On Tue, 06 Sep 2011, Russ Allbery wrote:
Lars Wirzenius l...@liw.fi writes:
Would one of the following work?
1. Merge perl and perl-modules. Some waste across architectures in the
archive, but not horribly much.
This is probably the simplest solution conceptually and in terms of
Le Tuesday 6 September 2011 08:36:15, Russ Allbery a écrit :
The problem is that perl and perl-modules really are one package that was
split apart solely to get the (large) architecture-independent parts into
an arch: all package.
Then, may be, we should flag them accordingly. Let's imagine a
On Mon, 05 Sep 2011, Josselin Mouette wrote:
Le dimanche 04 septembre 2011 à 17:30 -0700, Russ Allbery a écrit :
I would like to see someone sort out what to do about perl and
perl-modules before Policy makes it a requirement.
Is there anything special about this case? It looks to me that
Raphael Hertzog hert...@debian.org writes:
On Mon, 05 Sep 2011, Josselin Mouette wrote:
Le dimanche 04 septembre 2011 à 17:30 -0700, Russ Allbery a écrit :
I would like to see someone sort out what to do about perl and
perl-modules before Policy makes it a requirement.
Is there anything
]] Russ Allbery
| Yeah, perl-modules really does depend on perl, in that the package is
| actively broken and will break other things, not just merely unusable,
| without it. Note that perl-modules has a ton of Provides, and something
| depending on one of those provided modules without
[...]
The problem is that perl and perl-modules really are one package that was
split apart solely to get the (large) architecture-independent parts into
an arch: all package.
Wouldn't this problem be solved by moving the contents of perl to, e.g.,
perl-bin, making perl a dummy package,
On Tue, 06 Sep 2011, Michael Tautschnig wrote:
[...]
The problem is that perl and perl-modules really are one package that was
split apart solely to get the (large) architecture-independent parts into
an arch: all package.
Wouldn't this problem be solved by moving the contents of
On Tue, Sep 06, 2011 at 11:57:39AM +0200, Raphael Hertzog wrote:
On Tue, 06 Sep 2011, Michael Tautschnig wrote:
The problem is that perl and perl-modules really are one package that was
split apart solely to get the (large) architecture-independent parts into
an arch: all package.
Lars Wirzenius l...@liw.fi writes:
Would one of the following work?
1. Merge perl and perl-modules. Some waste across architectures in the
archive, but not horribly much.
This is probably the simplest solution conceptually and in terms of tool
changes required, but the Perl maintainers
Le dimanche 04 septembre 2011 à 17:30 -0700, Russ Allbery a écrit :
I would like to see someone sort out what to do about perl and
perl-modules before Policy makes it a requirement.
Is there anything special about this case? It looks to me that the
perl-modules→perl dependency just needs to be
Le samedi 03 septembre 2011 à 11:54 +0200, Bill Allombert a écrit :
Today circular dependencies in unstable reached an all-time low, with
only 40 circular dependencies.
I think it should now be clear that there aren’t any such issues that
cannot be fixed, with more or less complication.
On 09/05/2011 10:13 PM, Bill Allombert wrote:
Le samedi 03 septembre 2011 à 11:54 +0200, Bill Allombert a écrit :
Today circular dependencies in unstable reached an all-time low, with
only 40 circular dependencies.
I think it should now be clear that there aren’t any such issues that
cannot
Bill Allombert bill.allomb...@math.u-bordeaux1.fr (03/09/2011):
I would like to thanks all the developpers that have worked toward
removing circular dependencies, and especially the X Strike Force and
the Debian Java maintainers.
The X Strike Force would like to thank David Kalnischkies for
Le samedi 03 septembre 2011 à 11:54 +0200, Bill Allombert a écrit :
Today circular dependencies in unstable reached an all-time low, with
only 40 circular dependencies.
I think it should now be clear that there aren’t any such issues that
cannot be fixed, with more or less complication.
Given
On 2011-09-04 13:49, Josselin Mouette wrote:
I think it should now be clear that there aren’t any such issues that
cannot be fixed, with more or less complication.
Given the benefits for dependency resolvers to be able to guarantee the
dependency tree is actually a tree and not a DAG, isn’t
On 04/09/2011 14:44, Eugene V. Lyubimkin wrote:
While I also would want Debian to eventually get rid of circular
dependencies, I am not sure about (the value of) the benefits.
