout
> or the like...
> 
> 
> Just my 2€cent, you will judge...
> 
> -- 
> J.A. Magallon  \   Software is like sex:
>  \ It's better when it's free
-- 

Olivier Thauvin
CNRS  -  LATMOS
♖ ♘ ♗ ♕ ♔ ♗ ♘ ♖


pgppwn2IlhYIv.pgp
Description: PGP signature


Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread Romain d'Alverny
On Wed, Oct 20, 2010 at 18:34, Olivier Thauvin
 wrote:
>
> Now come the question: "what is a valid mirror ?", eg, what a mirror
> should have as file to be valid ?

Not sure if we discussed in depth MirrorBrain
(http://www.mirrorbrain.org/ ) for managing mirrors index and
redirections.

If we were going to use it, could we, for instance, leave mirrors some
liberty to mirror what branch they want (with some guidances and
preferences of course) and let our MirrorBrain instance check and
build the list of valid mirrors for the file actually requested?

This, provided that _consistent_ branches of the tree are mirrored,
and not only a file here, a file there.

On one hand, this would introduce at least to other things to check:
 - having enough distributed mirrors that map the whole tree;
 - having download/install tools take this into account.

On the other hand, this could allow more mirrors to take part in this,
in that it may require less storage space and less bandwidth usage.

It's not the only reason to use MirrorBrain anyway, but I wondered if
this could be a complementary reason.

Not sure, insights welcome.


Cheers,

Romain


Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread Maarten Vanraes
Op woensdag 20 oktober 2010 18:34:24 schreef Olivier Thauvin:
> Hi,
> 
> You can find here:
> http://distrib-coffee.ipsl.jussieu.fr/pub/linux/Mageia/
> the current mirror tree proposal.
> We now have to discuss it, I think.
> 
> Here notes:
> 
> The 'mageia_timestamp' is a file updated on the main server every 5
> minutes. It allow to check is a mirror synced or not
> 
> Comparing to mandriva there is no more split between devel/stable, all
> distributions goes into distribs/ and all isos file into iso/.
> 
> The updates/ tree disapear, avoiding some possible dependencies issues
> in updates rpms.
> 
> The peoples/ directory is dedicated to contributors and must allow
> anyone to share files related to the distribution (testing rpms,
> preworks) in an unofficial ways.
> How this can be setup in practice still have to be discuss.
> 
> Finally the software/ directory should allow you to distribute the
> tarball (not rpm) for software we do. As soon we do free software people
> must be able to distribute our code in the same way any projects does.
> Replies like "svn is readable" or "we have rpm" is not appropriated.
> 
> Now come the question: "what is a valid mirror ?", eg, what a mirror
> should have as file to be valid ?
> 
> I suggest to not give the choice and avoid mistake by saying except
> "peoples" a mirror must respect the whole tree to be valid.
> This mean everything must exists with this structure under the top level
> path.
> This way may avoid issues like mandriva on ibiblio (only 2005 and 2007.1
> seems to exists...)
> ftp://ibiblio.org/pub/linux/distributions/mandriva/Mandrivalinux/
> 
> Comment and idea welcome.
> 
> Regards.


i didn't follow the whole previous discussion about the mirror tree, but would 
it be advisable to have a noarch subdir, next to i586 and x86_64 ?

it would mean that if you use i586, you'd need noarch too; and the same thing 
with x86_64; but imo, the setup might use a little less storage or it'd be 
easier set up; and less hacks required?

if you'd do a dual arch CD; you could easily have noarch, i586 and x86_64 next 
to eachother; imo.

also, if we split up more packages to use noarch for content stuff or xxx-doc 
subpackages or whatever...


Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread Samuel Verschelde
Le jeudi 21 octobre 2010 00:44:16, Olivier Thauvin a écrit :
> In a nutshell: this must be discuss. Having a people is at time just an
> idea, but it solve a need I had in mandriva to provide some rpms for
> testing purpose.

I don't understand very well what kind of packages would go into "people" 
rather than in dedicated testing media. What would be the difference ?

Regards

Samuel Verschelde


Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread Samuel Verschelde
Le jeudi 21 octobre 2010 00:47:54, Olivier Thauvin a écrit :
> * Wolfgang Bornath (molc...@googlemail.com) wrote:
> > 2010/10/21 Olivier Thauvin :
> > >
> > > I am not sure our mirror will reach 700MB but sice the current cooker
> > > tree is near of 80Go (w/o iso), and if I take as base 2 release per
> > > year, then I obtain 700GB in three years.
> > 
> > So 350Go would be enough for the current version plus the previous 2
> > versions (including ISOs). Would that be a "valid" mirror?
> 
> If the rules I suggested are applied and at time, yes.
> But be aware the tree will grow release after release.
> 
> I am waiting the release cycle to be defined to know how much space
> we'll need. If we choose to do only one release per year, the 700GB size
> will occur only in 6 years. And if we don't import all rpms for mandriva
> the size need will be less again.

