| 1 |
Dear Luca, |
| 2 |
|
| 3 |
Luca Barbato <lu_zero@g.o> writes: |
| 4 |
|
| 5 |
> On 07/22/2013 09:19 AM, Benda Xu wrote: |
| 6 |
> |
| 7 |
> You have some flavours, e.g. |
| 8 |
> |
| 9 |
> +#define GLIBC_DYNAMIC_LINKER32 RUNTIME_ROOT_PREFIX "/lib/ld-linux.so.2" |
| 10 |
> +#define GLIBC_DYNAMIC_LINKER64 RUNTIME_ROOT_PREFIX "/lib64/ld-linux-x86-64.so.2" |
| 11 |
> |
| 12 |
> or |
| 13 |
> |
| 14 |
> +#define UCLIBC_DYNAMIC_LINKER RUNTIME_ROOT_PREFIX "/lib/ld-uClibc.so.0" |
| 15 |
> +#define UCLIBC_DYNAMIC_LINKER32 RUNTIME_ROOT_PREFIX "/lib/ld-uClibc.so.0" |
| 16 |
> +#define UCLIBC_DYNAMIC_LINKER64 RUNTIME_ROOT_PREFIX "/lib/ld64-uClibc.so.0" |
| 17 |
> |
| 18 |
> And you add some prefix to it. |
| 19 |
> |
| 20 |
> while the correct patch would just wipe all this cruft and have a |
| 21 |
> |
| 22 |
> DYNAMIC_LINKER_{32/64/default} and a mean to set it. |
| 23 |
> |
| 24 |
> Everything is still not exactly great since you could have more than 2 |
| 25 |
> abis per architecture. |
| 26 |
> |
| 27 |
> As said the patch is surely an improvement over the previous situation |
| 28 |
> but doesn't help solving the problem fully. |
| 29 |
|
| 30 |
Got what you meant. I am just afraid if upstream would reject a mean to |
| 31 |
fully override the dynamic linker for the sake of a /absolutely stable |
| 32 |
soname/. |
| 33 |
|
| 34 |
An alternative: Given the similar situation in binutils, I am thinking |
| 35 |
of another switch (like --{enable,with}-native-sysroot) to turn |
| 36 |
--with-sysroot into a native (non-cross) version for our purpose. (GLEP |
| 37 |
draft follows) What do you say? |
| 38 |
|
| 39 |
Cheers, |
| 40 |
Benda |
Archives