[gentoo-dev] torsmo has been recontinued named conky. plan on removing torsmo

[gentoo-dev] torsmo has been recontinued named conky. plan on removing torsmo

[Daniel]

[-- Attachment #1: Type: text/plain, Size: 725 bytes --]


Torsmo has been inactive upstream for well over a year 
http://torsmo.sourceforge.net/changelog.php

The large number of bugs
http://sourceforge.net/tracker/?group_id=108594&atid=650929
and large number of patches
http://sourceforge.net/tracker/?group_id=108594&atid=650931
resulted in a fork - conky http://conky.sourceforge.net/.

This is being actively developed by Brenden.

I propose removing torsmo and placing a package.move entry so that conky will 
be the logical upgrade (after all it was the same code base).

Objections with good logical argument or a name in torsmo/metadata.xml may be 
considered.

-- 
Daniel Black <dragonheart@gentoo.org>
Gentoo Crypto/PPC/dev-embedded/Forensics/NetMon

[-- Attachment #2: Type: application/pgp-signature, Size: 190 bytes --]

Re: [gentoo-dev] torsmo has been recontinued named conky. plan on removing torsmo

[Thomas de Grenier de Latour]

On Thu, 5 Jan 2006 23:08:13 +1100
Daniel <dragonheart@gentoo.org> wrote:

> I propose removing torsmo and placing a package.move entry so
> that conky will be the logical upgrade (after all it was the same
> code base).

How will users understand what happened to their good old torsmo,
which used to work fine for months but suddenly disapeared after
an "emerge -u world"?

I think it would be better to package.mask it with some comment
like "Deprecated, masked in favor of app-admin/conky", because
that's something emerge reports when it makes world updates fail
(whereas package moves are absolutly silent and implicit).

In short, for me it's just like any other package deprecation,
because actually the "same code base" argument doesn't really
matter from a user point of view.

--
TGL.
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] torsmo has been recontinued named conky. plan on removing torsmo

[Thomas de Grenier de Latour]

On Thu, 5 Jan 2006 13:55:48 +0100
Thomas de Grenier de Latour <degrenier@easyconnect.fr> wrote:

> I think it would be better to package.mask it with some comment
> like "Deprecated, masked in favor of app-admin/conky", because
> that's something emerge reports when it makes world updates fail

Oops, no, i'm wrong: when a masked package is in world but is not
depended-on by anything else, world update will not fail with the
comments from package.mask, but simply point to emaint and then
continue, ignoring the package. And `emaint --check  world` does
not seem to report comments from package.mask, so my solution is
not that good neither to inform users about what's happening in
such a case.

There is still the solution of the dummy ebuild with some relevant
einfo to replace the actual deprecated package (i think i've seen
that in the past, winex-cvs for instance IIRC), but that's ugly :/

--
TGL.
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] torsmo has been recontinued named conky. plan on removing torsmo

[Paweł Madej]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I got other idea but I don't know how to code it.
In ebuild you can set die on some reasons.
So make in conky ebuild which will be package-move of torsmo such error
that if torsmo is installed emerging conky should fail with die. And
user will be asked to unmerge a torsmo manually and emerge conky then.

What dou you think about such solution?

- --
Paweł Madej aka Nysander
http://quanteam.info          | http://forum-farmaceutyczne.org
http://nysander.quanteam.info | http://wiki.quanteam.info
GPG key: 5861680B             | keyserver: http://pgp.mit.edu
key fingerprint: 34A9 B8BB DFA2 4F0B EFB5  CE50 82F4 8C82 5861 680B
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (GNU/Linux)

iD8DBQFDvSkUgvSMglhhaAsRAm8KAKCHhRRc5v1iwTrp13iUg66UHI0qqwCfWKpJ
0q+qWyGU/jFWU5evKabY8AU=
=PLT4
-----END PGP SIGNATURE-----
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] torsmo has been recontinued named conky. plan2 on removing torsmo

[Daniel]

[-- Attachment #1: Type: text/plain, Size: 640 bytes --]



step 1.

package.mask torsmo

step 2.
get conky the same stable keywords as torsmo

step 3.
put a pre and post install on conky saying you may have got here from torsmo.

step 4.
package.move torsmo -> conky

step 5.
be happy that users have been aware of the changes,
be happy that users now have a more functional product that does the same 
thing and more with the same good interface.
be happy that another gentoo package doesn't need horrible patching and all 
the stuff that Brenden and joemyre fixed up with Conky.

