[gentoo-dev] Doing and then undoing slotmoves

[gentoo-dev] Doing and then undoing slotmoves

[Fabio Erculiani]

Hi,
6 days ago gienah committed a bunch of slotmoves for the haskell
glib/gtk stuff [1], basically moving the pkgs to slot 0 (from slot 2).
This was done in file 4Q-2013.
It turns out that the same gienah moved those pkgs to slot 2 (from
slot 0) in 2Q-2013 [2].

I have never seen something like that and this generated an
interesting bug in entropy (well, I fixed it...). What I am asking is
quite simple though.
Is this allowed? Should the previous slotmove be removed?

Thanks,

[1] http://sources.gentoo.org/cgi-bin/viewvc.cgi/gentoo-x86/profiles/updates/4Q-2013?view=log
[2] http://sources.gentoo.org/cgi-bin/viewvc.cgi/gentoo-x86/profiles/updates/2Q-2013?r1=1.1&r2=1.2
-- 
Fabio Erculiani

Re: [gentoo-dev] Doing and then undoing slotmoves

[Ulrich Mueller]

>>>>> On Wed, 18 Dec 2013, Fabio Erculiani wrote:

> I have never seen something like that and this generated an
> interesting bug in entropy (well, I fixed it...). What I am asking
> is quite simple though. Is this allowed?

The PMS does not allow it:
http://dev.gentoo.org/~ulm/pms/5/pms.html#x1-390004.4.4

Ulrich

Re: [gentoo-dev] Doing and then undoing slotmoves

["Paweł Hajdan, Jr."]

[-- Attachment #1: Type: text/plain, Size: 925 bytes --]

On 12/18/13, 10:19 AM, Ulrich Mueller wrote:
>>>>>> On Wed, 18 Dec 2013, Fabio Erculiani wrote:
> 
>> I have never seen something like that and this generated an
>> interesting bug in entropy (well, I fixed it...). What I am asking
>> is quite simple though. Is this allowed?
> 
> The PMS does not allow it:
> http://dev.gentoo.org/~ulm/pms/5/pms.html#x1-390004.4.4

I think it goes even further:

"Any name that has appeared as the origin of a move must not be reused
in the future. Any slot that has appeared as the origin of a slot move
may not be used by packages matching the spec of that slot move in the
future."

To me this means moving the packages in question back to slot 2 is never
allowed, because they were once removed from it.

Does this interpretation sound correct? What is the rationale for PMS logic?

Finally, do we have a good way now to automate checks against this?

Paweł


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 203 bytes --]

Re: [gentoo-dev] Doing and then undoing slotmoves

[Fabio Erculiani]

On Wed, Dec 18, 2013 at 1:13 PM, "Paweł Hajdan, Jr."
<phajdan.jr@gentoo.org> wrote:
[ snip ]
>
> Finally, do we have a good way now to automate checks against this?

The current PMS spec, as you quoted, allows one way moves only.
For this reason, I guess that simulating the updates twice should
result in no applicable updates on the second pass, unless a "circular
dependency" is found.
Assuming that the simulation step is more or less constant time (is
it?), this could only take O(2n), O(n) normalized.

I implemented something along these lines in entropy and it spotted
the faulty slotmove lines.

>
> Paweł
>



-- 
Fabio Erculiani

Re: [gentoo-dev] Doing and then undoing slotmoves

[Donnie Berkholz]

[-- Attachment #1: Type: text/plain, Size: 687 bytes --]

On 10:19 Wed 18 Dec     , Ulrich Mueller wrote:
> >>>>> On Wed, 18 Dec 2013, Fabio Erculiani wrote:
> 
> > I have never seen something like that and this generated an
> > interesting bug in entropy (well, I fixed it...). What I am asking
> > is quite simple though. Is this allowed?
> 
> The PMS does not allow it:
> http://dev.gentoo.org/~ulm/pms/5/pms.html#x1-390004.4.4

Seems we should add repoman support to check profiles/. Spec mandates 
that are not implemented in any tool are unlikely to be adhered to.

-- 
Thanks,
Donnie

Donnie Berkholz
Council Member / Sr. Developer, Gentoo Linux <http://dberkholz.com>
Analyst, RedMonk <http://redmonk.com/dberkholz/>

[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-dev] Doing and then undoing slotmoves

[Jeroen Roovers]

On Sun, 22 Dec 2013 20:07:21 -0600
Donnie Berkholz <dberkholz@gentoo.org> wrote:

> Seems we should add repoman support to check profiles/. Spec mandates 
> that are not implemented in any tool are unlikely to be adhered to.

repoman does not work in profiles, to my knowledge. It expects ebuild
in package directories in categorie directories.


     jer

Re: [gentoo-dev] Doing and then undoing slotmoves

[Donnie Berkholz]

[-- Attachment #1: Type: text/plain, Size: 656 bytes --]

On 12:31 Mon 23 Dec     , Jeroen Roovers wrote:
> On Sun, 22 Dec 2013 20:07:21 -0600
> Donnie Berkholz <dberkholz@gentoo.org> wrote:
> 
> > Seems we should add repoman support to check profiles/. Spec mandates 
> > that are not implemented in any tool are unlikely to be adhered to.
> 
> repoman does not work in profiles, to my knowledge. It expects ebuild
> in package directories in categorie directories.

I'm confused. Isn't that exactly what I just said? "add repoman support"

-- 
Thanks,
Donnie

Donnie Berkholz
Council Member / Sr. Developer, Gentoo Linux <http://dberkholz.com>
Analyst, RedMonk <http://redmonk.com/dberkholz/>

[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-dev] Doing and then undoing slotmoves

[Jeroen Roovers]

On Wed, 25 Dec 2013 20:15:00 -0600
Donnie Berkholz <dberkholz@gentoo.org> wrote:

> > repoman does not work in profiles, to my knowledge. It expects
> > ebuild in package directories in categorie directories.
> 
> I'm confused. Isn't that exactly what I just said? "add repoman
> support"

Yes, you did, but apparently that's non-trivial, since it has been
requested again and again over the years on the mailing lists. :)

There is even a bug about it. #390651


     jer

Re: [gentoo-dev] Doing and then undoing slotmoves

[Alex Alexander]

[-- Attachment #1: Type: text/plain, Size: 655 bytes --]

On Wed, Dec 18, 2013 at 11:11 AM, Fabio Erculiani <lxnay@gentoo.org> wrote:

> Hi,
> 6 days ago gienah committed a bunch of slotmoves for the haskell
> glib/gtk stuff [1], basically moving the pkgs to slot 0 (from slot 2).
> This was done in file 4Q-2013.
> It turns out that the same gienah moved those pkgs to slot 2 (from
> slot 0) in 2Q-2013 [2].
>
> I have never seen something like that and this generated an
> interesting bug in entropy (well, I fixed it...). What I am asking is
> quite simple though.
> Is this allowed? Should the previous slotmove be removed?
>

Did a few quick tests, seems to me that the old slotmove should be removed.

Alex

[-- Attachment #2: Type: text/html, Size: 1009 bytes --]