[gentoo-dev] Some important questions to the officals of www.gentoo.org

[gentoo-dev] Some important questions to the officals of www.gentoo.org

[Alexander Mieland]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello gentoo-team,

I've searched for an official email-adress to send my questions to, but I 
could not find one. So sorry for sending this over the list.

Also sorry because my terrible english. :)

We are two young (expose(16), dma147(32)) german guys who wants to 
reanimate a project, which was on gentoo.org and which is already dead 
since over a year.
We have already started with it und we have enough ideas and knowledge to 
finish it.

Now we want to ask, if we can have a subdomain of gentoo.org to publish 
it, when it is finished. The old, dead project had a subdomain, but it's 
down.

you will become much more Infos about the project, if you're an official 
of gentoo.org and you'll mail us.
Please answer to: dma147@mieland-programming.de

So we could be sure, that we are the only ones who work on that project.

So please email us and we'll explain, what we are doing and want to do.

Greetings and much thanks. :)

- -- 
Alexander Mieland                      System: Gentoo 2004.1, 2.6.7 SMP
APP - Another PHP Program                 Registered Linux-User #249600
http://www.php-programs.de                        http://counter.li.org
http://www.mieland-programming.de            s/Alexander Mieland/dma147
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFA8EWkkxKnzj3lcIERAh2QAJ4qXTnS/FugO3Pnfw5+KtKCZzUjuwCfb2ji
90ZB8tj/L9JA+PPHih8+gTU=
=nBQo
-----END PGP SIGNATURE-----

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Some important questions to the officals of www.gentoo.org

[Mike Frysinger]

On Saturday 10 July 2004 03:38 pm, Alexander Mieland wrote:
> I've searched for an official email-adress to send my questions to, but I
> could not find one. So sorry for sending this over the list.

development related activies is gentoo-dev

> We are two young (expose(16), dma147(32)) german guys who wants to
> reanimate a project, which was on gentoo.org and which is already dead
> since over a year.
> We have already started with it und we have enough ideas and knowledge to
> finish it.
>
> Now we want to ask, if we can have a subdomain of gentoo.org to publish
> it, when it is finished. The old, dead project had a subdomain, but it's
> down.

as far as i can tell, you neglect to mention anything about 'the project' or 
what you could possibly be refering to
-mike

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Some important questions to the officals of www.gentoo.org

[Grant Goodyear]

[-- Attachment #1: Type: text/plain, Size: 886 bytes --]

> Now we want to ask, if we can have a subdomain of gentoo.org to publish 
> it, when it is finished. The old, dead project had a subdomain, but it's 
> down.

Um, why all the secrecy?  

You'll want to talk to somebody on the infrastructure team, but I think
I can fairly say that the answer will be "no", unless (a) you are
willing to give the Gentoo infrastructure team access to your server,
and (b) the server satisfies the Gentoo infrastructure requirements.  On
the other hand, if there's a clear consensus that your project is worth
having (and you're willing to maintain it), then I'm sure we would be
happy to host your project on one of our servers and provide you
with access to it.

-g2boojum-
-- 
Grant Goodyear	
Gentoo Developer
g2boojum@gentoo.org
http://www.gentoo.org/~g2boojum
GPG Fingerprint: D706 9802 1663 DEF5 81B0  9573 A6DC 7152 E0F6 5B76

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

Re: [gentoo-dev] Some important questions to the officals of www.gentoo.org

[Alexander Mieland]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Am Sonntag, 11. Juli 2004 00:05 schrieb Grant Goodyear:
> > Now we want to ask, if we can have a subdomain of gentoo.org to
> > publish it, when it is finished. The old, dead project had a
> > subdomain, but it's down.
>
> Um, why all the secrecy?

that's because I've made enough experience with people who steals ideas.

> You'll want to talk to somebody on the infrastructure team, but I think
> I can fairly say that the answer will be "no", unless (a) you are
> willing to give the Gentoo infrastructure team access to your server,

what do you mean with "access"? SSH or FTP? Or simply an uri to the 
project to look at it?
At this time, we are planning on it. We don't have something you could 
look at.  