If we choose to do only one release per year, the backports media may grow 
faster, but indeed probably less rapidly than a whole release tree.

Samuel


Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread Remy CLOUARD
On Thu, Oct 21, 2010 at 12:44:16AM +0200, Olivier Thauvin wrote:
> * Remy CLOUARD (shikam...@mandriva.org) wrote:
> > On Wed, Oct 20, 2010 at 06:34:24PM +0200, Olivier Thauvin wrote:
> > > Hi,
> > > 
> > > The peoples/ directory is dedicated to contributors and must allow
> > > anyone to share files related to the distribution (testing rpms,
> > > preworks) in an unofficial ways.
> > > How this can be setup in practice still have to be discuss.
> > Hi, I would like to ask some details about this:
> > As some of you might know, I’m using awesome for a window manager and
> > have been providing RPMS for quite some time for it on my website.
> > 
> > May I use the peoples repository to provide it as well as a customized
> > cairo (with xcb enabled) ?
> 
> As a mirror manager I say yes: I'll discuss with sys-admin team about
> how we can setup this, with a per user disk space limit (to avoid
> issue).
> 
> From a project point of view it is veey different. The mirror tree is
> what some people will see about Mageia and I do think:
> - people content must be related to mageia
Well, RPMS compiled for mageia are by definition related to mageia :-)
> - everything that can be included in the distrib must provide as RPMs
>   using BS
Well this would allow the chroot to fetch packages from these
alternatives repositories and could automate the process.
> - people content must respect our policies.
> 
> In a nutshell: this must be discuss. Having a people is at time just an
> idea, but it solve a need I had in mandriva to provide some rpms for
> testing purpose.
> 
> > 
> > Will users be able to add this as another repository ?
> 
> I don't know any process to deny someone to push hdlists and rpms.
> However I don't think it is a good idea to use this space as an
> alternative to oficial mageia. The risk is to confuse people.
Well, I don’t think it’s about confusing people, after all these
packages are in a different subtree than regular packages, rather
giving choice. I know at least one person who said on #mageia-de that
mageia would be a no-go for him if he couldn’t have awesome.

In Gentoo you have the ability to define your own USE flags to tweak
your system to your needs. Here’s what it says when you activate the xcb
flag and emerge cairo:

WARN: postinst
You have enabled the Cairo XCB backend which is used only by
a select few apps. The Cairo XCB backend is presently
un-maintained and needs a lot of work to get it caught up
to the Xrender and Xlib backends, which are the backends used
by most applications. See:
http://lists.freedesktop.org/archives/xcb/2008-December/004139.html

I thought about using a README.urpmi to show in the %post section in a
similar way. Of course we’re not gentoo and as we provide precompiled
packages it doesn’t allow us to do things in a similar way (though we
could provide some mechanisms like AUR for arch).

The whole idea is that some people won’t feel “ghettoized” because they
use something different than KDE or GNOME, though the concept can be
extended to anything and not just window managers/desktop environments.
> 
> Also, I'd like to notice this space is not a way to provide your own
> project, there is platform for that (sourceforge, zarb, gnu, etc...)
Don’t get me wrong here, I was not talking about hosting a project, just
packages for mageia.

Regards,
-- 
Rémy CLOUARD
() ascii ribbon campaign - against html e-mail
/\ www.asciiribbon.org - against proprietary attachments


pgpkA37F3TE4n.pgp
Description: PGP signature


Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread Olivier Thauvin
* Maarten Vanraes (maarten.vanr...@gmail.com) wrote:
> Op woensdag 20 oktober 2010 18:34:24 schreef Olivier Thauvin:
> > Hi,
> > 
> > You can find here:
> > http://distrib-coffee.ipsl.jussieu.fr/pub/linux/Mageia/
> > the current mirror tree proposal.
> > We now have to discuss it, I think.
> > 
> i didn't follow the whole previous discussion about the mirror tree, but 
> would 
> it be advisable to have a noarch subdir, next to i586 and x86_64 ?

This is not on my side. The internal distribution tree structure is more
related to build system than the global mirror tree structure.

At least, the tree I am suggesting neither deny it, neither make it
easier.

Regards.

-- 

Olivier Thauvin
CNRS  -  LATMOS
♖ ♘ ♗ ♕ ♔ ♗ ♘ ♖


pgp0edinOgwxz.pgp
Description: PGP signature


Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread Olivier Blin
Olivier Thauvin  writes:

>> I've seen that the distribs directory has sub-directory called stable1
>> and cauldron, will this sub-directory will be present in iso/ too ?
>
> As soon stable1 means "our first distrib" yes, but I don't think
> cauldron will be needed under iso/.

