1 |
On Sat, Sep 10, 2011 at 1:22 AM, Markos Chandras <hwoarang@g.o> wrote: |
2 |
> -----BEGIN PGP SIGNED MESSAGE----- |
3 |
> Hash: SHA512 |
4 |
> |
5 |
> On 10/09/2011 11:15 πμ, Mike Gilbert wrote: |
6 |
>> On Sat, Sep 10, 2011 at 4:04 AM, Markos Chandras |
7 |
>> <hwoarang@g.o> wrote: |
8 |
>>> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512 |
9 |
>>> |
10 |
>>> On 10/09/2011 11:01 πμ, Michał Górny wrote: |
11 |
>>>> Hello, |
12 |
>>>> |
13 |
>>>> Is there a real reason to have <herd>no-herd</herd>? As I see |
14 |
>>>> it, it's just an ugly hack which all programs have to learn and |
15 |
>>>> hack for no benefit. |
16 |
>>>> |
17 |
>>> What is the problem with that? Why is it an ugly hack? |
18 |
>>> |
19 |
>> |
20 |
>> To a simple program/user, it basically says that the package |
21 |
>> belongs to a herd called "no-herd", rather than not belonging to |
22 |
>> any herd. |
23 |
> Well it is pretty obvious that no-herd means errr no herd. |
24 |
> |
25 |
>> |
26 |
>> For example, in IRC, willikins tries to look up the members of |
27 |
>> "no-herd" when you do !meta -v because the bot lacks a special |
28 |
>> exception for this odd-ball value. |
29 |
> Just because willikins behaves like that does not justify the removal |
30 |
> of this tag. God knows how many utilites and scripts are out there |
31 |
> using the no-herd tag from metadata. |
32 |
|
33 |
I want to comment briefly here because I think this affects the |
34 |
project a fair bit and I'm not talking specifically about |
35 |
metadata.dtd. Projects need to be able to fix mistakes made in the |
36 |
past. Backwards compatibility is important but it is not some holy |
37 |
grail that we should require all the time. Sometimes the choices we |
38 |
made were wrong and we should be free to fix them. |
39 |
|
40 |
We see this in upstream packages when ABI changes happen. Sometimes an |
41 |
old function was written poorly, or is no longer needed. Libraries |
42 |
change, packages are broken, and then they are fixed. |
43 |
|
44 |
I don't mean to say 'never worry about breaking stuff.' In this |
45 |
particular case I'm sure if important tools are broken by removing |
46 |
no-herd they will be fixed (because important tools need to work...as |
47 |
it turns out.) I find it difficult to conceive of a situation where a |
48 |
machine would be broken by this change; so I am less worried about it. |
49 |
|
50 |
-A |
51 |
|
52 |
> |
53 |
>> |
54 |
>> I would say that each package needs to have at least one herd or |
55 |
>> maintainer (which may be maintainer-needed or maintainer-wanted). |
56 |
>> |
57 |
> Well, you can easily assign your packages to dozen of herds and still |
58 |
> be the only one who touches them. So what's the point? Just to pretend |
59 |
> that the package is supported by an entire herd? |
60 |
> |
61 |
> - -- |
62 |
> Regards, |
63 |
> Markos Chandras / Gentoo Linux Developer / Key ID: B4AFF2C2 |
64 |
> -----BEGIN PGP SIGNATURE----- |
65 |
> Version: GnuPG v2.0.18 (GNU/Linux) |
66 |
> |
67 |
> iQIcBAEBCgAGBQJOax44AAoJEPqDWhW0r/LCdjMP/3pJ2vV9tkQcwwga+pocB6U8 |
68 |
> rpIm5nwEKUoTLucVUbyLuF0X8Z3u3RtAF1aJcq0UDN8EhBtzaAPFnvcRekqDF2R2 |
69 |
> 3Col31hyPUFGCXwyij8WSYwoTcVWYLwRw8m8jBBcXQmM2sLXCF1BNZQ83m0IAFv8 |
70 |
> l73nZkHclebxZ+Pxs6/1nZZ/ZOUHTAT/SaI+pqkDKu64PbEjIKoxhmW5b2vbEPcQ |
71 |
> gry3w/lsZsFLNxOM31H4/HkkB4HrOXBH9e8IKTO1dzZIWKWBA3ytwJFf78kWMNDo |
72 |
> pz3d++nzw6o0gc0Rr+n2b+qL1gy1FPcM7qzRuZ2kESCvmrMW+HRqIrylSGcToJBt |
73 |
> pKP+xF1h6/EJVa1IoQq0IvETms82nPnmV+uRoF5RZDD2gvQ5LIgQCtesHFEDAg8c |
74 |
> WyBV1WmGiMg235XQYg8qASfz4VAUuiHqPLtJ/SkyD78CqPNsVf+Ak6t8HLoF+v+d |
75 |
> 0SJGzC+CbvMJoSXWHcX9pHx1fi3Enw6OziKm1cGzu22TcxCyuMTxGvk9kyfCzT6L |
76 |
> QfQ0WQFDS8CO26L/O+uah/boeGYFvNciaEb7m+tQkCEyAIPbNXPFMdz6XzroOStj |
77 |
> e5iyj4mpq59MLD/6ULX04mb/4r/3OI3c9f2GfwZOCes+cjTmesMYeHOcMfpiqXIn |
78 |
> 9Z+DzlydAJ4Xy7iYtIbg |
79 |
> =jJ/i |
80 |
> -----END PGP SIGNATURE----- |
81 |
> |
82 |
> |