Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 08 Oct 2013 09:51:56 +0200
From:      Mathias Picker <Mathias.Picker@virtual-earth.de>
To:        Pascal Schmid <pascal@lechindianer.de>,freebsd-ports@freebsd.org
Subject:   Re: [HEADSUP] Staging, packaging and more
Message-ID:  <fa14e958-3b95-4983-9230-61a3a9cce2c0@email.android.com>
In-Reply-To: <52519F61.6090506@lechindianer.de>
References:  <20131003084814.GB99713@ithaqua.etoilebsd.net> <524D6059.2000700@FreeBSD.org> <524DD120.4000701@freebsd.org> <20131003203501.GA1371@medusa.sysfault.org> <CAGwOe2Ye2MLz3QpyMW3wyN9ew%2BiNnTETS1oOi_%2B8dPehUcWa0w@mail.gmail.com> <20131004061833.GA1367@medusa.sysfault.org> <20131004063259.GC72453@ithaqua.etoilebsd.net> <20131004065753.GV82824@droso.dk> <20131004070158.GE72453@ithaqua.etoilebsd.net> <20131004111256.GC98118@admin.xzibition.com> <CAJ9axoSF2%2BRys6MG078XCEkKEs2kEpVJegGgqFN3b2t2%2BR80kw@mail.gmail.com> <CAE-m3X1Uixj3anuROC3=kjG6j-LadEXMv_V%2BY=SD=SoDmNky1A@mail.gmail.com> <52519F61.6090506@lechindianer.de>

next in thread | previous in thread | raw e-mail | index | archive | help




Pascal Schmid <pascal@lechindianer=2Ede> schrieb:
>-----BEGIN PGP SIGNE=
D MESSAGE-----
>Hash: SHA1
>
>On 10/06/2013 07:21 PM, Bernhard Fr=C3=B6hlic=
h wrote:
>> On Sun, Oct 6, 2013 at 2:20 PM, Ulrich Sp=C3=B6rlein <uqs@freeb=
sd=2Eorg>
>wrote:
>>> 2013/10/4 Bryan Drewery <bryan@shatow=2Enet>:
>>>> On=
 Fri, Oct 04, 2013 at 09:01:58AM +0200, Baptiste Daroussin wrote:
>>>>> On =
Fri, Oct 04, 2013 at 08:57:53AM +0200, Erwin Lansing wrote:
>>>>>> On Fri, =
Oct 04, 2013 at 08:32:59AM +0200, Baptiste Daroussin
>wrote:
>>>>>>>>>>> 
>=
>>>>>>>>>> Please no devel packages=2E
>>>>>>>>>> 
>>>>>>>>>> Seconded=2E
>=
>>>>>>>> 
>>>>>>>>> What's wrong with devel packages?
>>>>>>>> 
>>>>>>>> It=
 complicates things for developers and custom software on
>FreeBSD=2E The t=
ypical
>>>>>>>> situation that I see on most Linux platforms is a lot of
>c=
onfusion by people, why
>>>>>>>> their custom software XYZ does not properl=
y build - the most
>common answer: they
>>>>>>>> forgot to install a tremen=
dous amount of dev packages,
>containing headers, build
>>>>>>>> tools and =
whatnot=2E On FreeBSD, you can rely on the fact that if
>you installed e=2E=
g=2E
>>>>>>>> libGL, you can start building your own GL applications withou=
t
>the need to install
>>>>>>>> several libGL-dev, libX11-dev, =2E=2E=2E pa=
ckages first=2E This is
>something, which I
>>>>>>>> personally see as a bi=
g plus of the FreeBSD ports system and
>which makes FreeBSD
>>>>>>>> attrac=
tive as a development platform=2E
>>>>>>>> 
>>>>>>> 
>>>>>>> On the other e=
nds, that makes the package fat for embedded
>systems, that also makes
>>>>=
>>> some arbitrary runtime conflicts between packages (because they
>both p=
rovide the same
>>>>>>> symlink on the =2Eso, while we could live with 2 ve=
rsion at
>runtime), that leads to
>>>>>>> tons of potential issue while bui=
lding locally, and that makes
>having sometime insane
>>>>>>> issues with d=
ependency tracking=2E Why having =2Ea, =2Ela, =2Eh etc in
>production serve=
rs? It
>>>>>>> could greatly reduce PBI size, etc=2E
>>>>>>> 
>>>>>>> Perso=
nnaly I do have no strong opinion in one or another
>direction=2E Should we=
 be 
>>>>>>> nicer with developers? with end users? with embedded world? Th=
at
>is the question to
>>>>>>> face to decide if -devel packages is where w=
e want to go or not=2E
>>>>>>> 
>>>>>> 
>>>>>> If we chose to go down that =
path, at least we should chose a
>different name as we've
>>>>>> used the -=
devel suffix for many years for developmental versions=2E
>>>>>> 
>>>>>> I =
must agree that it is one of the things high on my list of
>things that irr=
itate me
>>>>>> with several Linux distributions but I can see the point fo=
r for
>embedded systems as
>>>>>> well=2E  But can't we have both?  Create =
three packages, a default
>full package and split
>>>>>> packages of -bin, =
-lib, and even -doc=2E  My first though twas to
>make the full package
>>>>=
>> a meta-package that would install the split packages in the
>background,=
 but that would
>>>>>> probably be confusing for users at the end of the da=
y, so rather
>just have it be a real
>>>>>> package=2E
>>>>>> 
>>>>> I do l=
ike that idea very much, and it is easily doable with stage
>:)
>>>> 
>>>> =
+1 to splitting packages for embedded usage=2E
>>> 
>>> -1 for the split, a=
s it will not fix anybody's problem=2E
>>> 
>>> On regular machines, disk s=
pace is cheap and having to install more
>packages is just annoying
>>> to =
users=2E Think of the time wasted that people are told to apt-get
>libfoo-d=
ev before they can
>>> build anything from github, or similar=2E
>>> 
>>> I=
f you actually *are* space constricted on your tiny embedded
>machine, what=
 the fuck are you
>>> doing with the sqlite database and all the metadata a=
bout
>ports/packages anyway? Just rm
>>> /usr/include and /usr/share/doc, /=
usr/share/man, etc=2E when building
>your disk image=2E But you
>>> are doi=
ng that already anyway, so this solves no actual problem for
>you=2E
>>> 
>=
>> My two cents Uli
>> 
>> I also don't see why we need to optimize our pac=
kages for an embedded
>environment that is
>> usually very customized=2E Wo=
uldn't it make more sense to provide some
>proper port / packaging
>> optio=
ns/flags that help to optimize size of the packages without
>touching heade=
r files? People
>> could use that flags and poudriere to build their packag=
es together
>with all their other 
>> compiler flags and cpu optimisations=
=2E
>> 
>
>+1
>
>As far as I can see Daniel Nebdal's approach ("WITH_DEV_FI=
LES" flag,
>and defaulting to "yes")
>sounds promising=2E

