1 |
Compiler-config is pretty much 98% done, but I haven't had time to work |
2 |
on it these past few weeks because I'm in mid semester crunch mode. I |
3 |
was hoping to get it done before school got busy, but that's not going |
4 |
to happen. I'd like to solicit some help finishing it up, and since |
5 |
most of the remaining issues are pertinent to multilib toolchains, I |
6 |
figured it'd be best to try here than -dev. |
7 |
|
8 |
The remaining issues before release are: |
9 |
|
10 |
Fixup the migration code to intelligently choose the correct profile. |
11 |
Currently, it just deletes all the ones installed by toolchain.eclass, |
12 |
converts the gcc-config-1.X profiles (so it knows the profile names) and |
13 |
activates the one corresponding to the old profile. This isn't so great |
14 |
on multilib systems because it destroys the alternative target profiles |
15 |
(like i686 in our case). Additionally, the migration script should |
16 |
activate profiles for the alternative targets. |
17 |
|
18 |
toolchain.eclass needs to be updated to properly set the new profile for |
19 |
all targets of the toolchain, not just the current one. |
20 |
|
21 |
the eselect module could do with some cleanup... especially user |
22 |
feedback. |
23 |
|
24 |
Any help would be greatly appreciated... especially since I was hoping |
25 |
to get this done in time for use in 2006.0, then the corresponding |
26 |
binutils tool for 2006.1. |
27 |
|
28 |
Thanks, |
29 |
Jeremy |