1 |
On 07/01/2013 05:18 PM, heroxbd wrote: |
2 |
> Hey Luca, |
3 |
> |
4 |
> Luca Barbato <lu_zero@g.o> writes: |
5 |
> |
6 |
>> On 07/01/2013 04:48 PM, heroxbd wrote: |
7 |
>> |
8 |
>>> Gentoo on Android is about running Gentoo natively in a directory prefix |
9 |
>>> on Android devices[1] in parallel with native Android, mentored by Luca. |
10 |
>>> |
11 |
>>> Progress report for the past 3 days, |
12 |
>>> |
13 |
>>> 1. gcc specs hack is not that interesting as it first looks. |
14 |
>>> - parsing or generating specs script is no fun. |
15 |
>>> The best parser/generator is gcc itself |
16 |
>>> - appending simple specs (suches -dynamic-linker=/xx/xx.so) feel |
17 |
>>> not clever, appending --with-spec to configure either. |
18 |
>>> |
19 |
>>> Therefore, I decide not to use specs hack as the main building |
20 |
>>> block. At the same time, it can be handy to do manual tuning |
21 |
>>> with it. |
22 |
>> |
23 |
>> That must be reconsidered later to understand what's the problem and if |
24 |
>> upstream can help on that. |
25 |
> |
26 |
> Got it. |
27 |
> |
28 |
>>> Relying on binutils wrapper is no good either. The solution |
29 |
>>> finally falls on a mixed one between sysroot and native paths |
30 |
>>> methods (ref. table in [2]). Where I make use of eprefixify of |
31 |
>>> prefix.eclass to modify location of dynamic linker for *run |
32 |
>>> time* and sysroot for *compile time*. |
33 |
>> |
34 |
>> It isn't perfect, but as long it gives result I can live with that. |
35 |
> |
36 |
> The perfect solution is to use --with-runtime-root-prefix contributed by |
37 |
> Google[1][2], but not enough interest was gathered[3]. |
38 |
|
39 |
The patch itself is partially wrong (you might want to use a totally |
40 |
different layout) but surely looks interesting. |
41 |
|
42 |
lu |