[gentoo-dev] Make inotify a global USE flag?

[gentoo-dev] Make inotify a global USE flag?

[Ulrich Mueller]

These packages have inotify as local USE flag:

   app-editors/emacs-vcs
   app-misc/recoll
   app-misc/strigi
   dev-libs/ecore
   dev-vcs/git-annex
   media-sound/mpd
   net-im/ekg2
   net-misc/mediatomb
   sys-fs/aufs2
   sys-fs/aufs3
   sys-process/cronie

Should we make it a global flag? What description is better:

   inotify - Enable inotify filesystem monitoring support
   inotify - Enable inotify file change notification support

BTW, half of the local flag descriptions are "Enable inotify support".
Imagine that! ;-)

Ulrich

Re: [gentoo-dev] Make inotify a global USE flag?

[Samuli Suominen]

On 21/03/13 12:29, Ulrich Mueller wrote:
> These packages have inotify as local USE flag:
>
>     app-editors/emacs-vcs
>     app-misc/recoll
>     app-misc/strigi
>     dev-libs/ecore
>     dev-vcs/git-annex
>     media-sound/mpd
>     net-im/ekg2
>     net-misc/mediatomb
>     sys-fs/aufs2
>     sys-fs/aufs3
>     sys-process/cronie
>
> Should we make it a global flag? What description is better:
>
>     inotify - Enable inotify filesystem monitoring support
>     inotify - Enable inotify file change notification support

+1, ...

... despite that most of these shouldn't even have a USE flag for 
inotify, instead it should always be enabled with 'kernel_linux' and let 
the application itself do a runtime check if inotify is available or not
since these have no external dependencies (other than virtual/os-headers)
not everything should be put behind USE flag :-/

- Samuli

Re: [gentoo-dev] Make inotify a global USE flag?

[Peter Stuge]

Ulrich Mueller wrote:
> Should we make it a global flag?

Sure.

> What description is better:
> 
>    inotify - Enable inotify filesystem monitoring support
>    inotify - Enable inotify file change notification support

The former is more correct, because inotify provides notifications
for more than just file changes.


Samuli Suominen wrote:
> it should always be enabled with 'kernel_linux' and let the 
> application itself do a runtime check if inotify is available or not

I think it's great if you are working on such patches for upstreams!


//Peter

Re: [gentoo-dev] Make inotify a global USE flag?

[Alexander Berntsen]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On 21/03/13 11:29, Ulrich Mueller wrote:
> Should we make it a global flag?
Sure.

> What description is better:
> 
> inotify - Enable inotify filesystem monitoring support inotify - 
> Enable inotify file change notification support
> 
> BTW, half of the local flag descriptions are "Enable inotify 
> support". Imagine that! ;-)
Why not just make it "Enable inotify support", and write more detailed
local flags? The local description should tell you what that
particular ebuild uses the flag for.

- -- 
Alexander
alexander@plaimi.net
http://plaimi.net/~alexander
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iF4EAREIAAYFAlFK844ACgkQRtClrXBQc7UScQD+L0WBiRIIKTa4ifpXGzMK8FUK
fqKpcb0PhKa5gb5KavQA/3CWydciFb75kDtrkUy6nlLpDaOOoE6jB74+1yAQ+uos
=hgwR
-----END PGP SIGNATURE-----

Re: [gentoo-dev] Make inotify a global USE flag?

[Samuli Suominen]

On 21/03/13 13:43, Peter Stuge wrote:
> Ulrich Mueller wrote:
>> Should we make it a global flag?
>
> Sure.
>
>> What description is better:
>>
>>     inotify - Enable inotify filesystem monitoring support
>>     inotify - Enable inotify file change notification support
>
> The former is more correct, because inotify provides notifications
> for more than just file changes.
>
>
> Samuli Suominen wrote:
>> it should always be enabled with 'kernel_linux' and let the
>> application itself do a runtime check if inotify is available or not
>
> I think it's great if you are working on such patches for upstreams!

no, i'm talking from experience -- every package I've seen/touched/... 
with IUSE="inotify" has had code that made it safe to use also for
non-inotify'd kernels
i'm referring to the mistakes done by maintainers by adding 
unnecessarily the flag
cut the sarcasm :-)

- Samuli

Re: [gentoo-dev] Make inotify a global USE flag?