It could be if we build snapshots of the dev distro automatically

-- 
Olivier Blin - blino


Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread Olivier Thauvin
* Samuel Verschelde (sto...@laposte.net) wrote:
> Le jeudi 21 octobre 2010 00:44:16, Olivier Thauvin a écrit :
> > In a nutshell: this must be discuss. Having a people is at time just an
> > idea, but it solve a need I had in mandriva to provide some rpms for
> > testing purpose.
> 
> I don't understand very well what kind of packages would go into "people" 
> rather than in dedicated testing media. What would be the difference ?

I can give the example I had on Mandriva: I did some work on 'rpm'
package itself, but since breaking this pakage would have lock the whole
Bs I'd prefer to first have it tested by some people.
This directory is in this case the perfect place. At this time,
"testing" media did not exists thought.

Another example could be the XFCE live Cd done by some people in past,
hosted on distrib-coffee as mandriva where not able to push it on thier
mirror (this issue get solved latelly).

Pushing something on the mageia Mirror need some control. The people
place give the opportunity to contributors to distribute somethings w/o
needing "super user" privileges, but with a clear "unofficial" state.

I am pretty sure some people will find usage to it.

> 
> Regards
> 
> Samuel Verschelde
-- 

Olivier Thauvin
CNRS  -  LATMOS
♖ ♘ ♗ ♕ ♔ ♗ ♘ ♖


pgpM15teuq82x.pgp
Description: PGP signature


Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread Olivier Thauvin
* Remy CLOUARD (shikam...@mandriva.org) wrote:
> On Thu, Oct 21, 2010 at 12:44:16AM +0200, Olivier Thauvin wrote:
> > * Remy CLOUARD (shikam...@mandriva.org) wrote:
> > Also, I'd like to notice this space is not a way to provide your own
> > project, there is platform for that (sourceforge, zarb, gnu, etc...)
> Don’t get me wrong here, I was not talking about hosting a project, just
> packages for mageia.

I was just listing potential issue I was thinking to. I am not saying
people/ is not the right place for such thing neither ;)



-- 

Olivier Thauvin
CNRS  -  LATMOS
♖ ♘ ♗ ♕ ♔ ♗ ♘ ♖


pgpQUXBCYUeUa.pgp
Description: PGP signature


Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread Olivier Thauvin
* Olivier Blin (mag...@blino.org) wrote:
> Olivier Thauvin  writes:
> 
> >> I've seen that the distribs directory has sub-directory called stable1
> >> and cauldron, will this sub-directory will be present in iso/ too ?
> >
> > As soon stable1 means "our first distrib" yes, but I don't think
> > cauldron will be needed under iso/.
> 
> It could be if we build snapshots of the dev distro automatically

In this case, yes of course. "I don't think" does not mean it's a
reality for ever ;) It's just my current feeling according information I
have ;)

> 
> -- 
> Olivier Blin - blino
-- 

Olivier Thauvin
CNRS  -  LATMOS
♖ ♘ ♗ ♕ ♔ ♗ ♘ ♖


pgpHbBEBO83z2.pgp
Description: PGP signature


Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread Renaud MICHEL
On jeudi 21 octobre 2010 at 21:36, Remy CLOUARD wrote :
> On Thu, Oct 21, 2010 at 12:44:16AM +0200, Olivier Thauvin wrote:
> > I don't know any process to deny someone to push hdlists and rpms.
> > However I don't think it is a good idea to use this space as an
> > alternative to oficial mageia. The risk is to confuse people.
> 
> Well, I don’t think it’s about confusing people, after all these
> packages are in a different subtree than regular packages, rather
> giving choice. I know at least one person who said on #mageia-de that
> mageia would be a no-go for him if he couldn’t have awesome.

awesome is already in mandriva, why wouldn't it be in mageia?
If there is someone willing to package it, he may as well do it in the core 
repository (be it main or contrib, or whatever they will be called).

-- 
Renaud Michel


Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread Michael Scherer
Le mercredi 20 octobre 2010 à 18:34 +0200, Olivier Thauvin a écrit :
> Hi,
> 
> You can find here:
> http://distrib-coffee.ipsl.jussieu.fr/pub/linux/Mageia/
> the current mirror tree proposal.
> We now have to discuss it, I think.
> 
> Here notes:
> 
> The 'mageia_timestamp' is a file updated on the main server every 5
> minutes. It allow to check is a mirror synced or not
> 
> Comparing to mandriva there is no more split between devel/stable, all
> distributions goes into distribs/ and all isos file into iso/.
> 
> The updates/ tree disapear, avoiding some possible dependencies issues
> in updates rpms.
> 
> The peoples/ directory is dedicated to contributors and must allow
> anyone to share files related to the distribution (testing rpms,
> preworks) in an unofficial ways.
> How this can be setup in practice still have to be discuss.

Like quota, etc ?

> Finally the software/ directory should allow you to distribute the
> tarball (not rpm) for software we do. As soon we do free software people
> must be able to distribute our code in the same way any projects does.
> Replies like "svn is readable" or "we have rpm" is not appropriated.
> 
> Now come the question: "what is a valid mirror ?", eg, what a mirror
> should have as file to be valid ?
> 
> I suggest to not give the choice and avoid mistake by saying except
> "peoples" a mirror must respect the whole tree to be valid.

