1 |
On Sun, 20 Mar 2005 22:01:18 +0100 |
2 |
Paul de Vrieze <pauldv@g.o> wrote: |
3 |
|
4 |
> I'll add some nice tag to support this. |
5 |
|
6 |
I'm sorry to insist, but is there something wrong with using |
7 |
use.local.desc, as i've proposed in bug #84884? I don't understand |
8 |
the rational for choosing metadata.xml. In my opinion, it makes |
9 |
this detailed description less likely to be written (more |
10 |
complex syntax, whereas the one of .desc file is already |
11 |
well-known), and less likely to be read (lack of user tools |
12 |
whereas there are plently for use.local.desc, which would need at |
13 |
most small trivial changes, if any). |
14 |
|
15 |
Also, take the following scenario: |
16 |
|
17 |
- use.local.desc has: |
18 |
"cat/pkgA:foo - adds support for libfoo as a replacement of \ |
19 |
libbar. Do not enable it but if you really know what you do." |
20 |
|
21 |
- then comes cat/pkgB, which also support libfoo: |
22 |
"cat/pkgB:foo - adds support for libfoo, for playing .foo files" |
23 |
|
24 |
- And more packages start supporting libfoo, thus "foo" becomes |
25 |
a global flag: |
26 |
"foo - adds support for libfoo" |
27 |
With the "metadata.xml" approach, that will require moving the |
28 |
existing descriptions from use.local.desc to the various xml |
29 |
files. At the contrary, with the "keep it in use.local.desc" |
30 |
approach, no additional work is required... |
31 |
|
32 |
|
33 |
So could someone explain me that choice? |
34 |
|
35 |
Thanks, |
36 |
|
37 |
-- |
38 |
TGL. |
39 |
-- |
40 |
gentoo-dev@g.o mailing list |