1 |
On 12/08/2014 12:45 AM, Ulrich Mueller wrote: |
2 |
>>>>>> On Sun, 07 Dec 2014, Zac Medico wrote: |
3 |
> |
4 |
>> On 12/07/2014 07:06 PM, Christoph Junghans wrote: |
5 |
>>> I know, I am late to the party, I just wanted to say that in |
6 |
>>> unpacker.eclass I implemented a variant which uses neither deb2tags |
7 |
>>> nor ar on prefix, but just bash's read and head. |
8 |
> |
9 |
> The code uses "head -c" which is a GNU extension: |
10 |
> |
11 |
> $ man 1p head |
12 |
> HEAD(1P) POSIX Programmer's Manual HEAD(1P) |
13 |
> [...] |
14 |
> SYNOPSIS |
15 |
> head [−n number] [file...] |
16 |
> [...] |
17 |
> |
18 |
> For example, on OpenBSD and HPUX there is no -c option, or it is used |
19 |
> differently. Also Busybox doesn't have it. |
20 |
> |
21 |
> Are we better off by requiring GNU head instead of GNU ar? |
22 |
|
23 |
That's a good question. We might use dd instead of head, or port the |
24 |
shell code to python. |
25 |
|
26 |
>> Very nice. We should borrow that code and use it to eliminate the |
27 |
>> need for deb2targz in portage. |
28 |
> |
29 |
> In PMS, this would have to go into EAPI 6, I suppose? |
30 |
|
31 |
Yes, I suppose so, since PMS currently says deb2targz is required when |
32 |
ar is not compatible with GNU ar. However, earlier you said that other |
33 |
package managers currently use ar unconditionally. Also, I have to |
34 |
wonder if any of the other package managers are used on AIX. So, maybe a |
35 |
retroactive change to PMS would be acceptable? |
36 |
|
37 |
Maybe it's not worth the trouble, and we should simply keep the |
38 |
deb2targz usage for AIX. |
39 |
-- |
40 |
Thanks, |
41 |
Zac |