[gentoo-dev] [IMPORTANT] server/client USE flags

[gentoo-dev] [IMPORTANT] server/client USE flags

[Marius Mauch]

[-- Attachment #1: Type: text/plain, Size: 1213 bytes --]

Hi,

currently we have a lot of packages in the portage tree that provide
both server and client implementations of a package, big things like
mysql and samba but also small stuff like telnet-bsd. Often people don't
need the server part of these packages, but currently there is no way to
disable them. Some packages have splitted the package, but this is IMO
not a good idea in general, as often the ebuilds are the same except for
one or two commands.
So my idea (as well as others) is to add a new "server" USE flag that
controls wether the server part of a package is build. This idea is
outlined in more details on bug 12499, which also had the idea of a
"client" USE flag, so that it's possible to only build the libs or only
the server, however I think that such a flag will create more problems
than benefits (see the bug for explanations).
So, I'd like to get some opinions/comments from people on this idea, as
this is coming up about every month and there wasn't really any action
to solve this issue.

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] [IMPORTANT] server/client USE flags

[Andrew Gaffney]

Marius Mauch wrote:
> Hi,
> 
> currently we have a lot of packages in the portage tree that provide
> both server and client implementations of a package, big things like
> mysql and samba but also small stuff like telnet-bsd. Often people don't
> need the server part of these packages, but currently there is no way to
> disable them. Some packages have splitted the package, but this is IMO
> not a good idea in general, as often the ebuilds are the same except for
> one or two commands.
> So my idea (as well as others) is to add a new "server" USE flag that
> controls wether the server part of a package is build. This idea is
> outlined in more details on bug 12499, which also had the idea of a
> "client" USE flag, so that it's possible to only build the libs or only
> the server, however I think that such a flag will create more problems
> than benefits (see the bug for explanations).
> So, I'd like to get some opinions/comments from people on this idea, as
> this is coming up about every month and there wasn't really any action
> to solve this issue.

I think that it has the potential to work. The problem is that if someone were to emerge a 
package with USE="client -server" and then they decide they want the server part, they 
have to re-emerge the entire package. Not too big of an issue, but an issue nonetheless.

-- 
Andrew Gaffney


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[Philippe Coulonges]

Le Tue, 4 Nov 2003 00:20:17 +0100
Marius Mauch <genone@gentoo.org> écrivait :

> currently we have a lot of packages in the portage tree that provide
> both server and client implementations of a package, big things like
> mysql and samba but also small stuff like telnet-bsd. 

Humm, don't forget ntp.
I like to have a correct date on my machine, but for that, I dont need
CPU crunching to keep variation < 128µs, ntpdate
is enough.

CU
CPHIL
-- 
Article 19 
Everyone has the right to freedom of opinion and expression;
this right includes freedom to hold opinions without interference 
and to seek, receive and impart information and ideas through any media
and regardless of frontiers. 
	-- Universal Declaration of Human Rights - United Nations 1948

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[Troy Dack]

On Tue, 2003-11-04 at 10:20, Marius Mauch wrote:
> Hi,
> 
> currently we have a lot of packages in the portage tree that provide
> both server and client implementations of a package, big things like
> mysql and samba but also small stuff like telnet-bsd. Often people don't
> need the server part of these packages, but currently there is no way to
> disable them. Some packages have splitted the package, but this is IMO
> not a good idea in general, as often the ebuilds are the same except for
> one or two commands.
> So my idea (as well as others) is to add a new "server" USE flag that
> controls wether the server part of a package is build. This idea is
> outlined in more details on bug 12499, which also had the idea of a
> "client" USE flag, so that it's possible to only build the libs or only
> the server, however I think that such a flag will create more problems
> than benefits (see the bug for explanations).
> So, I'd like to get some opinions/comments from people on this idea, as
> this is coming up about every month and there wasn't really any action
> to solve this issue.
> 
> Marius

I like this idea, it makes sense and seems more logical than a client
USE flag.

I think that a *big fat notice* in /etc/make.conf should be included
just above the USE=... line though, something like:

# If you intend to use this machine as a server and require both
# client/server portions of packages like mysql or samba installed then
# ensure that you include  "server"  in your USE flags below.