I would also exclude software/ from regular mirror, because this may not
requires as much redundancy than regular mirror ( ie, I doubt many
people will download tarball on a regular basis, and most softwre
project are fine with 2 or 3 mirror ).

On the other hand, it may not add much overhead to mirrors anyway.


Something I would like to add is database dumps ( cleaned from password
and private information, of course ), and maybe svn backup, stuff like
that. So people who wish to fork our project will not have the same
problem as we did.

One of the issue we currently have is the fact we couldn't fork Mandriva
wiki, because we didn't have access to the server. The same goes for
several components ( ideas, maintainers database, full bugzilla dump,
etc ). One of my goal would be to have a forkable infrastructure, so
people can replicate ours ( using puppet/cfengine, as we plan to use ).
( replicate to study, replicate it to play with it, replicate to enhance
it if needed )

And this would also provides use with a "linus t" backup system :p


> This mean everything must exists with this structure under the top level
> path.
> This way may avoid issues like mandriva on ibiblio (only 2005 and 2007.1
> seems to exists...)
> ftp://ibiblio.org/pub/linux/distributions/mandriva/Mandrivalinux/

I would let people choose what they mirror based on version. Ie, someone
could mirror the last version only, the 3 last, etc, etc.

This will likely provides enough flexibility without sacrifing too much
the simplicity.
-- 
Michael Scherer



Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread J.A. Magallón
On Thu, 21 Oct 2010 17:16:06 +0100, Buchan Milne  wrote:

> On Thursday, 21 October 2010 06:37:37 Olivier Thauvin wrote:
> > * J.A. Magallón (jamagal...@ono.com) wrote:
> > > On Wed, 20 Oct 2010 18:34:24 +0200, Olivier Thauvin 
...
> > 
> > > - could be the arch names more uniform ? in my personal scripts/setups
> > > 
> > >   I use x86-32 and x86-64.
> 
> Is x86-32 a valid architecture for rpm etc.? While uniformity might be nice, 
> unfortunately vendors don't necessarily choose uniform architecure names, and 
> it might be better to match the repo structure to values that can be 
> determined directly (and not heuristcally) . 
> 
> I've also never seen 'uname -m' report x86-32 or x86_32.
> 
> > >   Moreover, perhaps in a not so near future some
> > >   adventurous soul builds Mageia on ARM or Sparc, so why not sort things
> > >   like
> > >   
> > > distrib/cauldron/srpm
> > > distrib/cauldron/x86/32/iso
> > > 
> > >/rpm
> > > 
> > > /64
> > > 
> > > /arm/32
> > > 
> > > /64
> 
> I don't know if memory address space is a useful differentiator here, as 
> features differ substantially in different ARM cores of the same family or 
> architecture version. E.g., Fedora has an 'armv5tel' architecture, N900 ships 
> .deb's with 'armel' as the architecture. See 
> http://en.wikipedia.org/wiki/ARM_architecture
> 
> > > 
> > > /sparc/32
> > > 
> > >   /64
> 
> AFAIK, the valid architecture names for sparc are sparc,sparc64,sparcv9.
> 

I was just thinking of people navigating the tree structure looking for what
to download and install. Directory names have not to be the same as RPMs
architecture, isn't it ? And is more neutral, you can decide you build the
distro for Pentium3 at least and dont have to change the folder names... ;)

And, btw, what brain-damaged mind got the names i386 and x86_64 ? Solaris on
x86 uses i86pc (arch -k). Somebody should try to push upstream (in rpm, perhaps)
a decent naming scheme [x86,x86-32] - x86-64, [sparc,sparc-32] - sparc-64.
You said that fedora on ARM uses esoteric arch codes, couldn't be possible
to use x86-32 and try to get it into upstream RPM ?

Looking for that info (where that thing was born), I found that this naming
scheme (x86-32,x86-64) is even described in Wikipedia:

http://en.wikipedia.org/wiki/X86
http://en.wikipedia.org/wiki/X86_64

so it looks not so insane after all...

-- 
J.A. Magallon  \   Software is like sex:
 \ It's better when it's free


Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread Olivier Thauvin
* Michael Scherer (m...@zarb.org) wrote:
> Le mercredi 20 octobre 2010 à 18:34 +0200, Olivier Thauvin a écrit :
> > Hi,
> > 
> > You can find here:
> > http://distrib-coffee.ipsl.jussieu.fr/pub/linux/Mageia/
> > the current mirror tree proposal.
> > We now have to discuss it, I think.
> > The peoples/ directory is dedicated to contributors and must allow
> > anyone to share files related to the distribution (testing rpms,
> > preworks) in an unofficial ways.
> > How this can be setup in practice still have to be discuss.
> 
> Like quota, etc ?

Yes.