[Peter Stuge]

Samuli Suominen wrote:
>> Samuli Suominen wrote:
>>> it should always be enabled with 'kernel_linux' and let the
>>> application itself do a runtime check if inotify is available or not
>>
>> I think it's great if you are working on such patches for upstreams!
>
> no, i'm talking from experience -- every package I've seen/touched/...
> with IUSE="inotify" has had code that made it safe to use also for
> non-inotify'd kernels
> i'm referring to the mistakes done by maintainers by adding
> unnecessarily the flag

That was not at all clear, but that's great! Then you could fix those
ebuilds. Except there's that rule to not fix bugs in others' ebuilds,
so even though you've found a bug you're not allowed to fix it.. :\


//Peter

Re: [gentoo-dev] Make inotify a global USE flag?

[Sergey Popov]

[-- Attachment #1: Type: text/plain, Size: 502 bytes --]

21.03.2013 16:10, Peter Stuge wrote:
> That was not at all clear, but that's great! Then you could fix those
> ebuilds. Except there's that rule to not fix bugs in others' ebuilds,
> so even though you've found a bug you're not allowed to fix it.. :\

To be honest - last statement is not correct, but, please, do not start
this useless discussion about common sense in touching others' packages
again.

-- 
Best regards, Sergey Popov
Gentoo Linux Developer
Desktop-effects project lead


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 555 bytes --]

Re: [gentoo-dev] Make inotify a global USE flag?

[Samuli Suominen]

On 21/03/13 14:10, Peter Stuge wrote:
> Samuli Suominen wrote:
>>> Samuli Suominen wrote:
>>>> it should always be enabled with 'kernel_linux' and let the
>>>> application itself do a runtime check if inotify is available or not
>>>
>>> I think it's great if you are working on such patches for upstreams!
>>
>> no, i'm talking from experience -- every package I've seen/touched/...
>> with IUSE="inotify" has had code that made it safe to use also for
>> non-inotify'd kernels
>> i'm referring to the mistakes done by maintainers by adding
>> unnecessarily the flag
>
> That was not at all clear, but that's great! Then you could fix those
> ebuilds.

Yes, I could, or I could just let other maintainers know about it, like 
on the ML, wait... that's what I just did

> Except there's that rule to not fix bugs in others' ebuilds,
> so even though you've found a bug you're not allowed to fix it.. :\

You cut down the important bit from my last post about 'cutting the sarcasm'
We really don't need this in the ML
Can we stop it before it begins? I hope so

- Samuli

Re: [gentoo-dev] Make inotify a global USE flag?

[Peter Stuge]

Samuli Suominen wrote:
>>> i'm referring to the mistakes done by maintainers by adding
>>> unnecessarily the flag
>>
>> That was not at all clear, but that's great! Then you could fix those
>> ebuilds.
>
> Yes, I could, or I could just let other maintainers know about it, like
> on the ML, wait... that's what I just did

..using the mailing list as bugtracker?

I completely understand and agree that submitting a bug takes certain
effort, sometimes too much for simple things.

I'm just ridiculing the fact that rules are chilling bugfixes.

Maybe it inspires someone, even if it isn't following the rules for
how to express opinion on rules.


>> Except there's that rule to not fix bugs in others' ebuilds,
>> so even though you've found a bug you're not allowed to fix it.. :\
>
> You cut down the important bit from my last post about 'cutting the 
> sarcasm'
> We really don't need this in the ML
> Can we stop it before it begins? I hope so

The emperor was naked all along. ;)


//Peter

Re: [gentoo-dev] Make inotify a global USE flag?