Hopefully this will go some way to reducing all the "I emerged samba but
didn't get smbd/nmbd" bugs that are likely to come about.

A GWN or simply a News item as well as a post to gentoo-announce should
be made when this is implemented in the ebuilds.

-- 
Troy Dack					http://linux.tkdack.com
<troy@tkdack.com>				http://webportage.sf.net

Public Key: http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x4D90BE3C
Key fingerprint = 1F3D 6C15 16AA 09D5 0C96  92E5 FD89 16F9 4D90 BE3C


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[Donny Davies]

[-- Attachment #1: Type: text/plain, Size: 1339 bytes --]

Hi Marius

On Tue, Nov 04, 2003 at 12:20:17AM +0100, Marius Mauch wrote:
[...]
>So my idea (as well as others) is to add a new "server" USE flag that
>controls wether the server part of a package is build. This idea is
>outlined in more details on bug 12499, which also had the idea of a
>"client" USE flag, so that it's possible to only build the libs or only
>the server, however I think that such a flag will create more problems
>than benefits (see the bug for explanations).

I do not support this.  Debian's SAMBA has at least subpackages for
- common
- doc
- smbclient
- winbind
- swat
- smbfs
- python
- pam

which is nuts.  Users dont know what to install to get what they need.

I recall looking at mod_php (or maybe it was php) subpackages for another
distribution, and there were _dozens_ of them.  Insane in the membrane.

I like the way Gentoo traditionally works; you emerge something and you
get it all, with some knobs to control build-time dependencies.

Further, you are overloading the intended function of USE variables.
Instead of controlling optional build-time functionality, now you
are abusing them to control optional install-time bits.

It is not natural to stop at "client" and "server" flags either.
What about "dev" for .a and .h things?  This is really going down
the slipperly slope in my opinion.

Donny


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

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[Jason Rhinelander]

On Mon, 2003-11-03 at 16:17, Donny Davies wrote:
> Further, you are overloading the intended function of USE variables.
> Instead of controlling optional build-time functionality, now you
> are abusing them to control optional install-time bits.
> 
> It is not natural to stop at "client" and "server" flags either.
> What about "dev" for .a and .h things?  This is really going down
> the slipperly slope in my opinion.

Not to mention the fact that you could easily want the MySQL server &
client, but only the Samba client.  Does this mean you'd have to have
"mysqlserver", "sambaclient", "sambadoc", etc. USE flags?  Or would you
juggle around with your use flags for various packages, preventing you
from being able to emerge -u world?  This is actually a greater problem
- per-package USE flags would be wonderfully USEful (pardon the pun) in
many other situations.

The only logical way I can see to do this is splitting everything up
into multiple packages, i.e. mysql-server mysql-client, mysql-docs,
samba-server, samba-client, samba-docs, etc. - and I really dislike that
approach (I'm against the vim-core/vim/gvim split as well), for the
reasons mentioned above (and the fact that "updating portage cache"
after an emerge sync takes long enough as is without shattering packages
7 or 8 ways each).

> Donny
-- 
-- Jason Rhinelander
-- Gossamer Threads, Inc.


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[Marius Mauch]

[-- Attachment #1: Type: text/plain, Size: 2559 bytes --]

On 11/03/03  Donny Davies wrote:

> Hi Marius
> 
> On Tue, Nov 04, 2003 at 12:20:17AM +0100, Marius Mauch wrote:
> [...]
> >So my idea (as well as others) is to add a new "server" USE flag that
> >controls wether the server part of a package is build. This idea is
> >outlined in more details on bug 12499, which also had the idea of a
> >"client" USE flag, so that it's possible to only build the libs or
> >only the server, however I think that such a flag will create more
> >problems than benefits (see the bug for explanations).
> 
> I do not support this.  Debian's SAMBA has at least subpackages for
> - common
> - doc
> - smbclient
> - winbind
> - swat
> - smbfs
> - python
> - pam
> 
> which is nuts.  Users dont know what to install to get what they need.

That's why I'm against package splits.