> and (b) the server satisfies the Gentoo infrastructure requirements. 
> On the other hand, if there's a clear consensus that your project is
> worth having (and you're willing to maintain it), then I'm sure we
> would be happy to host your project on one of our servers and provide
> you with access to it.

Okay, then I'll explain, what we want to do and what we have done until 
now.

Daniel and me, Alex, got the idea of a database which contains merge times   
of all currently available (unmasked) gentoo packages for x86 
architecture.
To realize this, we'll have a dedicated machine for just emerging all 
packages (and which is just now merging...*g*).
The time it takes on this machine will be the standard and will be 
multiplied with the value calculated by a client, which has information 
about the users hardware. This way we will be able to give the people a 
clue of how long it would take to emerge certain packages on their own 
machine.

With this idea and also the idea behind the two projects on 
stats.gentoo.org and www.mksoft.co.il/gentoo-stats/, we want to provide a 
service which is IMHO very informative and usefull.

What we have until now is surely the idea and the knowledge to code the 
needed stuff (client-side-script and web-interface). But we don't want to 
jump into the cold water, coding all the stuff without any official 
support from gentoo.org, because that makes no sense.

So, that's the actual state.
if we'll be allowed to use the official design of gentoo.org and, perhabs 
if we'll get a subdomain like state.gentoo.org, we would just begin to 
realize this project.

okay, now it's your turn. *g*

Yours sincerely,

- -- 
Alexander Mieland                      System: Gentoo 2004.1, 2.6.7 SMP
APP - Another PHP Program                 Registered Linux-User #249600
http://www.php-programs.de                        http://counter.li.org
http://www.mieland-programming.de            s/Alexander Mieland/dma147
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFA8HAHkxKnzj3lcIERAnQ2AKC1V2BeQTUVndxHPmxySAB1xQpaWACfdfJj
zB7I0Fx08z41wMoOwfpUC70=
=VYpS
-----END PGP SIGNATURE-----

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Some important questions to the officals of www.gentoo.org

[Ciaran McCreesh]

[-- Attachment #1: Type: text/plain, Size: 677 bytes --]

On Sun, 11 Jul 2004 00:39:03 +0200 Alexander Mieland
<dma147@mieland-programming.de> wrote:
| Daniel and me, Alex, got the idea of a database which contains merge
| times   of all currently available (unmasked) gentoo packages for x86 
| architecture.

...pretty useless, really, unless you plan to provide times for both SMP
and non-SMP kit. Otherwise you'll get *horridly* skewed results for
packages which can't handle -j>1.

Heck, even distcc is enough to upset this beyond the point of usability.

-- 
Ciaran McCreesh : Gentoo Developer (Sparc, MIPS, Vim, Fluxbox)
Mail            : ciaranm at gentoo.org
Web             : http://dev.gentoo.org/~ciaranm


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

Re: [gentoo-dev] Some important questions to the officals of www.gentoo.org

[Alexander Mieland]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Am Sonntag, 11. Juli 2004 00:44 schrieb Ciaran McCreesh:
> On Sun, 11 Jul 2004 00:39:03 +0200 Alexander Mieland
> <dma147@mieland-programming.de> wrote:
> | Daniel and me, Alex, got the idea of a database which contains merge
> | times   of all currently available (unmasked) gentoo packages for x86
> | architecture.
>
> ...pretty useless, really, unless you plan to provide times for both
> SMP and non-SMP kit. Otherwise you'll get *horridly* skewed results for
> packages which can't handle -j>1.
>
> Heck, even distcc is enough to upset this beyond the point of
> usability.

ACK.
it's surely in our mind to provide as much information as we can. I'm also 
planning on a second dedicated machine, which has a Pentium 4 with 
HT-technology to provide the SMP-part of the merge-times. In the future, 
if all is running and the response of the users is good enough, we hope 
that we can also provide the merge-times for other architectures than 
x86.