[Ulrich Mueller]

>>>>> On Thu, 21 Mar 2013, Alexander Berntsen wrote:

>> BTW, half of the local flag descriptions are "Enable inotify 
>> support". Imagine that! ;-)

> Why not just make it "Enable inotify support", and write more
> detailed local flags?

The point is that this so-called description adds no useful
information whatsoever. Most of the time USE="foo" will enable some
"foo" feature.

Ulrich

Re: [gentoo-dev] Make inotify a global USE flag?

[Tom Wijsman]

[-- Attachment #1: Type: text/plain, Size: 1045 bytes --]

On Thu, 21 Mar 2013 14:33:50 +0100
Peter Stuge <peter@stuge.se> wrote:

> ..using the mailing list as bugtracker?

No. In this context we are using the mailing list to discuss a patch
to a bug; this is completely normal use of the mailing list, you can
also see this behavior on other mailing lists, eg. the Linux kernel ML.

> I completely understand and agree that submitting a bug takes certain
> effort, sometimes too much for simple things.

All it takes is a single command line, check out PyBugz...

> I'm just ridiculing the fact that rules are chilling bugfixes.
>
> Maybe it inspires someone, even if it isn't following the rules for
> how to express opinion on rules.

If this bothers you so much, then please start a new separate thread
on it. Bringing this irrelevant matter to several threads is pointless.


With kind regards,

Tom Wijsman (TomWij)
Gentoo Developer

E-mail address  : TomWij@gentoo.org
GPG Public Key  : 6D34E57D
GPG Fingerprint : C165 AF18 AB4C 400B C3D2  ABF0 95B2 1FCD 6D34 E57D

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 490 bytes --]

Re: [gentoo-dev] Make inotify a global USE flag?

[Alexander Berntsen]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On 21/03/13 15:14, Ulrich Mueller wrote:
> Most of the time USE="foo" will enable some "foo" feature.
Most of the time is not all of the time, but I see your point. Maybe
"Enable inotify to notice filesystem changes". Or one of your
suggestions, both of which look adequate. As long as the local
descriptions say whether they are monitoring IN_ACCESS, IN_DELETE,
IN_MODIFY or whatever.

- -- 
Alexander
alexander@plaimi.net
http://plaimi.net/~alexander
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iF4EAREIAAYFAlFLMfgACgkQRtClrXBQc7VlJwD/RzcgbJGicmjwWqOREbEswrQK
yaf8yupQVLbDw2NaDdYA/1WYUfIqYEaxEMbO62kq21gq8f7xOLYPtlw/ypqZDAQ/
=JKsx
-----END PGP SIGNATURE-----

Re: [gentoo-dev] Make inotify a global USE flag?

[Michał Górny]

[-- Attachment #1: Type: text/plain, Size: 1214 bytes --]

On Thu, 21 Mar 2013 12:32:31 +0200
Samuli Suominen <ssuominen@gentoo.org> wrote:

> On 21/03/13 12:29, Ulrich Mueller wrote:
> > These packages have inotify as local USE flag:
> >
> >     app-editors/emacs-vcs
> >     app-misc/recoll
> >     app-misc/strigi
> >     dev-libs/ecore
> >     dev-vcs/git-annex
> >     media-sound/mpd
> >     net-im/ekg2
> >     net-misc/mediatomb
> >     sys-fs/aufs2
> >     sys-fs/aufs3
> >     sys-process/cronie
> >
> > Should we make it a global flag? What description is better:
> >
> >     inotify - Enable inotify filesystem monitoring support
> >     inotify - Enable inotify file change notification support
> 
> +1, ...
> 
> ... despite that most of these shouldn't even have a USE flag for 
> inotify, instead it should always be enabled with 'kernel_linux' and let 
> the application itself do a runtime check if inotify is available or not
> since these have no external dependencies (other than virtual/os-headers)
> not everything should be put behind USE flag :-/

net-im/ekg2 fixed. But you should seriously consider opening a bug, at
least a common one CC-ing all the remaining maintainers.


-- 
Best regards,
Michał Górny

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 966 bytes --]