1 |
On Tue, Feb 3, 2009 at 5:42 PM, Grant <emailgrant@×××××.com> wrote: |
2 |
>>>>>>> When this was asked a few weeks ago someone then asked why |
3 |
>>>>>>> --with-bdeps Y isn't the default? This seems to burn nearly everyone |
4 |
>>>>>>> once in awhile. |
5 |
>>>>>>> |
6 |
>>>>>> Because using --with-bdeps y causes unnecessary compilation of packages |
7 |
>>>>>> that don't need t0 be changed. They won't be used again until the |
8 |
>>>>>> dependent package is updated, so why waste time rebuilding them in the |
9 |
>>>>>> interim? |
10 |
>>>>>> |
11 |
>>>>>> No one really gets "burned" by this, they just wonder why installed |
12 |
>>>>>> packages aren't upgraded, nothing stops working. |
13 |
>>>>>> |
14 |
>>>>> |
15 |
>>>>> I added: |
16 |
>>>>> |
17 |
>>>>> EMERGE_DEFAULT_OPTS="--with-bdeps n" |
18 |
>>>>> |
19 |
>>>>> to make.conf and ran 'emerge --depclean' and it got rid of a bunch of |
20 |
>>>>> stuff, but I'm still confused by boost. --depclean didn't remove it, |
21 |
>>>>> 'emerge -avDuN world' doesn't downgrade it even --with-bdeps y, but |
22 |
>>>>> 'emerge -pv boost' would downgrade it. I also re-emerged twinkle and |
23 |
>>>>> rb_libtorrent which are the packages that depend on boost, but the |
24 |
>>>>> result is the same. |
25 |
>>>>> |
26 |
>>>>> Also man seems to be broken after that --depclean. When I try to use |
27 |
>>>>> it, I get errors starting with: |
28 |
>>>>> |
29 |
>>>>> sh: /usr/bin/unlzma: No such file or directory |
30 |
>>>>> |
31 |
>>>>> - Grant |
32 |
>>>>> |
33 |
>>>>> |
34 |
>>>>> |
35 |
>>>> |
36 |
>>>> |
37 |
>>>> This may help. |
38 |
>>>> |
39 |
>>>> root@smoker / # equery belongs /usr/bin/unlzma |
40 |
>>>> [ Searching for file(s) /usr/bin/unlzma in *... ] |
41 |
>>>> app-arch/lzma-utils-4.32.7 (/usr/bin/unlzma -> lzma) |
42 |
>>>> root@smoker / # |
43 |
>>>> |
44 |
>>>> I would rebuild that or see why it is not already installed. I would |
45 |
>>>> think that would be part of system??? I'm not sure tho. |
46 |
>>> |
47 |
>>> I seem to recall some switch from LZMA to BZ2 manpages in an |
48 |
>>> etc-update recently ... |
49 |
>> |
50 |
>> emerging lzma-utils fixed it, thank you. I always etc-update as soon |
51 |
>> as the packages are built. Should lzma-utils be a dependency of |
52 |
>> something? |
53 |
>> |
54 |
>> - Grant |
55 |
> |
56 |
> Weird, --depclean wants to remove lzma-utils again even though: |
57 |
> |
58 |
> # equery depends lzma-utils |
59 |
> [ Searching for packages depending on lzma-utils... ] |
60 |
> dev-libs/mpfr-2.3.2 (app-arch/lzma-utils) |
61 |
> media-libs/libpng-1.2.34 (app-arch/lzma-utils) |
62 |
> media-libs/netpbm-10.44.00-r1 (app-arch/lzma-utils) |
63 |
> net-dns/dnsmasq-2.45 (app-arch/lzma-utils) |
64 |
> net-misc/netkit-rsh-0.17-r9 (app-arch/lzma-utils) |
65 |
> sys-apps/coreutils-6.10-r2 (app-arch/lzma-utils) |
66 |
> sys-apps/net-tools-1.60_p20071202044231-r1 (app-arch/lzma-utils) |
67 |
> sys-devel/m4-1.4.11 (app-arch/lzma-utils) |
68 |
> sys-kernel/linux-headers-2.6.27-r2 (app-arch/lzma-utils) |
69 |
> sys-libs/gpm-1.20.5 (app-arch/lzma-utils) |
70 |
> |
71 |
> Maybe it's listed as a build-time dependency of coreutils when it |
72 |
> should be runtime? |
73 |
> |
74 |
> - Grant |
75 |
> |
76 |
> |
77 |
|
78 |
coreutils is an lzma archive, so lzma-utils are required to decompress |
79 |
it. So it seems proper that it's a build-time dep. |
80 |
|
81 |
I think there was something about man using lzma IF you had lzma-utils |
82 |
installed at the time of emerging man. So maybe you can try to unmerge |
83 |
lzma-utils, then re-emerge man (and maybe convert your lzma manpages |
84 |
to bz2). |
85 |
|
86 |
Also be sure you've got PORTAGE_COMPRESS set to what you'd like in |
87 |
your make.conf |