NetBSD source update | 1 Mar 02:14 2009
Picon

daily pkgsrc CVS update output


Updating pkgsrc tree:
? pkgsrc/INDEX
? pkgsrc/README-IPv6.html
? pkgsrc/README-all.html
P pkgsrc/devel/libextractor/Makefile
P pkgsrc/devel/libextractor/distinfo
P pkgsrc/devel/libextractor/patches/patch-ac
P pkgsrc/devel/p5-Test-Harness/Makefile
U pkgsrc/devel/p5-Test-Harness/distinfo
P pkgsrc/doc/CHANGES-2009
P pkgsrc/emulators/wine-devel/Makefile
P pkgsrc/emulators/wine-devel/PLIST.common
P pkgsrc/emulators/wine-devel/distinfo
P pkgsrc/emulators/wine-devel/patches/patch-ad
P pkgsrc/emulators/wine-devel/patches/patch-af
P pkgsrc/emulators/wine-devel/patches/patch-ba
P pkgsrc/graphics/MesaLib/glx-config.mk
P pkgsrc/graphics/graphviz/distinfo
P pkgsrc/graphics/graphviz/patches/patch-ab
P pkgsrc/lang/gcc3/distinfo
U pkgsrc/lang/gcc3/patches/patch-at
P pkgsrc/lang/gcc3-c/Makefile
P pkgsrc/mail/dkim-milter/Makefile
P pkgsrc/mail/dkim-milter/PLIST
P pkgsrc/mail/dkim-milter/distinfo
P pkgsrc/mail/dkim-milter/options.mk
P pkgsrc/mail/dkim-milter/patches/patch-aa
P pkgsrc/mail/roundcube/Makefile
P pkgsrc/pkgtools/pkg_install/files/add/main.c
(Continue reading)

Joerg Sonnenberger | 1 Mar 19:47 2009
Picon

Re: Half-installed packages

On Sat, Feb 28, 2009 at 10:36:27PM +0100, Havard Eidnes wrote:
> I think this points to a lack of robustness in the package tools,
> perhaps first and foremost in that pkg_add should not half-add a package
> if something goes amiss during installation, but also that pkg_delete
> ought to be able to cope with this sort of problem.

I think you are confusing an important part here. pkg_add is not used
for non-destdir installation, so it simply can't avoid/fix that issue.
The only situation in which a reasonable current pkg_add should leave an
uninstallable package in /var/db/pkg is when you reboot at the wrong
time. There is a small window for that. pkg_delete doesn't and IMO
shouldn't cope with that as it is a sign that your tree is quite
corrupt, so every assumption it make can be a bad one.

Joerg

Greg Troxel | 1 Mar 19:51 2009
Picon

amanda: expressing defaults again in Makefile?


I am needing amanda on a small-memory sparc, and am thus newly motivated
to make the default packages acceptable to me.  I am taking a cleanup
pass on the packages, and noticed that AMANDA_{USER,TMP,VAR} are in
/usr/pkgsrc/mk/default/mk.conf.  I am guessing that they should
therefore not be defined in the makefile, and have the following patch.
OK to commit, or am I confused?

Barring objections, I will also be making AMANDA_FQDN and AMANDA_SSH the default, perhaps
without even an option, because I now of no good reason why they should
ever be off.

Index: sysutils/amanda-common/Makefile.common
===================================================================
RCS file: /cvsroot/pkgsrc/sysutils/amanda-common/Makefile.common,v
retrieving revision 1.27
diff -u -p -r1.27 Makefile.common
--- sysutils/amanda-common/Makefile.common	29 Jul 2008 17:21:39 -0000	1.27
+++ sysutils/amanda-common/Makefile.common	1 Mar 2009 18:47:01 -0000
 <at>  <at>  -10,7 +10,7  <at>  <at>  DISTNAME?=		amanda-${VERS}
 CATEGORIES=		sysutils
 MASTER_SITES?=		${MASTER_SITE_SOURCEFORGE:=amanda/}

-MAINTAINER=		pkgsrc-users <at> NetBSD.org
+MAINTAINER=		gdt <at> NetBSD.org
 HOMEPAGE=		http://www.amanda.org/

 .include "../../mk/bsd.prefs.mk"
 <at>  <at>  -22,10 +22,6  <at>  <at>  AMANDA_GROUP?=		sysadmin
 AMANDA_GROUP?=		operator
(Continue reading)

Lloyd Parkes | 1 Mar 21:12 2009
Picon
Picon

Re: amanda: expressing defaults again in Makefile?

Greg Troxel wrote:
> I am needing amanda on a small-memory sparc, and am thus newly motivated
> to make the default packages acceptable to me.  I am taking a cleanup
> pass on the packages, and noticed that AMANDA_{USER,TMP,VAR} are in
> /usr/pkgsrc/mk/default/mk.conf.  I am guessing that they should
> therefore not be defined in the makefile, and have the following patch.
> OK to commit, or am I confused?

I think they should be commented out rather than deleted, with a one line 
comment above them so that it's obvious that they are defaults that can be 
overridden.

Cheers
--

-- 
Lloyd Parkes
Systems Consultant
Open Systems
Ph: +64 4 890 2437

Joerg Sonnenberger | 1 Mar 21:52 2009
Picon

Re: amanda: expressing defaults again in Makefile?