For example, even by default dependency resolvers have to consider
Depends-Recommends loops. And if the user has
On Sun, 04 Sep 2011 13:49:06 +0200, Josselin Mouette j...@debian.org wrote:
Non-text part: multipart/signed
Given the benefits for dependency resolvers to be able to guarantee the
dependency tree is actually a tree and not a DAG
Pardon my terminology nit-pick, but do you mean guarantee the
Le dimanche 04 septembre 2011 à 11:41 -0300, David Bremner a écrit :
Pardon my terminology nit-pick, but do you mean guarantee the
dependency graph is actually a DAG? The A in DAG is for acyclic.
Yes of course. (Not enough sleep last night…)
--
.''`. Josselin Mouette
: :' :
`. `'
`-
On 2011-09-04 15:42, Vincent Danjean wrote:
On 04/09/2011 14:44, Eugene V. Lyubimkin wrote:
While I also would want Debian to eventually get rid of circular
dependencies, I am not sure about (the value of) the benefits.
For example, even by default dependency resolvers have to consider
[Eugene V. Lyubimkin]
While I also would want Debian to eventually get rid of circular
dependencies, I am not sure about (the value of) the benefits.
One problem I run into every once in a while is how circulars
dependencies make upgrades fail to recover when a disk is full or
something else
Petter Reinholdtsen p...@hungry.com writes:
[Eugene V. Lyubimkin]
Therefore I think _for this moment_ mandating in the policy will be
too strict.
I definitely believe it is time to require no loops in policy.
I would like to see someone sort out what to do about perl and
perl-modules before
Dear developers,
Today circular dependencies in unstable reached an all-time low, with
only 40 circular dependencies.
Here the current list:
* acheck acheck-rules
* aide aide-common
* autodocktools mgltools-pmv
* bochs bochs-wx
* bootcd bootcd-i386 bootcd-ia64 (1 pre-depends)
*
[Bill Allombert]
Here the list of current circular dependencies:
[...]
* perl perl-modules
Heh, I suspect this one just caused a test upgrade I did of a KDE
desktop from Lenny to Squeeze to fail.
Upgrading perl-modules failed with this error message (in Norwegian,
sorry about that):
dpkg:
Hi Bill,
On Sonntag, 6. Juni 2010, Bill Allombert wrote:
Today circular dependencies in unstable reached an all-time low.
Really? I'm surprised. (But don't have any reason not to believe, I'm just
surprised from watching piuparts results (piuparts cant cope well with
circular depends (see the
Hi,
On Sonntag, 6. Juni 2010, Petter Reinholdtsen wrote:
Today circular dependencies in unstable reached an all-time low.
Very good to hear. If only we could get it down to zero, piuparts
would be able to test all the packages
As Steve already pointed out: this is (for the very most) a bug
Hi,
On 2010年06月06日 18:50, Neil Williams wrote:
Some might be too intrusive to be fixed for squeeze (lbc6, perl, g++)
libc6 is too difficult.
perl-modules - perl could be demoted to Recommends, libstdc++6-4.4-dev
- g++-4.4 likewise. Neither package provides functionality on its own
that
Hi,
On 2010年06月06日 15:21, Bill Allombert wrote:
Today circular dependencies in unstable reached an all-time low.
Excellent.
* libc6 libgcc1
This one is tricky, because it is an accurate reflection of the ELF
headers. The gcc and glibc build processes actually jump through a lot
of
On 06/06/2010 18:29, Josselin Mouette wrote:
Le dimanche 06 juin 2010 à 18:06 +0200, Petter Reinholdtsen a écrit :
[Bill Allombert]
Dear developers,
Today circular dependencies in unstable reached an all-time low.
Very good to hear. If only we could get it down to zero, piuparts
would be
On 07/06/2010 11:38, Simon Richter wrote:
The other symbols from libgcc are weak references to pthread functions
(i.e. no dependency there),
gcc-4.5 can change that (unless bug #577961 has been fixed: I did not
recheck recently)
Regards,
Vincent
--
Vincent Danjean GPG key ID
On 06/06/2010 14:21, Bill Allombert wrote:
* exim4 exim4-base exim4-daemon-heavy exim4-daemon-light fcron
exim4-base depends on cron | fcron and fcron depends on exim4 |
mail-transport-agent. Is this strictly a circular dependency? Is it one
that causes issues?
--
To UNSUBSCRIBE, email to
On Mon, Jun 07, 2010 at 11:49:54AM +0200, Simon Richter wrote:
perl-modules - perl could be demoted to Recommends, libstdc++6-4.4-dev
- g++-4.4 likewise. Neither package provides functionality on its own
that breaks if the other package is not installed, rather, they'd just
be dead
Adam Borowski wrote:
perl-modules is exactly the same case as foo-data - foo. I don't see why
it should be treated any different.