By the way, as long as we dont have a second machine with smp-support, it 
should be possibly to get reasonably accurate results if we multiply the 
number of cpus and then divide through a factor like 1.2. You can test 
it, it should be reasonably accurate.

But as I said, we are planning on a second machine with smp and also on 
merge-time information for other architectures.

(please overlook my horrible grammars)

Yours sincerely,

- -- 
Alexander Mieland                      System: Gentoo 2004.1, 2.6.7 SMP
APP - Another PHP Program                 Registered Linux-User #249600
http://www.php-programs.de                        http://counter.li.org
http://www.mieland-programming.de            s/Alexander Mieland/dma147
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFA8HblkxKnzj3lcIERAr0QAJ0TMsiXzDYp6TdZjDP9UaxDbUmp9gCePRH3
xNvjEIhIbfXQeJwkVtfdWw4=
=eN25
-----END PGP SIGNATURE-----

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Some important questions to the officals of www.gentoo.org

[Chris Gianelloni]

On Sat, 2004-07-10 at 19:08, Alexander Mieland wrote:
> ACK.
> it's surely in our mind to provide as much information as we can. I'm also 
> planning on a second dedicated machine, which has a Pentium 4 with 
> HT-technology to provide the SMP-part of the merge-times. In the future, 
> if all is running and the response of the users is good enough, we hope 
> that we can also provide the merge-times for other architectures than 
> x86.

Trust me when I tell you that a P4 w/ HT is *nowhere* near the speed of
true SMP.  You're better off going with your arbitrary calculations
below.

While I do think this information is useful, I still think the best
method is to come up with a new number, rather than time.  A good one
(stolen from LFS) is the BU or Bash Unit.  The time it takes to compile
bash is always 1BU.

SMP is ignored, since it skews the results.  So would distcc... The idea
is not to even try to give the exact time for emerging things for
everybody, but rather to have a strong baseline "average" that hits most
people.

> By the way, as long as we dont have a second machine with smp-support, it 
> should be possibly to get reasonably accurate results if we multiply the 
> number of cpus and then divide through a factor like 1.2. You can test 
> it, it should be reasonably accurate.
> 
> But as I said, we are planning on a second machine with smp and also on 
> merge-time information for other architectures.

-- 
Chris Gianelloni
Release Engineering QA Manager/Games Developer
Gentoo Linux

Is your power animal a pengiun?


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Some important questions to the officals of www.gentoo.org

[Alexander Mieland]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Am Sonntag, 11. Juli 2004 01:29 schrieb Chris Gianelloni:
> On Sat, 2004-07-10 at 19:08, Alexander Mieland wrote:
> > ACK.
> > it's surely in our mind to provide as much information as we can. I'm
> > also planning on a second dedicated machine, which has a Pentium 4
> > with HT-technology to provide the SMP-part of the merge-times. In the
> > future, if all is running and the response of the users is good
> > enough, we hope that we can also provide the merge-times for other
> > architectures than x86.
>
> Trust me when I tell you that a P4 w/ HT is *nowhere* near the speed of
> true SMP.  You're better off going with your arbitrary calculations
> below.
>
> While I do think this information is useful, I still think the best
> method is to come up with a new number, rather than time.  A good one
> (stolen from LFS) is the BU or Bash Unit.  The time it takes to compile
> bash is always 1BU.
>
> SMP is ignored, since it skews the results.  So would distcc... The
> idea is not to even try to give the exact time for emerging things for
> everybody, but rather to have a strong baseline "average" that hits
> most people.

sorry, I'll answer tomorrow. It's now 01:40am. :)

But that's a fantastic idea. why isn't it mine? *g*
I've also used lfs, a few years ago.


> > By the way, as long as we dont have a second machine with
> > smp-support, it should be possibly to get reasonably accurate results
> > if we multiply the number of cpus and then divide through a factor
> > like 1.2. You can test it, it should be reasonably accurate.
> >
> > But as I said, we are planning on a second machine with smp and also
> > on merge-time information for other architectures.

