1 |
* Fernando Canizo <conan@××××××××××.ar> [24/08/05 14:07]: |
2 |
> |
3 |
> El 24/ago/2005 a las 04:01 -0300, Jonas me decía: |
4 |
> > >Second, the whole idea is to do this for non-connected systems. Meaning, a |
5 |
> > >solution not involving the Internet... =) |
6 |
> > If you're refering to the fact that the person doesn't have an internet |
7 |
> > connection when he wants to view this information throught the emerge |
8 |
> > interface, then you're wrong. When you want to install or you're just |
9 |
> > searching for a package you're using the emerge interface , when you |
10 |
> > need more info you have to 1) open your browser 2) type in the url (if |
11 |
> > you're lucky you know the url or it's in your browsers cache) 3) search |
12 |
> > for the package on the website , while instead you could just do |
13 |
> > something like emerge --desc package. Now what's quicker and makes more |
14 |
> > sence ? |
15 |
> |
16 |
> Exactly! That's what i'm refering to. It's not an issues of connection it's an |
17 |
> issue of lazyness! It's like: "-- hey what are you doing this weekend? -- don't |
18 |
> know, it's cold and raining, i think i'm gonna sit in my throne and check what's |
19 |
> new on 'app-vim' to improve my vimyness, if find something interesting i'll |
20 |
> check the package webpage." |
21 |
> |
22 |
> For now i'm using 'questo': |
23 |
> === script ==== |
24 |
> #!/bin/bash |
25 |
> # lun may 2 20:57:24 ART 2005 |
26 |
> # conan - GPLed |
27 |
> # |
28 |
> # script to check for apps on rainy sunday morning |
29 |
> |
30 |
> BASE_DIR="/usr/portage/" |
31 |
> |
32 |
> [ $# -ne 1 ] && echo "Uso: questo <category>" && exit 0 |
33 |
> |
34 |
> cd $BASE_DIR$1 |
35 |
> for package in * ; do |
36 |
> eix "^$paquete\$" |
37 |
> echo "Looking for: $1/$paquete. ENTER to continue..." |
38 |
> read |
39 |
> done |
40 |
> === end script ==== |
41 |
> |
42 |
> Now i took sugestion from Ciaran to look in metadata.xml, i check a couple by |
43 |
> hand and see the 'longdesc' field but... with short descriptions! Now maybe i |
44 |
> just choose two with bad luck, but i got a feeling that gentoo maintainers |
45 |
> doesn't like to provide longdesc, although there is the posibility. |
46 |
> |
47 |
> I was thinking of doing a little script that gives longdesc found in metadata to |
48 |
> confirm this, but since i don't know nothing about xml, think it gonna take a |
49 |
> little bit more that if there were already an application to check |
50 |
> 'metadata.xml' |
51 |
|
52 |
If you have the XML::Simple perl module (which can be emerged as |
53 |
XML-Simple), you can do |
54 |
|
55 |
perl -MXML::Simple -le 'print XMLin(shift)->{"longdescription"}' <file> |
56 |
|
57 |
where <file> is the metadata.xml file. |
58 |
|
59 |
However, it appears that this field is not widely supported. Instead, |
60 |
you might use something like |
61 |
|
62 |
lynx -dump -nolist `esearch -o %h <package>` |
63 |
|
64 |
to dump the homepage of the package as text (of course, you still need |
65 |
to be online for that) |
66 |
|
67 |
Moshe |
68 |
|
69 |
PS. Sorry if this actually seems on-topic :) |