> I recall looking at mod_php (or maybe it was php) subpackages for
> another distribution, and there were _dozens_ of them.  Insane in the
> membrane.

First I don't see how mod_php would be affected by this and second I
think the 30something USE flags for mod_php can be confusing, too
(alhough half of the problems just are the usual "why do I need X for
php" problem).

> I like the way Gentoo traditionally works; you emerge something and
> you get it all, with some knobs to control build-time dependencies.
> 
> Further, you are overloading the intended function of USE variables.
> Instead of controlling optional build-time functionality, now you
> are abusing them to control optional install-time bits.

If we add the "server" flag to the default flags it will still install
everything by default, if not we just have to make a big announcement
(like Troy said).
I don't see how I would abuse USE flags, often server support is just an
option for a configure script. And many of the local USE flags don't
follow this rule as well (btw, where can I read about this rule ?).

> It is not natural to stop at "client" and "server" flags either.
> What about "dev" for .a and .h things?  This is really going down
> the slipperly slope in my opinion.

Maybe it's not natural to stop there (btw, I'm only for a "server" flag,
not for a "client" flag), but it's reasonable. To make the "dev" stuff
optional doesn't make much sense for Gentoo as dependencies usually need
them, but usually there aren't dependencies on the server part of
something.

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] [IMPORTANT] server/client USE flags

[Luke-Jr]

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

On Tuesday 04 November 2003 12:17 am, Donny Davies wrote:
> What about "dev" for .a and .h things?  This is really going down
> the slipperly slope in my opinion.
In an OS where nearly everything is compiled, headers and static libs are 
neccesary to use a library (which would be providing them). Libraries without 
these files can only be used by binaries, not actually compiling source code.
- -- 
Luke-Jr
Developer, Gentoo Linux
http://www.gentoo.org/
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)

iD8DBQE/pvsEZl/BHdU+lYMRAuEZAJ9kRQxWtPvkOmfQWASPE4Oi3RF1qQCeO0WT
za/icBVogSwJLCJXCeWtoqc=
=3Gcl
-----END PGP SIGNATURE-----


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[Jason Wever]

[-- Attachment #1: Type: text/plain, Size: 1095 bytes --]

On Mon, 03 Nov 2003 17:55:50 -0600
Andrew Gaffney <agaffney@technaut.darktalker.net> wrote:

> I think that it has the potential to work. The problem is that if
> someone were to emerge a package with USE="client -server" and then they
> decide they want the server part, they have to re-emerge the entire
> package. Not too big of an issue, but an issue nonetheless.

This could also require additional dependency checking logic in portage.  

Example;

Package A has the proposed client/server capability.  I choose to build
package A as a client.  I then choose to build package B, which depends on
package A.  However, a library/header/etc normally provided by package A
isn't included with-server+client set in my USE variable.  

Portage would have to be able to check to see how package A was compiled
and force a re-compilation of package A to satisfy package B.  This also
would violate the USE flags I've setup in make.conf (don't know if this
happens already or not).

Just something to consider.  I'm not overly partial one way or another.

-- 
Jason Wever
Gentoo/Sparc Team Co-Lead

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

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[Marius Mauch]

[-- Attachment #1: Type: text/plain, Size: 2279 bytes --]

On 11/03/03  Jason Rhinelander wrote:

> On Mon, 2003-11-03 at 16:17, Donny Davies wrote:
> > Further, you are overloading the intended function of USE variables.
> > Instead of controlling optional build-time functionality, now you
> > are abusing them to control optional install-time bits.
> > 
> > It is not natural to stop at "client" and "server" flags either.
> > What about "dev" for .a and .h things?  This is really going down
> > the slipperly slope in my opinion.
> 
> Not to mention the fact that you could easily want the MySQL server &
> client, but only the Samba client.  Does this mean you'd have to have
> "mysqlserver", "sambaclient", "sambadoc", etc. USE flags?  Or would
> you juggle around with your use flags for various packages, preventing
> you from being able to emerge -u world?  This is actually a greater
> problem- per-package USE flags would be wonderfully USEful (pardon the
> pun) in many other situations.

That's bug 13616, to be included in portage-2.0.50.

