[colord] Add conflict on icc-profiles-openicc

[ku.b-mal]

Hello,

I am the upstream maintainer of a number of colour management packages
including icc-profiles-openicc and new to this list. Due to the
speciality of matters, I jump in here.

Reason for my write:
[colord] Add conflict on icc-profiles-openicc :
https://lists.fedoraproject.org/pipermail/scm-commits/Week-of-Mon-20131209/1...

Problem:
The above Conflict tag breaks several packages, which rely on
icc-profiles-openicc, namely oyranos and depending graphics software.

We have no bug reports upstream, or in openSUSE, where I maintain many
colour management packages, nor see I such in Fedora in order to work
upstream to fix anything. In the past I was able to fix several issues
including those arrised by openSUSE, Fedora, Debian, Gentoo and other
distro packagers.

The Conflict tag causes now uncertaincy to packagers and forces me to
take from my development time on bug reports, finding out how packaging
works in Fedora etc.

Request:
remove the RPM spec Conflict tag immediately as it misuses the GNOME
requirement for colord to block Oyranos from Fedora without fixable
reasoning.

Bug reports:
https://bugzilla.redhat.com/show_bug.cgi?id=1069672
(https://bugzilla.redhat.com/show_bug.cgi?id=1042655)

Background:
colord author Richard Hughes sees himself as a competitor to the elder
Oyranos project, which helped paving the way for linux color management.
Not sure why he continues to repeat FUD[1] all over the place regarding
Oyranos. See the open bug as an example for that. But, as you might
imagine, that feels unfriendly.

kind regards and sorry for the noice
Kai-Uwe

[Rex Dieter]

ku.b-mal wrote:
> Hello,
>
> I am the upstream maintainer of a number of colour management packages
> including icc-profiles-openicc and new to this list. Due to the
> speciality of matters, I jump in here.
>
> Reason for my write:
> [colord] Add conflict on icc-profiles-openicc :
> https://lists.fedoraproject.org/pipermail/scm-commits/Week-of-Mon-20131209/1...
>
> Problem:
> The above Conflict tag breaks several packages, which rely on
> icc-profiles-openicc, namely oyranos and depending graphics software.

I reverted the Conflicts pending some hopefully better solution forthcoming 
in:
https://bugzilla.redhat.com/show_bug.cgi?id=1069672

[Richard Hughes]

On 28 February 2014 17:45, Rex Dieter  wrote:
> I reverted the Conflicts pending some hopefully better solution forthcoming
> in: https://bugzilla.redhat.com/show_bug.cgi?id=1069672

Well, I still stand by my logic; the user does not want two
*different* sRGB system profiles on one system and it's confusing to
have both in dialogs to choose from. The alternative is that I *hide*
at runtime any duplicate system profiles in the daemon.
> Background:
> colord author Richard Hughes sees himself as a competitor to the elder
> Oyranos project, which helped paving the way for linux color management.

I don't care so much for age, but I do care about adoption and users of the API:

http://qa.debian.org/popcon.php?package=colord
http://qa.debian.org/popcon.php?package=oyranos

colord does take some of the specifications you developed for Oyranos
and extends them significantly. I'm not sure how that's relevant here.
> Not sure why he continues to repeat FUD[1] all over the place regarding
> Oyranos.

I think you probably need to link to actual things I've said rather
than overarching statements. I'm trying to help build a cohesive
operating system which doesn't allow swapping of core parts of the
system. I'm happy to work with you on specifications but I'm not happy
supporting an OS where one CMS is interchangeable with another when
the design goals are so different.

Richard.

[ku.b-mal]

Am 28.02.2014 19:03, schrieb Richard Hughes:
> On 28 February 2014 17:45, Rex Dieter  wrote:
> > I reverted the Conflicts pending some hopefully better solution forthcoming
> > in: https://bugzilla.redhat.com/show_bug.cgi?id=1069672
>
> Well, I still stand by my logic; the user does not want two
> *different* sRGB system profiles on one system and it's confusing to
> have both in dialogs to choose from. The alternative is that I *hide*
> at runtime any duplicate system profiles in the daemon.

There are a number of situations your system could cover, in order to
not fail by over simplifying.
Two sRGB profiles in one system is maybe a user decission. Like a user
installs original HP sRGB.icm profile, or that from Argyll or lcms or
the official ICC web site.

You can easily *hide* any profile duplicates. I did so in the past in
Oyranos, but later decided against it for the above reasoning. But that
are design goals and that's entirely colord's and its users decission.

Given that colord is installed by default only for GNOME and otherwise
colord or Oyranos is installed by user or distributor decission, there
are few parallel installs likely.
> > Not sure why he continues to repeat FUD[1] all over the place regarding
> > Oyranos.
>
> I think you probably need to link to actual things I've said rather
> than overarching statements.

https://bugzilla.redhat.com/show_bug.cgi?id=1069672
> I'm happy to work with you on specifications but I'm not happy
> supporting an OS where one CMS is interchangeable with another when
> the design goals are so different.

Irritating attitude for a free and open source software developer. Specs
and software go hand in hand. Different concepts are needed to see what
is inspiring or fits best. One example for that statement is your own
desire to improve over existing software by starting a new project: colord.

Kai-Uwe