| 1 |
I finally got a Crystal Space ebuild that I can deal with. Unf. there is |
| 2 |
one step I need assistance with. |
| 3 |
|
| 4 |
When issuing a "make install", the system looks at either the default |
| 5 |
destination directory or the one you set. I've been trying |
| 6 |
/usr/local/crystal and if I use that, then "make install" gets a sandbox |
| 7 |
violation. If, however, I use ${D}/usr/local/crystal then it installs |
| 8 |
but right before that step it rewrites the cs-config script (like |
| 9 |
sdl-config, gtk-config, etc.) to say that the location of the CS libs |
| 10 |
and headers is /var/tmp/portage/crystal-0.94-r2/image//usr/local/crystal |
| 11 |
(or something like that, I don't have it in front of me right now). Of |
| 12 |
course, that'll be a bogus location when the package is actually |
| 13 |
installed. |
| 14 |
|
| 15 |
Do I need to do some sed magic to fix that somehow, by twiddling with |
| 16 |
the cs-config file that's been installed in the image? Or is there some |
| 17 |
sort of standard clever way to get around this problem with |
| 18 |
non-./configure-style packages? |
| 19 |
|
| 20 |
Random other info: |
| 21 |
Meanwhile, I'm still in a position where I can't build optimized builds |
| 22 |
with gcc 3.x and I can't use assembler at all. I've been told that CS |
| 23 |
requires a special version of nasm or something like that, but haven't |
| 24 |
found a direct reference for what works yet. Building a debug build with |
| 25 |
no assembler gets me a working install image, I believe, on gcc 3.x and |
| 26 |
gcc 2.95. |
| 27 |
|
| 28 |
Regards and thanks, |
| 29 |
matt |
| 30 |
|
| 31 |
-- |
| 32 |
Matt Matthews \ ph: 919.660.2811 \ Use GNU/Linux _o) w00t |
| 33 |
Duke Univ., Postdoc\ jvmatthe@×××××××××.edu \____________ /\\ |
| 34 |
Dept. of Mathematics\ http://www.math.duke.edu/~jvmatthe/ \ _\_V |
Archives