> The only logical way I can see to do this is splitting everything up
> into multiple packages, i.e. mysql-server mysql-client, mysql-docs,
> samba-server, samba-client, samba-docs, etc. - and I really dislike
> that approach (I'm against the vim-core/vim/gvim split as well), for
> the reasons mentioned above (and the fact that "updating portage
> cache" after an emerge sync takes long enough as is without shattering
> packages 7 or 8 ways each).

There are several reasons why I'm against package splits:
- redundant code: most times the ebuilds for the subpackages will be
very similar
- maintenance: if a new version is out you have to update n packages
instead of one, if there is a new bug you to fix it in all packages
- space overhead: the portage tree has about 80% filesystem overhead for
most people, coming from a lot of small files (digests, Manifests,
metadata.xml). Each new package makes this worse.
- slower portage: the more packages we have the slower portage will be.
One package won't do much, but a few hundred new packages can make a
noticable difference.

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] [IMPORTANT] server/client USE flags

[Matthieu Sozeau]

Hi all, 

At Tue, 4 Nov 2003 02:18:15 +0100, Marius Mauch wrote:
> 
> [1  <text/plain; US-ASCII (7bit)>]
> On 11/03/03  Jason Rhinelander wrote:
> 
> > On Mon, 2003-11-03 at 16:17, Donny Davies wrote:
> > > Further, you are overloading the intended function of USE variables.
> > > Instead of controlling optional build-time functionality, now you
> > > are abusing them to control optional install-time bits.
> > > 
> > > It is not natural to stop at "client" and "server" flags either.
> > > What about "dev" for .a and .h things?  This is really going down
> > > the slipperly slope in my opinion.
> >
> > Not to mention the fact that you could easily want the MySQL server &
> > client, but only the Samba client.  Does this mean you'd have to have
> > "mysqlserver", "sambaclient", "sambadoc", etc. USE flags?  Or would
> > you juggle around with your use flags for various packages, preventing
> > you from being able to emerge -u world?  This is actually a greater
> > problem- per-package USE flags would be wonderfully USEful (pardon the
> > pun) in many other situations.
> 
> That's bug 13616, to be included in portage-2.0.50.
> 
> > The only logical way I can see to do this is splitting everything up
> > into multiple packages, i.e. mysql-server mysql-client, mysql-docs,
> > samba-server, samba-client, samba-docs, etc. - and I really dislike
> > that approach (I'm against the vim-core/vim/gvim split as well), for
> > the reasons mentioned above (and the fact that "updating portage
> > cache" after an emerge sync takes long enough as is without shattering
> > packages 7 or 8 ways each).
> 
> There are several reasons why I'm against package splits:
> - redundant code: most times the ebuilds for the subpackages will be
> very similar
> - maintenance: if a new version is out you have to update n packages
> instead of one, if there is a new bug you to fix it in all packages
> - space overhead: the portage tree has about 80% filesystem overhead for
> most people, coming from a lot of small files (digests, Manifests,
> metadata.xml). Each new package makes this worse.
> - slower portage: the more packages we have the slower portage will be.
> One package won't do much, but a few hundred new packages can make a
> noticable difference.

I do not think server or client (or doc in some way) fits well in USE
flags either, so i was thinking of another way to parametrize ebuilds.
Why not add a new variable for ebuilds that could be set when you
use a particular extra name. My idea is:

in samba-*.ebuild:

FLAVORS="client server"

src_compile () {
        flavor client && ...
}

Then if you emerge samba you get all flavors, if samba-1.2#client then
only client and common code etc...

There would be virtually 3 packages but only one file, which could avoid
much of the problems you mention. 

Just my 0.02 (euro) cents.
--
Matthieu Sozeau
www: http://mattam.ath.cx

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[Donnie Berkholz]

[-- Attachment #1: Type: text/plain, Size: 483 bytes --]

On Mon, 2003-11-03 at 20:17, Jason Wever wrote:
> Portage would have to be able to check to see how package A was compiled
> and force a re-compilation of package A to satisfy package B.  This also
> would violate the USE flags I've setup in make.conf (don't know if this
> happens already or not).
> 
> Just something to consider.  I'm not overly partial one way or another.