No. IIRC already discusse ond -devel. 'perl-modules' contain arch:all Perl
modules some of which depend on arch:any modules in 'perl' and vice-versa.
Still can be
On Mon, Jun 07, 2010 at 04:20:22PM +0200, Holger Levsen wrote:
mind if we keep debian-devel in the loop? If not, please reply to
debian-devel@lists.debian.org and quote me freely :-)
of course not!
[...]
Do you plan to package this for Debian? Then I could easily re-use it on
On Mon, Jun 07, 2010 at 10:35:24AM +0200, Holger Levsen wrote:
Hi Bill,
On Sonntag, 6. Juni 2010, Bill Allombert wrote:
Today circular dependencies in unstable reached an all-time low.
Really? I'm surprised. (But don't have any reason not to believe, I'm just
surprised from watching
[Robert Lemmen]
so if anyone can think of good use cases for a tool that allows
analytics on the dependencies between entities in the debian space
(packages, bugs etc), then please send them my way. with a moderatly
long list of use cases, i could go on defining the query language...
One
Dear developers,
Today circular dependencies in unstable reached an all-time low.
Here the list of current circular dependencies:
* libc6 libgcc1
* perl perl-modules
* debconf debconf-english debconf-i18n
* abuse abuse-frabs abuse-lib
* ghostscript gs-common
* python-imaging python-imaging-tk
*
[Bill Allombert]
Dear developers,
Today circular dependencies in unstable reached an all-time low.
Very good to hear. If only we could get it down to zero, piuparts
would be able to test all the packages and a more deterministic
package installation order would be ensured. :)
Happy hacking,
Le dimanche 06 juin 2010 à 18:06 +0200, Petter Reinholdtsen a écrit :
[Bill Allombert]
Dear developers,
Today circular dependencies in unstable reached an all-time low.
Very good to hear. If only we could get it down to zero, piuparts
would be able to test all the packages and a more
--=20
Eugene V. Lyubimkin aka JackYF, JID: jackyf.devel(maildog)gmail.com
C++/Perl developer, Debian Developer
signature.asc
Description: OpenPGP digital signature
On Sun, 06 Jun 2010 18:29:01 +0200
Josselin Mouette j...@debian.org wrote:
Le dimanche 06 juin 2010 à 18:06 +0200, Petter Reinholdtsen a écrit :
[Bill Allombert]
Dear developers,
Today circular dependencies in unstable reached an all-time low.
Very good to hear. If only we could
Hi,
On Sun, Jun 06, 2010 at 06:29:01PM +0200, Josselin Mouette wrote:
Le dimanche 06 juin 2010 à 18:06 +0200, Petter Reinholdtsen a écrit :
[Bill Allombert]
Dear developers,
Today circular dependencies in unstable reached an all-time low.
Very good to hear. If only we could get it
On Sun, 6 Jun 2010 18:53:04 +0200
Rene Engelhard r...@debian.org wrote:
On Sun, Jun 06, 2010 at 06:29:01PM +0200, Josselin Mouette wrote:
Le dimanche 06 juin 2010 à 18:06 +0200, Petter Reinholdtsen a
écrit :
[Bill Allombert]
Dear developers,
Today circular dependencies in unstable
Hi,
On Sun, Jun 06, 2010 at 06:15:41PM +0100, Neil Williams wrote:
The bug report doesn't explain why this needs to be a Depends: either -
it could be a Recommends AFAICT. To quote the report, which for some
stuff needs - the definition of a Recommends in my book. If
some stuff in the
On Sun, 6 Jun 2010 19:18:19 +0200
Rene Engelhard r...@debian.org wrote:
On Sun, Jun 06, 2010 at 06:15:41PM +0100, Neil Williams wrote:
The bug report doesn't explain why this needs to be a Depends:
either - it could be a Recommends AFAICT. To quote the report,
which for some stuff
Le dimanche 06 juin 2010 à 18:53 +0200, Rene Engelhard a écrit :
No. Besides that I think it would be bad to forbid correct dependencies it
would break some subpolicies. (I have the cli-uno-bridge -
libuno-cppuhelper1.0-cil
one in mind, see #495748).
This is a very classical case of circular
On Sun, Jun 06, 2010 at 06:29:01PM +0200, Josselin Mouette wrote:
Le dimanche 06 juin 2010 à 18:06 +0200, Petter Reinholdtsen a écrit :
[Bill Allombert]
Dear developers,
Today circular dependencies in unstable reached an all-time low.
Very good to hear. If only we could get it down
46 matches
Mail list logo