-- 
Daniel Black <dragonheart@gentoo.org>
Gentoo Crypto/PPC/dev-embedded/Forensics/NetMon

[-- Attachment #2: Type: application/pgp-signature, Size: 190 bytes --]

Re: [gentoo-dev] torsmo has been recontinued named conky. plan2 on removing torsmo

[Thomas Matthijs]

* Daniel (dragonheart@gentoo.org) wrote:
 
> step 4.
> package.move torsmo -> conky

This will do _bad_ things if someone has both installed

-- 
 Thomas Matthijs (axxo, knu) 

-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] torsmo has been recontinued named conky. plan2 on removing torsmo

[Thomas de Grenier de Latour]

On Thu, 5 Jan 2006 23:39:13 +0100
Thomas Matthijs <axxo@gentoo.org> wrote:

> * Daniel (dragonheart@gentoo.org) wrote:
>  
> > step 4.
> > package.move torsmo -> conky
> 
> This will do _bad_ things if someone has both installed
> 

Bah, here it's ~ okay because torsmo versions are lower than the
ones of conky: after the move, there will be two conky packages
installed in the same SLOT, and thus the oldest one (actually
torsmo) will get autocleaned soon after by emerge.


I still don't really see the benefit of the package.move related
steps though.  Moving from torsmo to conky will require some manual
work by the users anyway (~/.xinitrc change, new config location,
etc.), and doing half the job automagically at some arbitrary time
just adds confusion for users:
 - those who will have notice the package.mask comments during the
wait after step 1 will already have done the full transition by
themselves.
 - those who didn't will just see their torsmo binary suddenly
disapearing (and given their unlikely visibility, i don't think
einfos in conky ebuild solve that), which imo is worst than just
keeping an obsolete package installed.

Imho, package.move should only be used for changes that are
strictly internal to the portage tree, and not for those which
"break" users' applications.

So i still prefer the simpler, usual, deprecation plan, like what
was done for xawdecode->xdtv for instance, despite they were
also sharing the same code base (an upstream name change actually):
1. package.mask torsmo with a comment about conky
2. wait long enough
3. drop torsmo from the tree

--
TGL.
-- 
gentoo-dev@gentoo.org mailing list

[gentoo-dev] torsmo has been removed from portage

[Daniel]

[-- Attachment #1: Type: text/plain, Size: 1014 bytes --]

I've taken the simple approach and just removed torsmo. It was in the 
package.mask for 2-3 weeks.

Users can emerge conky and rename their configuration file (~/.torsmorc to 
~/.conkyrc) and should get similar result.

On Thu, 5 Jan 2006 11:08 pm, Daniel wrote:
> Torsmo has been inactive upstream for well over a year
> http://torsmo.sourceforge.net/changelog.php
>
> The large number of bugs
> http://sourceforge.net/tracker/?group_id=108594&atid=650929
> and large number of patches
> http://sourceforge.net/tracker/?group_id=108594&atid=650931
> resulted in a fork - conky http://conky.sourceforge.net/.
>
> This is being actively developed by Brenden.
>
> I propose removing torsmo and placing a package.move entry so that conky
> will be the logical upgrade (after all it was the same code base).
>
> Objections with good logical argument or a name in torsmo/metadata.xml may
> be considered.

-- 
Daniel Black <dragonheart@gentoo.org>
Gentoo Crypto/dev-embedded/Forensics/NetMon

[-- Attachment #2: Type: application/pgp-signature, Size: 190 bytes --]