I believe this is on the list for portage .50 with things like
>=category/package-version[USE].

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

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[Matthew Kennedy]

Donny Davies <woodchip@gentoo.org> writes:

> Hi Marius
>

[...]

>
> It is not natural to stop at "client" and "server" flags either.
> What about "dev" for .a and .h things?  This is really going down
> the slipperly slope in my opinion.

Exactly.  What a maintenance nightmare.

Matt

-- 
Matthew Kennedy
Gentoo Linux Developer

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[Heiko Vogel]

[-- Attachment #1: Type: text/plain, Size: 1968 bytes --]

Donny Davies wrote:
> Hi Marius
> 
> On Tue, Nov 04, 2003 at 12:20:17AM +0100, Marius Mauch wrote:
> [...]
> >So my idea (as well as others) is to add a new "server" USE flag that
> >controls wether the server part of a package is build. This idea is
> >outlined in more details on bug 12499, which also had the idea of a
> >"client" USE flag, so that it's possible to only build the libs or only
> >the server, however I think that such a flag will create more problems
> >than benefits (see the bug for explanations).
> 
> I do not support this.  Debian's SAMBA has at least subpackages for
> - common
> - doc
> - smbclient
> - winbind
> - swat
> - smbfs
> - python
> - pam
> 
> which is nuts.  Users dont know what to install to get what they need.

Absolutely agreed -- I hate the Debian-way too!

> I recall looking at mod_php (or maybe it was php) subpackages for another
> distribution, and there were _dozens_ of them.  Insane in the membrane.
> 
> I like the way Gentoo traditionally works; you emerge something and you
> get it all, with some knobs to control build-time dependencies.

I share this opinion -- if a freak really doesn't like the
server/client parts of a package he should be cute enough to wipe it
out manually.
But a newbie who just wants Samba/MySQL to be installed will probably
trial-and-error for hours until he finds out that certain parts of the
package he wanted are distributed in other packages.

> Further, you are overloading the intended function of USE variables.
> Instead of controlling optional build-time functionality, now you
> are abusing them to control optional install-time bits.

Absolutely right ...

> It is not natural to stop at "client" and "server" flags either.
> What about "dev" for .a and .h things?  This is really going down
> the slipperly slope in my opinion.

Leave it as it is -- 
Do it the Gentoo way AND NOT the Debian/SuSE way!

greetz,
haaner

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

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[Mike Williams]

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

On Tuesday 04 November 2003 01:00, Marius Mauch wrote:

> > It is not natural to stop at "client" and "server" flags either.
> > What about "dev" for .a and .h things?  This is really going down
> > the slipperly slope in my opinion.
>
> Maybe it's not natural to stop there (btw, I'm only for a "server" flag,
> not for a "client" flag), but it's reasonable. To make the "dev" stuff
> optional doesn't make much sense for Gentoo as dependencies usually need
> them, but usually there aren't dependencies on the server part of
> something.

As just "another technical user" I feel the client/server flags would a killer 
feature.
I have several Gentoo desktops, some mildly slow, and a number of servers. On 
all of the desktops I was mysql support to talk to my servers. It has always 
annoyed me that I have to install the server too, it's such a pointless waste 
of time and space.

If the client is always built, and the server too with the default flags 
no-one would know the difference until they went and turned server support 
off.

- -- 
Mike Williams
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)

iD8DBQE/p4RxInuLMrk7bIwRAidsAJ0ZaUfEtG1M7dC/k3dQEvaWnwdrdACghK9i
yjcnp6zW14KKVHLPyyXb8jo=
=ES/J
-----END PGP SIGNATURE-----

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[William Hubbs]

