FAQ Search Today's Posts Mark Forums Read
» Video Reviews

» Linux Archive

Linux-archive is a website aiming to archive linux email lists and to make them easily accessible for linux users/developers.


» Sponsor

» Partners

» Sponsor

Go Back   Linux Archive > Gentoo > Gentoo Development

 
 
LinkBack Thread Tools
 
Old 01-16-2011, 02:05 PM
"Paweł Hajdan, Jr."
 
Default Blockers and package moves

On 1/16/11 2:49 PM, Ciaran McCreesh wrote:
> Second, when performing updates, Paludis also rewrites dependencies of
> installed packages to use the names.

This seems to imply that portage behaves differently. Should we update
PMS when we determine what's the correct behavior?
 
Old 01-16-2011, 02:17 PM
Ciaran McCreesh
 
Default Blockers and package moves

On Sun, 16 Jan 2011 16:05:22 +0100
"Paweł Hajdan, Jr." <phajdan.jr@gentoo.org> wrote:
> On 1/16/11 2:49 PM, Ciaran McCreesh wrote:
> > Second, when performing updates, Paludis also rewrites dependencies
> > of installed packages to use the names.
>
> This seems to imply that portage behaves differently. Should we update
> PMS when we determine what's the correct behavior?

Last time I looked, Portage simply left the old dependencies lying
around, and then silently ignored them most but not all of the time.

PMS has very little to say about how to deal with installed stuff, and
specifies only the format of the updates file, not how it is to be
handled.

That's almost certainly a good thing, since for historical reasons
Portage has some highly perverse behaviour when it comes to packages
where you've got the same version both installed and available in a
repository or overlay (and different Portage versions are perverse to
different extents on that). Requiring emulation of early Portage design
mistakes would just stop Portage from gradually fixing things over
time as has been happening up to now.

--
Ciaran McCreesh
 
Old 01-17-2011, 12:47 AM
Ryan Hill
 
Default Blockers and package moves

On Sun, 16 Jan 2011 13:49:38 +0000
Ciaran McCreesh <ciaran.mccreesh@googlemail.com> wrote:

> People seem to have started using blockers with package moves recently.
> For example, if cat/a is being moved to cat/b, people have started
> putting !cat/a as a dependency in cat/b. This is bad, for two reasons.

It's bad for a couple more too. No one should be doing this.

> Option the first is that people stop writing stupid blockers on package
> moves. Unless someone can come up with a convincing reason to keep them
> there, this is the option that should be taken, and repoman should
> enforce it.

Can repoman get at that info? It could be mentioned in
http://www.gentoo.org/proj/en/devrel/handbook/handbook.xml?part=2&chap=5#doc_chap6


--
fonts, gcc-porting, it makes no sense how it makes no sense
toolchain, wxwidgets but i'll take it free anytime
@ gentoo.org EFFD 380E 047A 4B51 D2BD C64F 8AA8 8346 F9A4 0662
 
Old 01-17-2011, 03:10 PM
Donnie Berkholz
 
Default Blockers and package moves

On 13:49 Sun 16 Jan , Ciaran McCreesh wrote:
> People seem to have started using blockers with package moves recently.
> For example, if cat/a is being moved to cat/b, people have started
> putting !cat/a as a dependency in cat/b. This is bad, for two reasons.
>
> First, you shouldn't have to do that. If package moves aren't working,
> we've got bigger problems, and throwing in some blockers won't help
> there.
>
> Second, when performing updates, Paludis also rewrites dependencies of
> installed packages to use the names. This means that your block on
> cat/a will be rewritten to a block on cat/b, which means the package
> ends up blocking itself.
>
> We've got two options here.
>
> Option the first is that people stop writing stupid blockers on package
> moves. Unless someone can come up with a convincing reason to keep them
> there, this is the option that should be taken, and repoman should
> enforce it.
>
> Option the second is that I make Paludis stop rewriting blockers for
> package moves. This is bad, because it means legitimate blockers which
> should be honoured will end up disappearing.
>
> Does anyone care to justify their "block the old name" habits?

How about playing nicely with overlays where the moves didn't happen
(yet)?

--
Thanks,
Donnie

Donnie Berkholz
Sr. Developer, Gentoo Linux
Blog: http://dberkholz.com
 
Old 01-17-2011, 03:30 PM
Zac Medico
 
Default Blockers and package moves

On 01/17/2011 08:10 AM, Donnie Berkholz wrote:
> On 13:49 Sun 16 Jan , Ciaran McCreesh wrote:
>> Does anyone care to justify their "block the old name" habits?
>
> How about playing nicely with overlays where the moves didn't happen
> (yet)?

That seems like a somewhat justifiable use case. If such a blocker is
placed in RDEPEND and specifies the SLOT (if the package happens to be
slotted), then it shouldn't cause a problem. Even if the package move
gets applied to the blocker, it will be obvious that a slotted circular
blocker in RDEPEND should simply be ignored.
--
Thanks,
Zac
 
Old 01-17-2011, 03:38 PM
Ciaran McCreesh
 
Default Blockers and package moves

On Mon, 17 Jan 2011 10:10:54 -0600
Donnie Berkholz <dberkholz@gentoo.org> wrote:
> How about playing nicely with overlays where the moves didn't happen
> (yet)?

It doesn't play nicely, though, since the overlay won't be mutually
blocking.

--
Ciaran McCreesh
 
Old 01-20-2011, 01:07 PM
Jacob Godserv
 
Default Blockers and package moves

On Mon, Jan 17, 2011 at 11:10, Donnie Berkholz <dberkholz@gentoo.org> wrote:
> How about playing nicely with overlays where the moves didn't happen
> (yet)?

I'm not a developer, but based on my experience, playing nicely with
overlays in general is a can of worms, and as far as I can tell has to
do with the fact that overlays have always needed to react to changes
in portage tree. (For example, if an overlay has pkg-0.1-r1 that fixes
a bug, and in portage tree 0.2 is available but still has the bug,
then the overlay needs to have a 0.2-r1.)

I'm not sure you have to worry about overlays. The developers of the
overlays will worry about them.

Now, for the educated: is this fair or accurate?

--
* * Jacob

* * "For then there will be great distress, unequaled
* * from the beginning of the world until now — and never
* * to be equaled again. If those days had not been cut
* * short, no one would survive, but for the sake of the
* * elect those days will be shortened."

* * Are you ready?
 
Old 01-20-2011, 06:44 PM
Jeroen Roovers
 
Default Blockers and package moves

On Thu, 20 Jan 2011 09:07:20 -0500
Jacob Godserv <jacobgodserv@gmail.com> wrote:

> On Mon, Jan 17, 2011 at 11:10, Donnie Berkholz <dberkholz@gentoo.org>
> wrote:
> > How about playing nicely with overlays where the moves didn't happen
> > (yet)?

> I'm not sure you have to worry about overlays. The developers of the
> overlays will worry about them.
>
> Now, for the educated: is this fair or accurate?

I maintain just the one overlay, which is my own, non-public playpen.
On bugs.gentoo.org, we generally deal with invalid bugs filed because
of out of date / moved / broken overlays, or the maintainers of the now
broken packages in the main tree will, by marking them as invalid, or
perhaps by assigning the bugs to the overlay's maintainer, if [1]
lists the overlay, thereby making it official.

Generally, fixing the main tree because of problems in overlays isn't
desirable, and shouldn't be done, provided it actually could be done at
all.


jer


[1] http://www.gentoo.org/proj/en/overlays/layman-global.txt
 

Thread Tools




All times are GMT. The time now is 12:51 PM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.
Copyright 2007 - 2008, www.linux-archive.org