1 |
Andrew wrote: |
2 |
|
3 |
> Relative paths don't really work unless you specify somewhere what they are |
4 |
> relative to (current dir isn't really a good option). For example, relative |
5 |
> paths work in the apache config, because the config root is specified as |
6 |
> /usr/lib/apache2/conf (which is a symlink to /etc/apache2). |
7 |
> |
8 |
> With catalyst, there are so many paths that could be the "root" |
9 |
> (/var/tmp/catalyst, /etc/catalyst, some random dir where you store your |
10 |
> specs, |
11 |
> etc.) that it just doesn't work. |
12 |
|
13 |
What I am in the process of doing is putting together a pre-configured |
14 |
Catalyst-based LiveCD build environment that is newbie-friendly and housed in a |
15 |
CVS. When a Catalyst newbie checks this build environment out of the CVS, I |
16 |
was hoping to have the Catalyst configuration file and the spec files already |
17 |
pointing to the resources they need to ( which are also present in the checked |
18 |
out environment ). |
19 |
|
20 |
The ability to use relative paths in the Catalyst configuration file and the |
21 |
spec files seems like it would work for this use. |
22 |
|
23 |
I have experimented with placing "print os.getcwd()" in the catalyst2 script |
24 |
and it indicates that the current working directory is set to whatever |
25 |
directory this script is launched from. |
26 |
|
27 |
This is the behavior I am interested in using because, after checking the |
28 |
LiveCD build environment out of the CVS, I would like the user to change into a |
29 |
specific directory which contains short scripts that will allow them to build |
30 |
various targets. |
31 |
|
32 |
I am currently experimenting with removing "expand=0" from the calls to the |
33 |
file_locate function on lines 159 and 162 in generic_stage_target.py to see if |
34 |
this breaks anything. |
35 |
|
36 |
Thanks :-) |
37 |
|
38 |
Ted Kosan |
39 |
tkosan@××××××××.net |
40 |
|
41 |
-- |
42 |
gentoo-catalyst@g.o mailing list |