Hi all,


On Tue, Nov 04, 2003 at 10:50:18AM +0000, Mike Williams wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> On Tuesday 04 November 2003 01:00, Marius Mauch wrote:
> 
> > > It is not natural to stop at "client" and "server" flags either.
> > > What about "dev" for .a and .h things?  This is really going down
> > > the slipperly slope in my opinion.
> >
> > Maybe it's not natural to stop there (btw, I'm only for a "server" flag,
> > not for a "client" flag), but it's reasonable. To make the "dev" stuff
> > optional doesn't make much sense for Gentoo as dependencies usually need
> > them, but usually there aren't dependencies on the server part of
> > something.
> 
> As just "another technical user" I feel the client/server flags would a killer 
> feature.
> I have several Gentoo desktops, some mildly slow, and a number of servers. On 
> all of the desktops I was mysql support to talk to my servers. It has always 
> annoyed me that I have to install the server too, it's such a pointless waste 
> of time and space.
> 
> If the client is always built, and the server too with the default flags 
> no-one would know the difference until they went and turned server support 
> off.

I agree with this.

If the "server" flag is in the defaults anyway, it will not affect anything unless server support is turned off.

Someone voiced a concern about someone coming up with a "dev" flag.  That isn't really a possible option in gentoo because everything is compiled anyway.

We also have one other use flag that I can think of that controls installation options instead of compile time options -- it is the doc use flag.

William


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[Bob Miller]

Marius Mauch wrote:

> So my idea (as well as others) is to add a new "server" USE flag that
> controls wether the server part of a package is build.

Two objections.

1. USE flags are global for the host.  "Serverness" is per-service,
   not per-host.  The host that is an apache server may be a mysql
   client.

2. Overridden USE flags get blown away by updates.  A given package
   is installed once, then upgraded numerous times over the life
   of a system.  It shouldn't need manual intervention every time.

I think Jason's idea, having package-local USE flags like
samba-server, ntp-client, etc. might make better sense.

The portage gods could generalize this so that USE flags have scope
and you could define something like USE="-server server:net-fs/samba"
for a desktop with shared filesystems. (actual syntax TBD)
That would be package-specific and persistent.

-- 
Bob Miller                              K<bob>
kbobsoft software consulting
http://kbobsoft.com                     kbob@jogger-egg.com

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[John Davis]

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

Bob Miller wrote:
| Marius Mauch wrote:
|
|
|>So my idea (as well as others) is to add a new "server" USE flag that
|>controls wether the server part of a package is build.
|
|
| Two objections.
|
| 1. USE flags are global for the host.  "Serverness" is per-service,
|    not per-host.  The host that is an apache server may be a mysql
|    client.
|
| 2. Overridden USE flags get blown away by updates.  A given package
|    is installed once, then upgraded numerous times over the life
|    of a system.  It shouldn't need manual intervention every time.
|
| I think Jason's idea, having package-local USE flags like
| samba-server, ntp-client, etc. might make better sense.
|
| The portage gods could generalize this so that USE flags have scope
| and you could define something like USE="-server server:net-fs/samba"
| for a desktop with shared filesystems. (actual syntax TBD)
| That would be package-specific and persistent.
|
To avoid over-complication, why don't we just separate the client and
server ebuilds completely? (eg, samba-server, samba-client, etc)? I
think woodchip pounced all over this one, but why avoid the most simple
and elegant solution?

Cheers,
//zhen
- --
John Davis
Gentoo Linux Developer
<http://dev.gentoo.org/~zhen>

- ----
Knowledge can be more terrible than ignorance if you're powerless to
change your world.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQE/p+6+ZlASNRlGLUcRAjlJAKCrw8XWHGuoqhFHEgCx3fjka81xvgCfQmAb
SzVYuFqsaHQsxgDs0HAl3wE=
=vSUJ
-----END PGP SIGNATURE-----


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[Jon Portnoy]

On Tue, Nov 04, 2003 at 01:23:58PM -0500, John Davis wrote:
> To avoid over-complication, why don't we just separate the client and
> server ebuilds completely? (eg, samba-server, samba-client, etc)? I
> think woodchip pounced all over this one, but why avoid the most simple
> and elegant solution?
> 

Because of all the previously-stated reasons as to why it's a bad idea.

-- 
Jon Portnoy
avenj/irc.freenode.net

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[Donny Davies]

[-- Attachment #1: Type: text/plain, Size: 778 bytes --]

Sorry about the messed up Reply-To on this one.  Slip of the D key.

