1 |
On Tue, 25 Feb 2014 17:28:41 +0200 |
2 |
Samuli Suominen <ssuominen@g.o> wrote: |
3 |
|
4 |
> |
5 |
> On 20/02/14 18:27, Tom Wijsman wrote: |
6 |
> > On Thu, 20 Feb 2014 11:26:18 +0200 |
7 |
> > Samuli Suominen <ssuominen@g.o> wrote: |
8 |
> > |
9 |
> >> On 20/02/14 10:47, Steev Klimaszewski wrote: |
10 |
> >>> On Thu, 2014-02-20 at 10:40 +0200, Samuli Suominen wrote: |
11 |
> >>>> On 20/02/14 09:44, Steev Klimaszewski wrote: |
12 |
> >>>>> On Thu, 2014-02-20 at 07:55 +0200, Samuli Suominen wrote: |
13 |
> >>>>>> On 20/02/14 00:23, Ulrich Mueller wrote: |
14 |
> >>>>>>> Following up to today's QA meeting: The gtk3 USE flag is used |
15 |
> >>>>>>> by 27 packages, so I suggest making it a global flag: |
16 |
> >>>>>>> |
17 |
> >>>>>>> gtk3 - Add support for x11-libs/gtk+ (The GIMP Toolkit) |
18 |
> >>>>>>> version 3 |
19 |
> >>>>>>> |
20 |
> >>>>>>> Ulrich |
21 |
> >>>>>> that would suggest it's fine to use, and is anything but |
22 |
> >>>>>> temporary |
23 |
> >>>>>> |
24 |
> >>>>>> -1 from here |
25 |
> >>>>>> |
26 |
> >>>>> MATE desktop (which I hope to bring in to Portage soon) can be |
27 |
> >>>>> built against gtk+ 2 or gtk+ 3, and upstream supports doing |
28 |
> >>>>> both, so +1 from me. Just because gtk+ 3 is the latest, does |
29 |
> >>>>> not mean it's the greatest, and I really wish people would |
30 |
> >>>>> realize that newest != bestest. |
31 |
> >>>>> |
32 |
> >>>>> |
33 |
> >>>> Then you pick whatever is best supported for MATE, and ship it |
34 |
> >>>> using that. Later when they have completed their support for |
35 |
> >>>> GTK+-3, and it's the best supported, you ship that. It's not |
36 |
> >>>> rocket science. |
37 |
> >>>> |
38 |
> >>> OR, since I'm the maintainer, I decide that I'm willing to deal |
39 |
> >>> with both, instead of you telling me that I need to pick one or |
40 |
> >>> the other. Upstream says both are supported and viable, and I'm |
41 |
> >>> willing to deal with the headaches. Just because you're |
42 |
> >>> unwilling doesn't mean others aren't. kthx. |
43 |
> >>> |
44 |
> >>> |
45 |
> >> Bye bye distribution level consistency :-( |
46 |
> >> |
47 |
> >> It's sad that few stubborn developers can do that. |
48 |
> >> |
49 |
> >> - Samuli |
50 |
> >> |
51 |
> > "'Ey! Have you heard about it. Gentoo doesn't provide X with support |
52 |
> > for Y, then what are their USE flags even for; what a shame, ..." |
53 |
> > |
54 |
> > If people want to support and use multiple things, let them do so. |
55 |
> > It is pretty much what Gentoo and its philosophy are about; which |
56 |
> > somewhat can be summarized as providing choices such that we fit |
57 |
> > the users' need, and not force our one true way upon them... |
58 |
> > |
59 |
> > Greetings from someone who runs GNOME 3 and MATE simultaneously; |
60 |
> > you can intentionally break it, but why would you? It takes away |
61 |
> > our happiness. On the other hand, there's the part where you want |
62 |
> > to break it for a reason, perhaps for your happiness; but then I'd |
63 |
> > like to hear why. |
64 |
> > |
65 |
> |
66 |
> So, no more setting USE="gtk" and assuming the best packaged software |
67 |
> will be get installed, be it with what version of the toolkit, 1, 2 |
68 |
> or 3 |
69 |
|
70 |
What is "best"? What you would deem best, could be worst for the user. |
71 |
|
72 |
> Instead, now you have to selectively do the maintainers job for |
73 |
> figuring out which one is the best supported one |
74 |
|
75 |
The maintainer can use IUSE flag defaults for this; but, there are |
76 |
users that want to control which toolkit is installed. Going further, |
77 |
they want to decide which toolkit the software is built with. |
78 |
|
79 |
> Despite already picking up a modern theme with both GTK+ 2.x and GTK+ |
80 |
> 3.x looks, now you might end up with half-crippled software just |
81 |
> because some stubborn people choose the looks, not the functionality, |
82 |
> to be their motivation |
83 |
|
84 |
You need to mask that as a packager. |
85 |
|
86 |
> Such people really don't deserve to own a packager status if they |
87 |
> can't take the time to determine / examine the package's best |
88 |
> supported graphical toolkit |
89 |
|
90 |
What is "best"? If upstream provides both and claims they both are |
91 |
well supported, I consider both are. |
92 |
|
93 |
> If multiple ones with similar feature set is supported, then the |
94 |
> latest toolkit is preferred |
95 |
|
96 |
What is "preferred"? What you prefer can be different than the user. |
97 |
|
98 |
> But seems like I'm repeating common sense which the GNOME guideline |
99 |
> layed out long ago |
100 |
|
101 |
The GNOME guideline fits just the GNOME team and developers whom follow |
102 |
that; however, in the bigger picture, there are maintainers that do not |
103 |
follow that and do something different. Looking at the tracker [1]; it |
104 |
appears that it is more common for such a bug to be marked as INVALID |
105 |
or WONTFIX, than it is to be marked as FIXED. |
106 |
|
107 |
As a result of this, the users get an inconsistent USE flag presented. |
108 |
|
109 |
The "common sense" is just another wording of "opinion" here; if it |
110 |
wants to be real "common sense" shared amongst almost everyone, it'll |
111 |
need a bit more than a guideline. |
112 |
|
113 |
[1]: |
114 |
https://bugs.gentoo.org/buglist.cgi?quicksearch=ALL%20blocked%3A420493 |
115 |
|
116 |
-- |
117 |
With kind regards, |
118 |
|
119 |
Tom Wijsman (TomWij) |
120 |
Gentoo Developer |
121 |
|
122 |
E-mail address : TomWij@g.o |
123 |
GPG Public Key : 6D34E57D |
124 |
GPG Fingerprint : C165 AF18 AB4C 400B C3D2 ABF0 95B2 1FCD 6D34 E57D |