Mailing List Archive

moving USE=server to global
On Saturday 17 March 2007, Steve Long wrote:
> Piotr Jaroszy?ski wrote:
> > I have heard about the magic limit of 5, but whatever...
>
> Is there a *technical* objection then to server?

if you're going to change the topic mid-thread, then you should update the
subject

you already know the state of the server/client debate ... moving it to global
doesnt fix any of the short comings, so it should stay local (and removed
where possible)
-mike
Re: moving USE=server to global [ In reply to ]
Mike Frysinger wrote:
> On Saturday 17 March 2007, Steve Long wrote:
>
>> Piotr Jaroszy?ski wrote:
>>
>>> I have heard about the magic limit of 5, but whatever...
>>>
>> Is there a *technical* objection then to server?
>>
>
> if you're going to change the topic mid-thread, then you should update the
> subject
>
> you already know the state of the server/client debate ... moving it to global
> doesnt fix any of the short comings, so it should stay local (and removed
> where possible)
> -mike
>
Yeah -- "server" is way too generic. I've forgotten where else I use it,
but when I build "vnc" I use it to get a VNC server. Maybe make a local
"vnc-server" USE flag for that one.

--
M. Edward (Ed) Borasky, FBG, AB, PTA, PGS, MS, MNLP, NST, ACMC(P)
http://borasky-research.blogspot.com/

If God had meant for carrots to be eaten cooked, He would have given rabbits fire.

--
gentoo-dev@gentoo.org mailing list
Re: Re: moving USE=server to global [ In reply to ]
Steve Long wrote:
> I don't see what is so dangerous about a server flag. After all I
> don't set doc globally, but it is a useful global flag, with clear
> intent, as would be server.

Does it mean:

A) "Compile just the server, as leaving it off compiles both client
AND server"

B) "Compile the client AND server, as leaving it off compiles ONLY the
client"

C) All of the above

--
Jim Ramsay
Gentoo/Linux Developer (rox,gkrellm)
Re: Re: Re: moving USE=server to global [ In reply to ]
On Mon, 2007-03-19 at 17:19 +0000, Steve Long wrote:
> Jim Ramsay wrote:
> > Steve Long wrote:
> >> I don't see what is so dangerous about a server flag. After all I
> >> don't set doc globally, but it is a useful global flag, with clear
> >> intent, as would be server.
> >
> > Does it mean:
> >
> > A) "Compile just the server, as leaving it off compiles both client
> > AND server"
> >
> > B) "Compile the client AND server, as leaving it off compiles ONLY the
> > client"
> >
> > C) All of the above
> >
> B, as asked for by many usrs. Does that cause other issues I am simply not
> seeing?

Do me a favor, please. Fix your "e" key. They're users, not usrs.

Thanks,

--
Chris Gianelloni
Release Engineering Strategic Lead
Alpha/AMD64/x86 Architecture Teams
Games Developer/Council Member/Foundation Trustee
Gentoo Foundation
Re: Re: moving USE=server to global [ In reply to ]
Steve Long wrote:
>> Yeah -- "server" is way too generic. I've forgotten where else I use it,
>> but when I build "vnc" I use it to get a VNC server. Maybe make a local
>> "vnc-server" USE flag for that one.
>>
>>
> I don't see what is so dangerous about a server flag. After all I don't set
> doc globally, but it is a useful global flag, with clear intent, as would
> be server.
>
> If usr sets server on a box in make.conf, against advice, they still have to
> actually emerge the pkgs they want, after all. So it's not like it's going
> to lead to a mass of bloat (unlike the current setup.)
>
Actually, on my systems, about the only USE flags I *don't* set globally
in make.conf are "doc", "examples" and "source". There are very few
conflicts from this, and just about everything in my package.use file is
either making the documentation and examples or suppressing an option
for a package where it doesn't work. It turns out to be easier for me to
manage things that way than to have a humongous package.use and a few
options in make.conf.
>
>


--
M. Edward (Ed) Borasky, FBG, AB, PTA, PGS, MS, MNLP, NST, ACMC(P)
http://borasky-research.blogspot.com/

If God had meant for carrots to be eaten cooked, He would have given rabbits fire.

--
gentoo-dev@gentoo.org mailing list
Re: moving USE=server to global [ In reply to ]
>> if you're going to change the topic mid-thread, then you should update
>> the subject
>>
Ah apologies for that; I just wanted to get the technical objection.

>> you already know the state of the server/client debate ... moving it to
>> global doesnt fix any of the short comings, so it should stay local (and
>> removed where possible)
>>
Well, I'm aware of the debate; I still haven't heard a convincing
*technical* argument against (and i have searched this list.) Your position
is fairly clear; can you enlighten usrs as to why you take that stance? I
must be missing something..

http://forums.gentoo.org/viewtopic-t-525893.html if you feel this is
cluttering the dev list with a usr prob.

> Yeah -- "server" is way too generic. I've forgotten where else I use it,
> but when I build "vnc" I use it to get a VNC server. Maybe make a local
> "vnc-server" USE flag for that one.
>
I don't see what is so dangerous about a server flag. After all I don't set
doc globally, but it is a useful global flag, with clear intent, as would
be server.

If usr sets server on a box in make.conf, against advice, they still have to
actually emerge the pkgs they want, after all. So it's not like it's going
to lead to a mass of bloat (unlike the current setup.)


--
gentoo-dev@gentoo.org mailing list
Re: Re: moving USE=server to global [ In reply to ]
Jim Ramsay wrote:
> Steve Long wrote:
>> I don't see what is so dangerous about a server flag. After all I
>> don't set doc globally, but it is a useful global flag, with clear
>> intent, as would be server.
>
> Does it mean:
>
> A) "Compile just the server, as leaving it off compiles both client
> AND server"
>
> B) "Compile the client AND server, as leaving it off compiles ONLY the
> client"
>
> C) All of the above
>
B, as asked for by many usrs. Does that cause other issues I am simply not
seeing?


--
gentoo-dev@gentoo.org mailing list
Re: Re: moving USE=server to global [ In reply to ]
M. Edward (Ed) Borasky wrote:
>> I don't see what is so dangerous about a server flag. After all I don't
>> set doc globally, but it is a useful global flag, with clear intent, as
>> would be server.
>>
>> If usr sets server on a box in make.conf, against advice, they still have
>> to actually emerge the pkgs they want, after all. So it's not like it's
>> going to lead to a mass of bloat (unlike the current setup.)
>>
> Actually, on my systems, about the only USE flags I *don't* set globally
> in make.conf are "doc", "examples" and "source". There are very few
> conflicts from this, and just about everything in my package.use file is
> either making the documentation and examples or suppressing an option
> for a package where it doesn't work. It turns out to be easier for me to
> manage things that way than to have a humongous package.use and a few
> options in make.conf.

With respect, I don't see how this affects anything to do with the flag? Eg
for my personal use, I need KRB5, LDAP, SMB and mySQL. The only server I
want is mySQL for local web-development with apache2, so I'd set server for
dev-db/mySQL in package.use and leave it at that.

I can understand that USE-conditional deps might be required, eg if a user
(grr like you ppl don't know what usr means ;) wants a web-app it might
require a db server. Firstly, I thought those were coming, and secondly in
that case the user would I would think know that server software was
needed, as s/he would be setting up a server app. A simple ewarn would
suffice imo, since the db server might well be on another host.


--
gentoo-dev@gentoo.org mailing list