- -- 
Alexander Mieland                      System: Gentoo 2004.1, 2.6.7 SMP
APP - Another PHP Program                 Registered Linux-User #249600
http://www.php-programs.de                        http://counter.li.org
http://www.mieland-programming.de            s/Alexander Mieland/dma147
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFA8H64kxKnzj3lcIERAui6AJ9ecN6EJr8Uk+qjY5xpnHAJOnS2igCfav9w
UvZdBjHfpA+Tnfpe661Nr2Q=
=c23c
-----END PGP SIGNATURE-----

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Some important questions to the officals of www.gentoo.org

[Ciaran McCreesh]

[-- Attachment #1: Type: text/plain, Size: 855 bytes --]

On Sun, 11 Jul 2004 01:08:20 +0200 Alexander Mieland
<dma147@mieland-programming.de> wrote:
| By the way, as long as we dont have a second machine with smp-support,
| it should be possibly to get reasonably accurate results if we
| multiply the number of cpus and then divide through a factor like 1.2.
| You can test it, it should be reasonably accurate.

Sadly, not the case. See, ./configure doesn't parallelise *at all*, and
some packages either override make to not run in parallel (because the
ebuild maintainer is too lazy to fix buggy upstream makefiles) or
because the source files have lots of dependencies. So, it depends
pretty much entirely upon the package as to how SMP affects it.

-- 
Ciaran McCreesh : Gentoo Developer (Sparc, MIPS, Vim, Fluxbox)
Mail            : ciaranm at gentoo.org
Web             : http://dev.gentoo.org/~ciaranm


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

Re: [gentoo-dev] Some important questions to the officals of www.gentoo.org

[marduk]

On Sun, 2004-07-11 at 00:39 +0200, Alexander Mieland wrote:

> that's because I've made enough experience with people who steals ideas.

But this is Linux/free software/whatever you call it.  We're supposed to
be about open exchange of ideas, not trade secrets.

> At this time, we are planning on it. We don't have something you could 
> look at.  

Okay, this is where I really think you got it wrong.  My understanding
is what really turns people on in the Linux/free software/whatever you
call it community is the product, not the idea.  You seem to be thinking
like a corporate person.  I know of a few projects in the free software
community that failed because a few persons had a good idea, but failed
to gain outside support because they didn't have the product to back it
up.  Put your code where your mouth is and if the infrastructure group
finds it worthy and appropriate then it might become a .gentoo.org.

But still I think you're going about it the wrong way.  If you really
want to do something useful then just do it.  It shouldn't matter
whether it's .gentoo.org worthy or not.  When I started writing
packages.gentoo.org (fresh ebuilds) it was originally just to satisfy a
personal itch.  I didn't see any way to tell when a new package was
added to portage.  So originally it just ran on my machine.  I had no
intention of creating an "official" gentoo.org site.  But then I thought
others "might" find it useful so I put it up on my web site.  Then it
became a .gentoo.org site. 

You seem to be going the opposite direction.  You want the site first
and then you start coding.  But I think you miss the idea.

As far as the idea itself.  I think it's been mentioned many many times
before why these things aren't accurate.  If one were to implement
something like this, I'd think they shouldn't use "real" numbers because
they will almost always "just plain wrong".  I think an arbitrary
coefficient would be more accurate, though still not precise.  But make
up a value ala BogoMips, call it Gentoo Hours or something.  Make the
time it takes to compile glibc (version x) to be 1 Gentoo Hour.  And the
time it takes to compile any other program will be a factor of that base
Gentoo Hour.  That "might" be more accurate but is still "just plain
wrong" ;-).

