portability.eclass: dead egethome, egetshell, is-login-disabled funcs ?

10-26-2011, 11:40 PM

i can't see any ebuild/eclass using egethome, egetshell,
is-login-disabled from portability.eclass. anyone have a reason for
keeping these before i punt them ?
-mike

10-30-2011, 09:33 PM

On 27.10.2011 2.40, Mike Frysinger wrote:
> i can't see any ebuild/eclass using egethome, egetshell,
> is-login-disabled from portability.eclass. anyone have a reason for
> keeping these before i punt them ?
> -mike
>

Breaking overlays. Isn't the standing policy still to not break
backwards compatibility as long as an eclass exists?

Regards,
Petteri

10-30-2011, 09:37 PM

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512

On 10/30/2011 10:33 PM, Petteri Räty wrote:
> On 27.10.2011 2.40, Mike Frysinger wrote:
>> i can't see any ebuild/eclass using egethome, egetshell,
>> is-login-disabled from portability.eclass. anyone have a reason
>> for keeping these before i punt them ? -mike
>>
>
> Breaking overlays. Isn't the standing policy still to not break
> backwards compatibility as long as an eclass exists?
>
> Regards, Petteri
>

We have no have direct control on overlays. It is up to maintainers to
ensure that their packages work with portage eclasses or they can have
their own eclasses locally.

- --
Regards,
Markos Chandras / Gentoo Linux Developer / Key ID: B4AFF2C2
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.18 (GNU/Linux)

iQIcBAEBCgAGBQJOrdGOAAoJEPqDWhW0r/LCvOsP/RDCgBYuZhgy8e3/MBqD7nyx
wnWyGeLBV7q8n8kVnRCMIKlgiKxKnphPUZhkcddxstzoGv3nAz ValxXVkfkyhl4G
Z2Kr8FXZZWOSKVDWTaEcZcOvAnBKZISs8kAhH/3wBzw4SrbocxHw+8i1Xs7ZW0nk
3gZDf+3sRKuKzg1rMf8uO0FE82DB5E5LZSuelEXovF7oqH/Re6jurdsoAdV4umD5
GZM3sC2bLzArwRdXQNgcx5t596eUDzg8AHEsu+tUxblbSR/xeBV7uMKluZVvD+Mf
93LTWZSuHg0bRKY+Kei4F5+thEq39MA4J2/07Rz/s3AgVYHNX+O7K//AuJvNc3Gl
C8hLkl2XsUZPlEuaCveuJahb/U84q+Bi2+6albP/hRzZ+Gsf8OrY8vksBRgy//Ew
YVPz9lZn6kiJVGJkVAvdAqbKpJrMBADBFJM8twzjEXD/b0oAK/oiCx2c136dAta7
zWB4izhuT2waShV0Fa6LIQG41tYMMPw/2l/4QMH92mFUt88aJE20wP3c92xIWZ7j
N6n2XZrU9R5zuhTMvbc6EZTS80dG0Cu07VJdJs7TFSIViqJZ4b Ke547VcnhjbSBM
r8DZTi9s4nbR8N7igN8q3sr/o3fUsuHh0hMFgUe/HKauHBliiTTML2wwXtZjkXWz
kNFxOQgozKg3BMyXNggc
=IwIg
-----END PGP SIGNATURE-----

10-30-2011, 09:50 PM

On 10/31/2011 12:37 AM, Markos Chandras wrote:
> On 10/30/2011 10:33 PM, Petteri Räty wrote:
>> On 27.10.2011 2.40, Mike Frysinger wrote:
>>> i can't see any ebuild/eclass using egethome, egetshell,
>>> is-login-disabled from portability.eclass. anyone have a reason
>>> for keeping these before i punt them ? -mike
>>>
>
>> Breaking overlays. Isn't the standing policy still to not break
>> backwards compatibility as long as an eclass exists?
>
>> Regards, Petteri
>
>
> We have no have direct control on overlays. It is up to maintainers to
> ensure that their packages work with portage eclasses or they can have
> their own eclasses locally.

There could be an deprecation warning that the functions are going away
to be nice with overlay maintainers but there is no requirement for such
thing far as I'm aware.

Punt away.

- Samuli

10-30-2011, 11:17 PM

On Sunday 30 October 2011 18:33:51 Petteri Räty wrote:
> On 27.10.2011 2.40, Mike Frysinger wrote:
> > i can't see any ebuild/eclass using egethome, egetshell,
> > is-login-disabled from portability.eclass. anyone have a reason for
> > keeping these before i punt them ?
>
> Breaking overlays. Isn't the standing policy still to not break
> backwards compatibility as long as an eclass exists?

do you have any actual examples ?
-mike

10-31-2011, 03:44 AM

On Sun, Oct 30, 2011 at 08:17:47PM -0400, Mike Frysinger wrote:
> On Sunday 30 October 2011 18:33:51 Petteri Räty wrote:
> > On 27.10.2011 2.40, Mike Frysinger wrote:
> > > i can't see any ebuild/eclass using egethome, egetshell,
> > > is-login-disabled from portability.eclass. anyone have a reason for
> > > keeping these before i punt them ?
> >
> > Breaking overlays. Isn't the standing policy still to not break
> > backwards compatibility as long as an eclass exists?

Not that I'm aware of; we only have to worry about the main tree.
Overlays are on their own.

William

10-31-2011, 05:56 AM

On 10/30/11 23:33, Petteri Räty wrote:
> On 27.10.2011 2.40, Mike Frysinger wrote:
>> i can't see any ebuild/eclass using egethome, egetshell,
>> is-login-disabled from portability.eclass. anyone have a reason for
>> keeping these before i punt them ?
>> -mike
>>
>
> Breaking overlays.

grepping over all overlays that I can checkout I find only the openbsd
overlay having a copy of the eclass, but zero ebuilds using it

> Isn't the standing policy still to not break
> backwards compatibility as long as an eclass exists?
>
> Regards,
> Petteri
>

10-31-2011, 04:40 PM

On Wednesday 26 October 2011 19:40:24 Mike Frysinger wrote:
> i can't see any ebuild/eclass using egethome, egetshell,
> is-login-disabled from portability.eclass. anyone have a reason for
> keeping these before i punt them ?

http://sources.gentoo.org/eclass/user.eclass?r1=1.3&r2=1.4
http://sources.gentoo.org/eclass/portability.eclass?r1=1.18&r2=1.19
-mike

11-02-2011, 11:13 PM

On Wednesday 26 October 2011 19:40:24 Mike Frysinger wrote:
> i can't see any ebuild/eclass using egethome, egetshell,
> is-login-disabled from portability.eclass. anyone have a reason for
> keeping these before i punt them ?

hmm, seems a few packages in the tree want this functionality. but they've
been calling getent directly and parsing the result rather than using the
helper funcs. so i've cleaned up & restored the relevant funcs, and converted
the ebuilds in the tree that want this info to use these rather than getent.
-mike