1 |
On 26 June 2011 21:53, Patrick Lauer <patrick@g.o> wrote: |
2 |
> |
3 |
> I disagree. If I put postgresql in x11-libs that's just wrong, and then |
4 |
> you fix it with a package move. Doesn't mean the category system is |
5 |
> broken, just means that it was in the wrong place at the wrong time. |
6 |
> |
7 |
>> |
8 |
>> As far as app-xemacs is concerned (and probably why you commented here), it |
9 |
>> should be sufficient to prepend "xemacs-" to package names from app-xemacs |
10 |
>> category in order to make them distinguished from the rest. |
11 |
>> It would be elegant and correct - after all when you "emerge ocaml" you don't |
12 |
>> expect to be installing objective caml mode for Emacs, but ocaml interpreter |
13 |
>> itself. |
14 |
> |
15 |
> Please don't do that. It leads to the kind of funny package names that |
16 |
> some legacy distros carry around, which make it exquisitely hard to |
17 |
> guess what their intent is. |
18 |
> |
19 |
> Right now "emerge openoffice" does what I want. Don't change that to |
20 |
> "openoffice-core-core" or other aneurisms that can be found in the wild. |
21 |
> Do not try to make me do extra work, that's rude and inconsiderate :) |
22 |
> |
23 |
>> |
24 |
> And actually we do have unique package names, just that we don't |
25 |
> obfuscate with random distractions but with a mostly working category |
26 |
> system. So think what you are trying to fix, and no, XML is not the |
27 |
> answer :) |
28 |
> |
29 |
|
30 |
I have to agree for the most part, I mean, ultimately were' going to |
31 |
have a bit of "convention" for the sake of convenience. |
32 |
|
33 |
If you want to see what a flat portage system would look like, just do |
34 |
ls /usr/bin/ |
35 |
|
36 |
A lot of thing that are perl modules will inevitably, in a flat |
37 |
topology, be prefixed with "perl-" , like so many other distros do ( |
38 |
and its ugly , debian's liblocal-libperl for local::lib is an |
39 |
example of this catastrophe ) |
40 |
|
41 |
Why do we need to devine a /new/ convention for giving packages a |
42 |
unique name when we have one that is *presently* working just fine. ( |
43 |
for the purpose of uniquely defining a package that is, its failing in |
44 |
a few places for reasons related to discoverability, but I'm no longer |
45 |
arguing that ). |
46 |
|
47 |
the difference between |
48 |
|
49 |
dev-perl-local-lib |
50 |
|
51 |
and |
52 |
|
53 |
dev-perl/local-lib |
54 |
|
55 |
is really not that big in practice, but one sucks infinitely more than |
56 |
the other. |
57 |
|
58 |
With this "new" system however, pkgmoves will be a thing of the past, |
59 |
even if we keep the legacy category system around. |
60 |
|
61 |
pkgmoves take place in my understanding, largely because the category |
62 |
they're in makes them proves hard to find for discoverability |
63 |
purposes. |
64 |
|
65 |
In the new system, categories are no longer part of the discovery |
66 |
system, and exist soley as a method to keep the files organised with |
67 |
distinct names. |
68 |
|
69 |
ie: Currently, were' proposing moving some things around , splitting |
70 |
media-sound into several sub categories, ie: |
71 |
|
72 |
sound-players (and maybe sound-radio) |
73 |
sound-editors |
74 |
sound-plugins |
75 |
sound-libs |
76 |
|
77 |
"Flattening" does not in any way solve this issue in the least, |
78 |
because in the case of a flattening, the difference would be primarily |
79 |
that you have a load of packages with unuseful prefixes instead of a |
80 |
directory, so you'd be doing |
81 |
|
82 |
pkgmove media-sound-amarok sound-players-amarok |
83 |
|
84 |
and you would still have to go and update all the dependencies etc. |
85 |
|
86 |
And I think you'll agree that that sounds a little daft. |
87 |
|
88 |
So in the new system, you put a package in a category that makes some |
89 |
degree of sense, and then leave it there, for all eternity, and rely |
90 |
on tags to make them discoverable instead of moving things around or |
91 |
renaming them. |
92 |
|
93 |
> Right now "emerge openoffice" does what I want. Don't change that to |
94 |
> "openoffice-core-core" or other aneurisms that can be found in the wild. |
95 |
> Do not try to make me do extra work, that's rude and inconsiderate :) |
96 |
|
97 |
I'd hope that with a new tag based system, you could "emerge |
98 |
openoffice" and get told "Yeah, I know that once upon a time there was |
99 |
only 1 openoffice, but now there's more than one, did you mean |
100 |
openoffice-org or libreoffice?" |
101 |
|
102 |
Same goes for the various forms of virtualbox ( did you mean |
103 |
virtualbox from source, virtualbox binary, the opensource edition or |
104 |
the commercial edition etc ) java ( did you mean 'sun-jdk? |
105 |
blackdown-jdk? ' etc ) flash ( did you mean adobe-flash? ) xine ( |
106 |
did you mean xine-ui or xine-lib? , or perhaps even gxine ? ) and a |
107 |
host of other things that are presently like this. |
108 |
|
109 |
|
110 |
|
111 |
-- |
112 |
Kent |
113 |
|
114 |
perl -e "print substr( \"edrgmaM SPA NOcomil.ic\\@tfrken\", \$_ * 3, |
115 |
3 ) for ( 9,8,0,7,1,6,5,4,3,2 );" |
116 |
|
117 |
http://kent-fredric.fox.geek.nz |