On Tue, Nov 04, 2003 at 01:23:58PM -0500, John Davis wrote:
> To avoid over-complication, why don't we just separate the client and
> server ebuilds completely? (eg, samba-server, samba-client, etc)? I
> think woodchip pounced all over this one, but why avoid the most simple
> and elegant solution?

That's not going to happen.  Not as long as I am maintaining SAMBA :-)

Separate ebuilds is not clean.  Overloading USE flags is also way off
base; they are not meant to designate subpackages.

Unless subpackages are supported by Portage in some official fashion,
I'm not playing ball.  That's really what the people are asking for.

So, design subpackages support.  Then the ebuilds can be changed.

Donny


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

[gentoo-dev] Gentoo Rocks My Socks

[Stroller]

On Nov 4, 2003, at 10:50 am, Mike Williams wrote:
>>
>> Maybe it's not natural to stop there (btw, I'm only for a "server" 
>> flag,
>> not for a "client" flag), but it's reasonable....
>
> As just "another technical user" I feel the client/server flags would 
> a killer
> feature.

Mee, too! Combined with per-package USE flags..
   *joy*
   *oh!*
   *joy*
   *I*
   *can't*
   *wait*
   *to*
   *USE*
   *these*
<does a little dance>
...this would be just awesome.

Thanks for all the hard work, devs. I can' wait to see Portage 2.0.50, 
now.

Stroller.


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[Terje Kvernes]

Donny Davies <woodchip@gentoo.org> writes:

  [ ... ]

> Unless subpackages are supported by Portage in some official
> fashion, I'm not playing ball.  That's really what the people are
> asking for.

  subpackages would be both neat and practical.  one thing we have to
  think about though, are dependencies.  certain ebuilds might depend
  on just a part of an ebuild.  some applications might require a
  binary called "ssh" to be available, others need the whole nine
  yards.  

  but, if you depend on 'openssh+client'[1], which is a subpackage of
  openssh, having 'openssh' installed will do.

  also, portage will have to support virtual subpackages.  ;-) 
 
> So, design subpackages support.  Then the ebuilds can be changed.

  totally agreed.  USE-flags have noting to do with client /
  server-support.


  [1] I'm not sure how we'd like to denote subpackages, '-' and '_'
      are sort of taken.

-- 
Terje

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[Luke-Jr]

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

On Wednesday 05 November 2003 01:24 am, Terje Kvernes wrote:
>   [1] I'm not sure how we'd like to denote subpackages, '-' and '_'
>       are sort of taken.
emerge kde
emerge kde/kdenetwork
emerge kde/kdenetwork/kmail

I wouldn't expect anything fancy like that until Portage 2.0 though...
- -- 
Luke-Jr
Developer, Gentoo Linux
http://www.gentoo.org/
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)

iD8DBQE/qFTJZl/BHdU+lYMRAjuCAJ4xQ4MdlBJZOa8i/Xj275RFfZN1mwCfZniS
ysJlY+jpEzq06+S7vOix+xE=
=uFZB
-----END PGP SIGNATURE-----


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[Chris Smith]

On Wednesday 05 November 2003 14:39, Luke-Jr wrote:
> emerge kde
> emerge kde/kdenetwork
> emerge kde/kdenetwork/kmail
>
> I wouldn't expect anything fancy like that until Portage 2.0 though...

Isn't that already the "catagory/package" delimeter? Could maybe cause some 
problems. I think the + sign could work well.

Also, isn't Portage 2.0 already out :D Ya must mean 2.1 :)

Cheers,
Chris.
-- 
Failure is not an option. It comes bundled with your Microsoft product.


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[Terje Kvernes]

Luke-Jr <luke-jr@gentoo.org> writes:

> On Wednesday 05 November 2003 01:24 am, Terje Kvernes wrote:
> >   [1] I'm not sure how we'd like to denote subpackages, '-' and '_'
> >       are sort of taken.
>
> emerge kde
> emerge kde/kdenetwork
> emerge kde/kdenetwork/kmail

  hm, could we ever get conflict between a subpackage and a catagory?
  or are we more looking to say that conceptually we're looking at the
  same ting?  ;-)

> I wouldn't expect anything fancy like that until Portage 2.0
> though...