--m




--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Some important questions to the officals of www.gentoo.org

[Mike Frysinger]

On Saturday 10 July 2004 06:39 pm, Alexander Mieland wrote:
> Daniel and me, Alex, got the idea of a database which contains merge times
>   of all currently available (unmasked) gentoo packages for x86
> architecture.

another useful thing you could do is stick all the CONTENTS files into a 
database and put a front end on it ...
then users can search it and ask the question 'what package provides this 
file' ... granted the results wont be 100% (different USE/arch/etc...) but 
it's better than nothing
-mike

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Some important questions to the officals of www.gentoo.org

[Joseph Booker]

On Sat, July 10, 2004 7:37 pm, Mike Frysinger said:
> On Saturday 10 July 2004 06:39 pm, Alexander Mieland wrote:
>> Daniel and me, Alex, got the idea of a database which contains merge
>> times
>>   of all currently available (unmasked) gentoo packages for x86
>> architecture.
>
> another useful thing you could do is stick all the CONTENTS files into a
> database and put a front end on it ...
> then users can search it and ask the question 'what package provides this
> file' ... granted the results wont be 100% (different USE/arch/etc...) but
> it's better than nothing
> -mike
>
> --
> gentoo-dev@gentoo.org mailing list
>
>

Better yet, use loops with equery to find packages installing into the
same files to prevent this:

[19:44][pts/0][joe@Athena ~] % equery belongs /etc/mailcap
Searching for file '/etc/mailcap' in *...
mail-client/pine-4.60 (/etc/mailcap)
net-mail/mailbase-0.00-r6 (/etc/mailcap)
net-p2p/bittorrent-3.4.2 (/etc/mailcap)


-- 
Joseph Booker
joe @ irc.neoturbine.net
jj110888 @ irc.freenode.net


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Some important questions to the officals of www.gentoo.org

[Marius Mauch]

[-- Attachment #1: Type: text/plain, Size: 717 bytes --]

On 07/10/04  Joseph Booker wrote:

> Better yet, use loops with equery to find packages installing into the
> same files to prevent this:
> 
> [19:44][pts/0][joe@Athena ~] % equery belongs /etc/mailcap
> Searching for file '/etc/mailcap' in *...
> mail-client/pine-4.60 (/etc/mailcap)
> net-mail/mailbase-0.00-r6 (/etc/mailcap)
> net-p2p/bittorrent-3.4.2 (/etc/mailcap)

Well, those packages will hopefully vanish over time when people start
using the collision-protect FEATURE in portage-2.0.51 (only in >=
_pre13).

Marius

-- 
Public Key at http://www.genone.de/info/gpg-key.pub

In the beginning, there was nothing. And God said, 'Let there be
Light.' And there was still nothing, but you could see a bit better.

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

Re: [gentoo-dev] Some important questions to the officals of www.gentoo.org

[Aron Griffis]

[-- Attachment #1: Type: text/plain, Size: 590 bytes --]

I can't speak for the merit of your idea or implementation, but here's
an idea for you...  portage should keep track of how long it has taken
in the past to merge various packages on the current machine and use
that as an indicator for how long it will take to build the current
package.  This cache could be preloaded with guesstimated numbers if
you want, but the nice thing is that it would be customized to the
local machine's typical performance on a per-package basis, and
wouldn't require continued network access to function.

Regards,
Aron

--
Aron Griffis
Gentoo Linux Developer


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

Re: [gentoo-dev] Some important questions to the officals of www.gentoo.org

[Don Seiler]

[-- Attachment #1: Type: text/plain, Size: 852 bytes --]

emerge splat
splat -s <package>

Don.

On Mon, Jul 12, 2004 at 04:25:49PM -0400, Aron Griffis wrote:
> I can't speak for the merit of your idea or implementation, but here's
> an idea for you...  portage should keep track of how long it has taken
> in the past to merge various packages on the current machine and use
> that as an indicator for how long it will take to build the current
> package.  This cache could be preloaded with guesstimated numbers if
> you want, but the nice thing is that it would be customized to the
> local machine's typical performance on a per-package basis, and
> wouldn't require continued network access to function.
> 
> Regards,
> Aron
> 
> --
> Aron Griffis
> Gentoo Linux Developer
> 



-- 
Don Seiler             rizzo@gentoo.org
Gentoo Linux Developer
http://dev.gentoo.org/~rizzo/

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