1 |
Jason Stubbs wrote: |
2 |
> A. Khattri wrote: |
3 |
> |
4 |
>> On Thu, 1 Dec 2005, Robert Larson wrote: |
5 |
>> |
6 |
>>> If I understand you correctly, then this may help: |
7 |
>>> # obtain a list of installed packages |
8 |
>>> emerge -vep world | grep ^'\[ebuild ' | while read a; do |
9 |
>>> # extract the package name |
10 |
>>> pkg=$(echo $a | cut -d']' -f2 | awk '{print $1}') |
11 |
>>> # build a binary of the package |
12 |
>>> quickpkg "/var/db/pkg/${pkg}" |
13 |
>>> done |
14 |
>> |
15 |
>> |
16 |
>> I came up with this (yes, granted it is tied to the file-system layout |
17 |
>> but |
18 |
>> its damn fast!): |
19 |
>> |
20 |
>> find /var/db/pkg -mindepth 2 -maxdepth 2 -print -exec quickpkg {} \; |
21 |
> |
22 |
> |
23 |
> I was thinking about extending quickpkg to support syntax like: |
24 |
> quickpkg [--deep [--build-deps] ] [--overwrite] <atom>|<target> |
25 |
> |
26 |
> <atom> is one or more regular atoms. <target> is any supported target |
27 |
> (currently "world" or "system"). --deep would recursively make packages |
28 |
> of runtime dependencies. --build-deps would include build time |
29 |
> dependencies (which of course aren't actually needed). --overwrite |
30 |
> would cause packages to be created even if they already exist in PKGDIR. |
31 |
> |
32 |
> Can I take this thread as an expression of interest in the above? |
33 |
> Anything missing or would be better done differently? |
34 |
|
35 |
Yes! |
36 |
|
37 |
And, but not sure how, adding something like --update and --update-only will save some cpu cycles. |
38 |
Actually, may be you are right. --overwrite might be better, but I was thinking about emerge syntax. |
39 |
|
40 |
Kalin. |
41 |
|
42 |
-- |
43 |
|[ ~~~~~~~~~~~~~~~~~~~~~~ ]| |
44 |
+-> http://ThinRope.net/ <-+ |
45 |
|[ ______________________ ]| |
46 |
|
47 |
-- |
48 |
gentoo-server@g.o mailing list |