> 
> > Finally the software/ directory should allow you to distribute the
> > tarball (not rpm) for software we do. As soon we do free software people
> > must be able to distribute our code in the same way any projects does.
> > Replies like "svn is readable" or "we have rpm" is not appropriated.
> > 
> > Now come the question: "what is a valid mirror ?", eg, what a mirror
> > should have as file to be valid ?
> > 
> > I suggest to not give the choice and avoid mistake by saying except
> > "peoples" a mirror must respect the whole tree to be valid.
> 
> I would also exclude software/ from regular mirror, because this may not
> requires as much redundancy than regular mirror ( ie, I doubt many
> people will download tarball on a regular basis, and most softwre
> project are fine with 2 or 3 mirror ).
> 
> On the other hand, it may not add much overhead to mirrors anyway.

I like the idea to centralize everything from external point of view: eg
if you find a mageia mirror, you have everything.

> 
> 
> Something I would like to add is database dumps ( cleaned from password
> and private information, of course ), and maybe svn backup, stuff like
> that. So people who wish to fork our project will not have the same
> problem as we did.

I was thinking to push svn on mirror.

As potential problem, all these backup can add a lot of work and
bandwidth on mirrors. I am not against, I am in favor of this, but
just listing issues we can encoutered.

> And this would also provides use with a "linus t" backup system :p

BTW: the misc-backup.rpm is taking a lot of place :)=

> 
> 
> > This mean everything must exists with this structure under the top level
> > path.
> > This way may avoid issues like mandriva on ibiblio (only 2005 and 2007.1
> > seems to exists...)
> > ftp://ibiblio.org/pub/linux/distributions/mandriva/Mandrivalinux/
> 
> I would let people choose what they mirror based on version. Ie, someone
> could mirror the last version only, the 3 last, etc, etc.
> 
> This will likely provides enough flexibility without sacrifing too much
> the simplicity.

In fact we have no way to deny to someone to do a partial mirror. The
question is from our point of view, do we encourage people to create
non testable mirror or untrusted mirror (not update enough to ensure
last security update get sync).

It also make difficult the listing (eg having a gazillon of url for each
distrib, which is the current problem of easyurpmi).

-- 

Olivier Thauvin
CNRS  -  LATMOS
♖ ♘ ♗ ♕ ♔ ♗ ♘ ♖


pgptakxeoSUkW.pgp
Description: PGP signature


Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread Wolfgang Bornath
2010/10/22 Olivier Thauvin :
>
> In fact we have no way to deny to someone to do a partial mirror. The
> question is from our point of view, do we encourage people to create
> non testable mirror or untrusted mirror (not update enough to ensure
> last security update get sync).
>
> It also make difficult the listing (eg having a gazillon of url for each
> distrib, which is the current problem of easyurpmi).

I think this is an easy one. Do it like it was done at Mandriva.
Define what a "valid" mirror is, then list only "valid" mirrors.
Besides that people can put up mirrors in any way they want and use
them in any way they want. They are just not listed as "official
mirrors of Mageia".
Same happened with our Mandriva mirror. I wouldn't mirror the /debug
branches, so I was told that my mirror was not official and it was
taken from the list. Everybody was happy, the maintainers of the list,
me, and all people who used my mirror (it was not crowded, even at
release time). A solution suitable for everybody.

-- 
wobo


Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread Maarten Vanraes
Op donderdag 21 oktober 2010 22:38:19 schreef Olivier Thauvin:
> * Maarten Vanraes (maarten.vanr...@gmail.com) wrote:
> > Op woensdag 20 oktober 2010 18:34:24 schreef Olivier Thauvin:
> > > Hi,
> > > 
> > > You can find here:
> > > http://distrib-coffee.ipsl.jussieu.fr/pub/linux/Mageia/
> > > the current mirror tree proposal.
> > > We now have to discuss it, I think.
> > 
> > i didn't follow the whole previous discussion about the mirror tree, but
> > would it be advisable to have a noarch subdir, next to i586 and x86_64 ?
> 
> This is not on my side. The internal distribution tree structure is more
> related to build system than the global mirror tree structure.
> 
> At least, the tree I am suggesting neither deny it, neither make it
> easier.

That is true, however from a globabl mirror pov; it kind of makes sense:

suppose people would want to do the arm or other ports, the noarch packages 
can be exactly the same; therefor there is no need to list the noarch packages 
in i586 and having them symlinked or whatever to x86_64 which will have larger 
.cz and people will need to mirror both if they want x86_64 only, etc...



Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread andré

Wolfgang Bornath a écrit :

2010/10/22 Olivier Thauvin:
   

In fact we have no way to deny to someone to do a partial mirror. The
question is from our point of view, do we encourage people to create
non testable mirror or untrusted mirror (not update enough to ensure
last security update get sync).

It also make difficult the listing (eg having a gazillon of url for each
distrib, which is the current problem of easyurpmi).
 