+1 

This doesn'=
t change things in the standard case and follows existing patterns, so I li=
ke it, too=2E

Mathias

>
>Pascal
>-----BEGIN PGP SIGNATURE-----
>Version: =
GnuPG v2=2E0=2E21 (GNU/Linux)
>Comment: Using GnuPG with Thunderbird - http=
://www=2Eenigmail=2Enet/
>
>iQIcBAEBAgAGBQJSUZ9hAAoJEAWefonBOgAfDlUP/3117hV=
dZ6WhrygIGnctSb49
>V+i0SggAFxXuvFFYlkjexrWFpjMPN2H7vBtR9DVbLNwqb4En+mVj/LVY=
1ejS9TAQ
>gj/nKlK6HNdVQWQD8qLfzFUAzWwnSBco/rIOiGkOrHuvFSUCTV5gPehoJ+Vg8Qnz
=
>dyUp5SByePNpY1MGMTJZh9gKWJFtTe8DcanDBCVL65rZf/eOVPyiMwlQK+Fy2AQj
>OQgJxhkW=
Jzvl5V9THsMGiSCzJ+9EMoC620F9WEs3MvO0Ky2zIercFJ2bDaks6CXn
>arNTsqTT1zI0sZNGN=
QMrnxYtQPgV3oCEAggj4ZOG0FkhmBkxWNOPUyahBUE/V8ds
>tvLvugzVzqeaIJWg3IKDNEfGGh=
0ZnAMhUakUHyJPDhuCLgb498uwElesmgaSvlky
>eotS4cWGVp2lquuf/xPRRl82K4ciozZi3mt=
tRmrfoznK69p1HJbepCn9maIhFkii
>WqLTjKVkeZ778is8mw8dom/Qb8OEj+XR6Vetq7cLg4Is=
//zieKzSvMWm7QrW1dAI
>zohAjP+lMP5d3TEmeVqvSZhQ9ticzqGGaW4U7zxxRZ0Y/zxkBwe3c=
IBEpjTpnW9p
>/a0DJ3JodVBo79N2JheIqweCK9RPn8rOK5HxujnWcJ3jbQAgCxOdLd9iyN6IxO=
jI
>3pHI9pO++Am9ReFvL/Uy
>=3Dqm+q
>-----END PGP SIGNATURE-----
>___________=
____________________________________
>freebsd-ports@freebsd=2Eorg mailing l=
ist
>http://lists=2Efreebsd=2Eorg/mailman/listinfo/freebsd-ports
>To unsubs=
cribe, send any mail to
>"freebsd-ports-unsubscribe@freebsd=2Eorg"




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?fa14e958-3b95-4983-9230-61a3a9cce2c0>