On Sun, Mar 01, 2009 at 01:51:54PM -0500, Greg Troxel wrote:
> I am needing amanda on a small-memory sparc, and am thus newly motivated
> to make the default packages acceptable to me.  I am taking a cleanup
> pass on the packages, and noticed that AMANDA_{USER,TMP,VAR} are in
> /usr/pkgsrc/mk/default/mk.conf.  I am guessing that they should
> therefore not be defined in the makefile, and have the following patch.
> OK to commit, or am I confused?

Stupid question, why is this user called backup at all? I think the
defaults should also be in the package Makefile(.common) and not
defaults/mk.conf, given that we now all tag them with BUILD_DEFS.
Please give others a chance to comment on that first though.

> Barring objections, I will also be making AMANDA_FQDN and AMANDA_SSH
> the default, perhaps without even an option, because I now of no good
> reason why they should ever be off.

At the very least they should be converted to the option framework.
Just guessing, but how does amanda behave on a host without FQDN when
compiled with that option?

Joerg

Greg Troxel | 1 Mar 22:10 2009
Picon

Re: amanda: expressing defaults again in Makefile?


Joerg Sonnenberger <joerg <at> britannica.bec.de> writes:

> On Sun, Mar 01, 2009 at 01:51:54PM -0500, Greg Troxel wrote:
>> I am needing amanda on a small-memory sparc, and am thus newly motivated
>> to make the default packages acceptable to me.  I am taking a cleanup
>> pass on the packages, and noticed that AMANDA_{USER,TMP,VAR} are in
>> /usr/pkgsrc/mk/default/mk.conf.  I am guessing that they should
>> therefore not be defined in the makefile, and have the following patch.
>> OK to commit, or am I confused?
>
> Stupid question, why is this user called backup at all? I think the

Separate issue; I was going to change it to 'amanda'.

> defaults should also be in the package Makefile(.common) and not
> defaults/mk.conf, given that we now all tag them with BUILD_DEFS.
> Please give others a chance to comment on that first though.

That would be ok with me, but I'm not sure on the norms here.

>> Barring objections, I will also be making AMANDA_FQDN and AMANDA_SSH
>> the default, perhaps without even an option, because I now of no good
>> reason why they should ever be off.
>
> At the very least they should be converted to the option framework.

Agreed, but better still to just remove choices that no one wants.

> Just guessing, but how does amanda behave on a host without FQDN when
(Continue reading)

Joerg Sonnenberger | 1 Mar 22:19 2009
Picon

Re: amanda: expressing defaults again in Makefile?

On Sun, Mar 01, 2009 at 04:10:50PM -0500, Greg Troxel wrote:
> > Just guessing, but how does amanda behave on a host without FQDN when
> > compiled with that option?
> 
> I am not quite sure; it seems obvious to me that always using FQDNs is
> the right thing to do :-)

E.g. postfix requires you to have a FQDN, does it do the same? This
might surprise people.

Joerg

David Laight | 1 Mar 23:06 2009
Picon

Re: amanda: expressing defaults again in Makefile?

On Sun, Mar 01, 2009 at 04:10:50PM -0500, Greg Troxel wrote:
> > Stupid question, why is this user called backup at all? I think the
> 
> Separate issue; I was going to change it to 'amanda'.

Well, that is a more stupid default.
For instance my sister is 'Amanda', and that could easily be her username!

	David

--

-- 
David Laight: david <at> l8s.co.uk

Adrian Portelli | 1 Mar 23:18 2009
Picon

Re: amanda: expressing defaults again in Makefile?

On 1/3/09 22:06, David Laight wrote:
> On Sun, Mar 01, 2009 at 04:10:50PM -0500, Greg Troxel wrote:
>>> Stupid question, why is this user called backup at all? I think the
>> Separate issue; I was going to change it to 'amanda'.
> 
> Well, that is a more stupid default.
> For instance my sister is 'Amanda', and that could easily be her username!
> 
> 	David
> 

Sorry for dragging this thread further off subject . . .

Wasn't there talk a little while ago about prefixing pkgsrc created user
names and group names so they could be easily identified from
system/user accounts ?

adrian.

David Brownlee | 1 Mar 23:23 2009
Picon

Re: pkg_info(1) and wildcards / by-file listenings

On Sat, 28 Feb 2009, Joerg Sonnenberger wrote:

> Hi all,
> currently pkg_info will incrementally list all packages that match each
> argument when used to query the pkgdb. This has the advantage of
> starting earlier, but has the disadvantage of printing the same package
> possibly multiple times and also without sorting. What do you think of
> changing this to build a list of unique packages matching the arguments
> first, optionally sorting it and then creating output?

 	That sounds like a very a good idea, would make the output
 	more useful for scripts and humans.

> Possible sort criterions are lexical and install time. Selection
> criterions are all, manual installed, automatically installed, leaves
> as well as packages matching the given list.

 	Would you default all output to one sort (say alpha) and then
 	provide a sort by 'x' option?

 	On a side note, would there be any sense in make -e take
 	multiple package names?

--

-- 
 		David/absolute       -- www.NetBSD.org: No hype required --


Gmane