I think this is an easy one. Do it like it was done at Mandriva.
Define what a "valid" mirror is, then list only "valid" mirrors.
Besides that people can put up mirrors in any way they want and use
them in any way they want. They are just not listed as "official
mirrors of Mageia".
Same happened with our Mandriva mirror. I wouldn't mirror the /debug
branches, so I was told that my mirror was not official and it was
taken from the list. Everybody was happy, the maintainers of the list,
me, and all people who used my mirror (it was not crowded, even at
release time). A solution suitable for everybody.
   


Excellent point.  As you say, a solution suitable for everyone.
Although for elements which are unlikely to be useful and take up a lot 
of space, it could still be useful to exclude them from the mirrors.


A variation of this could be, for mirrors constrained by disk space, to 
not keep as many older releases.  For example, suppose releases are 
every 6 months, supported for 3  years, for 6 supported releases.  Some 
mirrors might only keep 4 releases, or maybe 2, and still be considered 
official mirrors by Mageia.
This would be workable, as long as Mageia required a reasonable minimum 
of releases.

(I would say at least 2.)
Of course, Mageia would have to specify which mirrors only contained 
some supported releases.


my 2 cents :)
- André (andre999)


Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread andré

Romain d'Alverny a écrit :

On Wed, Oct 20, 2010 at 18:34, Olivier Thauvin
  wrote:
   

Now come the question: "what is a valid mirror ?", eg, what a mirror
should have as file to be valid ?
 

Not sure if we discussed in depth MirrorBrain
(http://www.mirrorbrain.org/ ) for managing mirrors index and
redirections.

If we were going to use it, could we, for instance, leave mirrors some
liberty to mirror what branch they want (with some guidances and
preferences of course) and let our MirrorBrain instance check and
build the list of valid mirrors for the file actually requested?

This, provided that _consistent_ branches of the tree are mirrored,
and not only a file here, a file there.

On one hand, this would introduce at least to other things to check:
  - having enough distributed mirrors that map the whole tree;
  - having download/install tools take this into account.

On the other hand, this could allow more mirrors to take part in this,
in that it may require less storage space and less bandwidth usage.
   

This is an excellent idea, at least to some degree.
Many users (like myself) like to download ISOs, which like the release 
repositories, will change every 6 months (or whatever the release period).
Otherwise they remain static, which means no maintenance for the mirror 
sites.

These elements tend to have relatively large space requirements.
As well, they would have a very large demand (bandwidth) at release 
time, to lessen considerably afterwards.


The update/backport/testing/debug repositories (or whatever Mageia will 
call them) are considerably smaller, but in constant flux, thus 
requiring regular updates on the mirror sites.
These transient repositories would tend to have a relatively constant, 
lower bandwidth demand.


So at least static/transient would be a natural separation.
As long as the division is by Mageia's repositories, it should be 
workable if mirrors select repositories on other bases.


I think it would be useful to keep the update/backport repositories 
together, since they are likely to be used together, and maybe all 
transient repositories.
Having the ISOs on separate mirrors should cause no problem, as users 
are likely to download an ISO by itself, one at a time.



It's not the only reason to use MirrorBrain anyway, but I wondered if
this could be a complementary reason.

Not sure, insights welcome.
   

After a quick look at the MirrorBrain features page, I say let's get it.

There is a problem with the current Mandriva mirror system, using Rpmdrake.
I'm in Montréal, and the nearest Mandriva mirror is not very reliable.  
It also allows only one direct download at a time.  So when I download 
ISOs, I go directly to another mirror, use multiple connexions, to 
download the ISO relatively fast, which is fine.


For updates and incremental downloads it is much easier to use Rpmdrake.
However, often, after having selected a package, I get an error message 
saying the package could not be downloaded because of a problem with the 
mirror site.

(I suspect that usually the site is just not in sync.)

MirrorBrain seems to be the perfect solution for this problem, as well 
as dealing with partial mirrors.


Interestingly, if for some reason a mirror site cannot carry a 
particular package because of restrictions imposed by the country where 
it is located, MirrorBrain will be able to cope.  So that simplifies the 
problem of dealing with legislative/copyright/patent restrictions as well.


So at least 3 reasons to get MirrorBrain :)

Cheers,

Romain
   


my 2 cents

- André (andre999)



Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread Ahmad Samir
On 21 October 2010 22:48, Olivier Thauvin  wrote:
> * Samuel Verschelde (sto...@laposte.net) wrote:
>> Le jeudi 21 octobre 2010 00:44:16, Olivier Thauvin a écrit :
>> > In a nutshell: this must be discuss. Having a people is at time just an
>> > idea, but it solve a need I had in mandriva to provide some rpms for
>> > testing purpose.
>>
>> I don't understand very well what kind of packages would go into "people" 
>> rather than in dedicated testing media. What would be the difference ?
>
> I can give the example I had on Mandriva: I did some work on 'rpm'
> package itself, but since breaking this pakage would have lock the whole
> Bs I'd prefer to first have it tested by some people.
> This directory is in this case the perfect place. At this time,
> "testing" media did not exists thought.
>
> Another example could be the XFCE live Cd done by some people in past,
> hosted on distrib-coffee as mandriva where not able to push it on thier
> mirror (this issue get solved latelly).
>
> Pushing something on the mageia Mirror need some control. The people
> place give the opportunity to contributors to distribute somethings w/o
> needing "super user" privileges, but with a clear "unofficial" state.
>
> I am pretty sure some people will find usage to it.
>
>>
>> Regards
>>
>> Samuel Verschelde
> --
>
> Olivier Thauvin
> CNRS  -  LATMOS
> ♖ ♘ ♗ ♕ ♔ ♗ ♘ ♖
>