  3.0?

-- 
Terje

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[Luke-Jr]

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

On Wednesday 05 November 2003 02:01 am, Chris Smith wrote:
> On Wednesday 05 November 2003 14:39, Luke-Jr wrote:
> > emerge kde
> > emerge kde/kdenetwork
> > emerge kde/kdenetwork/kmail
> >
> > I wouldn't expect anything fancy like that until Portage 2.0 though...
>
> Isn't that already the "catagory/package" delimeter? Could maybe cause some
> problems. I think the + sign could work well.
That'd replace categories...
>
> Also, isn't Portage 2.0 already out :D Ya must mean 2.1 :)
Bleh.. whatever portage-ng is =p
- -- 
Luke-Jr
Developer, Gentoo Linux
http://www.gentoo.org/
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)

iD8DBQE/qGGaZl/BHdU+lYMRAsUPAJ9VW6kXB1dCEp2o9bAYFi59pRrFRwCfUmZk
9fPXTK4uhqTzar6OAcJpchg=
=tczZ
-----END PGP SIGNATURE-----


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[Paul de Vrieze]

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

On Wednesday 05 November 2003 03:23, Terje Kvernes wrote:
> Luke-Jr <luke-jr@gentoo.org> writes:
> > On Wednesday 05 November 2003 01:24 am, Terje Kvernes wrote:
> > >   [1] I'm not sure how we'd like to denote subpackages, '-' and
> > > '_' are sort of taken.
> >
> > emerge kde
> > emerge kde/kdenetwork
> > emerge kde/kdenetwork/kmail
>
>   hm, could we ever get conflict between a subpackage and a catagory?
>   or are we more looking to say that conceptually we're looking at the
>   same ting?  ;-)

There are enough symbols on the keyboard to have a good delimiter. 
However what is needed too is a concept on how to do it. Are subpackages 
compiletime or installtime. Or do we allow both?

For example if you want kmail only from kdenetwork many parts of 
kdenetwork can be left alone (some support libs not), and don't even 
need to be compiled. In that case it would be possible to have dependant 
packages specify which subpackages they need, and have them added 
on-demand if needed (be aware of compiletimes).

Installtime is also possible, where packages are fully compiled but 
partially installed. If a binary package exists it then is possible to 
do ondemand installation of subpackages without recompiles being 
necessary. However with installtime subpackages there is the problem of 
identifying which files should be part of the subpackages.

This does not mean that I am principally against subpackages, but that 
subpackages are indeed complex and probably need to wait for portage-ng 
for support.

Paul

- -- 
Paul de Vrieze
Gentoo Developer
Mail: pauldv@gentoo.org
Homepage: http://www.devrieze.net
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)

iD8DBQE/qMbfbKx5DBjWFdsRAo+AAKCRxnv1bBiAdZ5JmN06AOKBDzeW/QCgzZvJ
vdHsUO1U6hQbBjituNnbN4Y=
=feWv
-----END PGP SIGNATURE-----


--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[Patrick Kursawe]

[-- Attachment #1: Type: text/plain, Size: 331 bytes --]

On Wed, Nov 05, 2003 at 10:46:07AM +0100, Paul de Vrieze wrote:

> This does not mean that I am principally against subpackages, but that 
> subpackages are indeed complex and probably need to wait for portage-ng 
> for support.

Having a look at bugzilla, I don't think this should be our top
priority :-)

Bye, Patrick

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

Re: [gentoo-dev] [IMPORTANT] server/client USE flags

[C. Brewer]

[-- Attachment #1: signed data --]
[-- Type: text/plain, Size: 458 bytes --]

On Wednesday 05 November 2003 1:46, Paul de Vrieze wrote:

> This does not mean that I am principally against subpackages, but that
> subpackages are indeed complex and probably need to wait for portage-ng
> for support.
>
> Paul

Won't this make dep/rev-dep checking an even more unlikely feature with 
extended subpackages?
-- 
Chuck Brewer
Registered Linux User #284015
Get my gpg public key at pgp.mit.edu!! Encrypted e-mail preferred.



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