1 |
Fixed now (I think). |
2 |
BTW, it's hard to use a binary installer inside portage - it's not set |
3 |
up for interactivity. I figured out that tailing off the installer |
4 |
script at the start leaves the tbz2 intact, which has the added |
5 |
advantage of letting us use the rest of the x86 ebuild without changes. |
6 |
I don't know why Blackdown doesn't distribute a tbz2 for ppc, but since |
7 |
they do so for x86, I think this is a fair hack (licence is still in the |
8 |
tbz2). The QT ebuilds deal with interactivity by just sed-ing the |
9 |
script to include the required answers, which is what I started trying |
10 |
to do; but then this seemed a lot easier. |
11 |
|
12 |
David |
13 |
|
14 |
Mark Guertin wrote: |
15 |
|
16 |
>Hello All |
17 |
> |
18 |
>I noticed last night that the blackdown jre (dev-java) build is broken, |
19 |
>someone might want to have a look at it. They dist their file as a .bin |
20 |
>file, that is self extracting bz2, so it's a little strange to start with, |
21 |
>but it looks like someone has done some weird things in that build and |
22 |
>didn't finish the ARCH stuff. This and it doesn't use $ARCH (which in my |
23 |
>experiments seems to work ok in that script). |
24 |
> |
25 |
>For example the SOURCE_URI doesn't pull the correct file for ppc... they use |
26 |
>${FCS}/${MY_P}.tar.bz2 when according to their setup it should use |
27 |
>${FCS}/${A}. It almost looks as though someone pushed their personal ebuild |
28 |
>file to CVS with some strange hacks at beginning. Also they should be using |
29 |
>${A} to extract the file as well instead of ${MY_P}.tar.bz2. Is it possible |
30 |
>to have the build just execute the file it downloaded as that is the proper |
31 |
>way to install from their readme.. it's self extracting to the same dir it |
32 |
>is in. |
33 |
> |
34 |
>Mark |
35 |
> |
36 |
> |
37 |
> |