Adding to what Nanar said, the way I see it, the stuff in people/ are
like the stuff we put in http://kenobi.mandriva.com/~contributor-name
sometimes; i.e. rpm packages to test (e.g. during cooker freeze before
asking for a push request on maintainers@)... etc. Just it'll be a bit
of a bigger place AND it won't put a load on the build clusters
themselves rather the load will be on mirrors, which is good IMHO.

-- 
Ahmad Samir


Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread herman
On Thu, 2010-10-21 at 17:07 -0700, Wolfgang Bornath wrote:
> 2010/10/22 Olivier Thauvin :
> >
> > In fact we have no way to deny to someone to do a partial mirror. The
> > question is from our point of view, do we encourage people to create
> > non testable mirror or untrusted mirror (not update enough to ensure
> > last security update get sync).
> Besides that people can put up mirrors in any way they want and use
> them in any way they want. They are just not listed as "official
> mirrors of Mageia".
> Everybody was happy, the maintainers of the list,
> me, and all people who used my mirror (it was not crowded, even at
> release time). A solution suitable for everybody.
So should we have an 'Official mirror list' and an 'Unofficial mirror
list'?  Because a mirror that is not listed at all, is rather hard to
find.
 




Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread Ahmad Samir
On 22 October 2010 08:11, herman  wrote:
> On Thu, 2010-10-21 at 17:07 -0700, Wolfgang Bornath wrote:
>> 2010/10/22 Olivier Thauvin :
>> >
>> > In fact we have no way to deny to someone to do a partial mirror. The
>> > question is from our point of view, do we encourage people to create
>> > non testable mirror or untrusted mirror (not update enough to ensure
>> > last security update get sync).
>> Besides that people can put up mirrors in any way they want and use
>> them in any way they want. They are just not listed as "official
>> mirrors of Mageia".
>> Everybody was happy, the maintainers of the list,
>> me, and all people who used my mirror (it was not crowded, even at
>> release time). A solution suitable for everybody.
> So should we have an 'Official mirror list' and an 'Unofficial mirror
> list'?  Because a mirror that is not listed at all, is rather hard to
> find.
>

I think wobo meant "local" mirrors, e.g. the mandrivauser.de mirror
was mainly used by German users who found out about it from
mandrivauser.de forums. (and having a list of "unofficial mirrors"
will be a pain, how would one differentiate "unofficial" mirrors that
just don't mirror some sub-trees, e.g. debug, and "unofficial" mirrors
that are just plain old/don't-get-synced-regularly).

-- 
Ahmad Samir


Re: [Mageia-dev] Mirror tree structure

2010-10-21 Thread Michael Scherer
Le vendredi 22 octobre 2010 à 00:26 +0200, J.A. Magallón a écrit :
> On Thu, 21 Oct 2010 17:16:06 +0100, Buchan Milne  
> wrote:
> 
> > On Thursday, 21 October 2010 06:37:37 Olivier Thauvin wrote:
> > > * J.A. Magallón (jamagal...@ono.com) wrote:
> > > > On Wed, 20 Oct 2010 18:34:24 +0200, Olivier Thauvin 
> ...
> > > 
> > > > - could be the arch names more uniform ? in my personal scripts/setups
> > > > 
> > > >   I use x86-32 and x86-64.
> > 
> > Is x86-32 a valid architecture for rpm etc.? While uniformity might be 
> > nice, 
> > unfortunately vendors don't necessarily choose uniform architecure names, 
> > and 
> > it might be better to match the repo structure to values that can be 
> > determined directly (and not heuristcally) . 
> > 
> > I've also never seen 'uname -m' report x86-32 or x86_32.
> > 
> > > >   Moreover, perhaps in a not so near future some
> > > >   adventurous soul builds Mageia on ARM or Sparc, so why not sort things
> > > >   like
> > > >   
> > > > distrib/cauldron/srpm
> > > > distrib/cauldron/x86/32/iso
> > > > 
> > > >/rpm
> > > > 
> > > > /64
> > > > 
> > > > /arm/32
> > > > 
> > > > /64
> > 
> > I don't know if memory address space is a useful differentiator here, as 
> > features differ substantially in different ARM cores of the same family or 
> > architecture version. E.g., Fedora has an 'armv5tel' architecture, N900 
> > ships 
> > .deb's with 'armel' as the architecture. See 
> > http://en.wikipedia.org/wiki/ARM_architecture
> > 
> > > > 
> > > > /sparc/32
> > > > 
> > > >   /64
> > 
> > AFAIK, the valid architecture names for sparc are sparc,sparc64,sparcv9.
> > 
> 
> I was just thinking of people navigating the tree structure looking for what
> to download and install. Directory names have not to be the same as RPMs
> architecture, isn't it ? And is more neutral, you can decide you build the
> distro for Pentium3 at least and dont have to change the folder names... ;)
> 

This will have a impact on the usability of the $ARCH keyword in url for
urpmi. Ie, the name must be derived from what we use for name of arch in
rpm, given the way this feature was coded.

Which in turn is arbitrary, iirc, but we may need to keep compatibility
with previous mandriva version, as we plan to offer a upgrade path.

-- 
Michael Scherer



Re: [Mageia-dev] Mirror tree structure

2010-10-22 Thread Remy CLOUARD
On Fri, Oct 22, 2010 at 02:56:04AM +0200, Maarten Vanraes wrote:
> Op donderdag 21 oktober 2010 22:38:19 schreef Olivier Thauvin:
> > * Maarten Vanraes (maarten.vanr...@gmail.com) wrote:
> > > Op woensdag 20 oktober 2010 18:34:24 schreef Olivier Thauvin:
> > > > Hi,
> > > > 
> > > > You can find here:
> > > > http://distrib-coffee.ipsl.jussieu.fr/pub/linux/Mageia/
> > > > the current mirror tree proposal.
> > > > We now have to discuss it, I think.
> > > 
> > > i didn't follow the whole previous discussion about the mirror tree, but
> > > would it be advisable to have a noarch subdir, next to i586 and x86_64 ?
> > 
> > This is not on my side. The internal distribution tree structure is more
> > related to build system than the global mirror tree structure.
> > 
> > At least, the tree I am suggesting neither deny it, neither make it
> > easier.
> 
> That is true, however from a globabl mirror pov; it kind of makes sense:
> 
> suppose people would want to do the arm or other ports, the noarch packages 
> can be exactly the same; therefor there is no need to list the noarch 
> packages 
> in i586 and having them symlinked or whatever to x86_64 which will have 
> larger 
> .cz and people will need to mirror both if they want x86_64 only, etc...
Actually, that’s exactly the point where it’s uncomfortable to have
noarch outside the arch tree.

If we only had i586 and x86_64 things would be quite simple, because
packages are built at the same time. Now consider more exotic arch: You
have to have machines that will compile for this arch. The problem is
that they might be lagging behind a bit if they’re not powerful enough,
or if we don’t have a sufficient number of machines.

If some noarch packages depend on arch dependant packages, and these
packages are not uptodate for arch $foo, that will lead to broken
packages in this arch.

Thus, having noarch packages outside the $foo tree is not feasible at
the moment.

Please also note that cross-compile is not an option either (but I’m not
an expert in this area to explain it in details).

Regards,
-- 
Rémy CLOUARD
() ascii ribbon campaign - against html e-mail
/\ www.asciiribbon.org - against proprietary attachments


Re: [Mageia-dev] Mirror tree structure

2010-10-22 Thread Remy CLOUARD
On Thu, Oct 21, 2010 at 10:55:32PM +0200, Renaud MICHEL wrote:
> On jeudi 21 octobre 2010 at 21:36, Remy CLOUARD wrote :
> > On Thu, Oct 21, 2010 at 12:44:16AM +0200, Olivier Thauvin wrote:
> > > I don't know any process to deny someone to push hdlists and rpms.
> > > However I don't think it is a good idea to use this space as an
> > > alternative to oficial mageia. The risk is to confuse people.
> > 
> > Well, I don’t think it’s about confusing people, after all these
> > packages are in a different subtree than regular packages, rather
> > giving choice. I know at least one person who said on #mageia-de that
> > mageia would be a no-go for him if he couldn’t have awesome.
> 
> awesome is already in mandriva, why wouldn't it be in mageia?
> If there is someone willing to package it, he may as well do it in the core 
> repository (be it main or contrib, or whatever they will be called).
This is off-topic, I was just taking awesome as an example.

Awesome in current Mandriva is outdated (v2.3.6) whereas 3.4.8 was
released. It’s just like comparing kde3 or kde4.

Regards,
-- 
Rémy CLOUARD
() ascii ribbon campaign - against html e-mail
/\ www.asciiribbon.org - against proprietary attachments


pgpt3nHbQkD0b.pgp
Description: PGP signature


Re: [Mageia-dev] Mirror tree structure

2010-10-22 Thread Wolfgang Bornath
There has always been the occasional user who needs just this special
application. Instead of telling the world that $DISTRO is not for him
if it does not serve this special app he can always ask a packager to
pack it.

Many communities do have a "Package Requests" section in their forums
for such occasions.

-- 
wobo