| 1 |
-----BEGIN PGP SIGNED MESSAGE----- |
| 2 |
Hash: SHA256 |
| 3 |
|
| 4 |
On 05/05/14 01:42 PM, Micha³ Górny wrote: |
| 5 |
> Dnia 2014-05-05, o godz. 09:23:56 Ian Stakenvicius <axs@g.o> |
| 6 |
> napisa³(a): |
| 7 |
> |
| 8 |
>> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 |
| 9 |
>> |
| 10 |
>> On 05/05/14 04:29 AM, Micha³ Górny wrote: |
| 11 |
>> |
| 12 |
>>> 3. deprecates multilib_for_best_abi() since having two separate |
| 13 |
>>> concepts of 'best ABI' and 'default ABI' is confusing, and |
| 14 |
>>> mostly doesn't serve any real purpose. |
| 15 |
>>> |
| 16 |
>>> For improved consistency, we would like people to use |
| 17 |
>>> multilib-minimal and multilib_is_native_abi() tests if |
| 18 |
>>> necessary. |
| 19 |
>>> |
| 20 |
>>> |
| 21 |
>>> I will submit the patches in replies to this mail. |
| 22 |
>>> |
| 23 |
>> |
| 24 |
>> multilib_for_best_abi was introduced to deprecate |
| 25 |
>> multilib_is_native_abi though, aren't we going backwards? |
| 26 |
> |
| 27 |
> Honestly, I don't remember why it was introduced. I just checked |
| 28 |
> the commit message and relevant mails, and it's all quite laconic. |
| 29 |
> It was introduced as part of multibuild_for_best_variant(), and |
| 30 |
> that benefited mostly distutils-r1 for its *_all() phases. |
| 31 |
> |
| 32 |
> I think multilib_for_best_abi() was mostly intended to help |
| 33 |
> getting autotools-multilib to work properly. Now it is built on top |
| 34 |
> of multilib-minimal, and people are encouraged to redefine the |
| 35 |
> multilib_* phases rather than try to hack on top of |
| 36 |
> 'autotools-utils_src_compile' and stuff. This makes most of |
| 37 |
> multilib_for_best_abi() irrelevant. |
| 38 |
> |
| 39 |
> So, I don't think we are really going backwards here. We've |
| 40 |
> changed direction over the past year. We've seen what caught better |
| 41 |
> and I'm mostly trying to make things simpler. As part of that, I'd |
| 42 |
> like to remove redundant APIs and focus on supporting one |
| 43 |
> best-supported interface for multilib. At the point, |
| 44 |
> multilib-minimal seems to be the way forward. |
| 45 |
> |
| 46 |
> Do you agree with me on this? Do you have another ideas? |
| 47 |
> |
| 48 |
|
| 49 |
Nope, this makes sense now. I have a sneaky suspicion that my memory |
| 50 |
had some cross-talk between multilib_for_best_abi and |
| 51 |
multilib_build_binaries, too... if multilib_for_best_abi was always |
| 52 |
based on multilib_is_native_abi, then I expect it will be fine to |
| 53 |
deprecate it. |
| 54 |
|
| 55 |
|
| 56 |
|
| 57 |
|
| 58 |
|
| 59 |
|
| 60 |
-----BEGIN PGP SIGNATURE----- |
| 61 |
Version: GnuPG v2.0.22 (GNU/Linux) |
| 62 |
|
| 63 |
iF4EAREIAAYFAlNn2LEACgkQ2ugaI38ACPDddAEAthcqIbx9/4TBM0rfqlDnXdk7 |
| 64 |
ZeFzOkHlUYv7xNGBoFMBALZItkBcJVO8VNQ1bvUYVf+j8W98JWmUt6MBgZdiZZm2 |
| 65 |
=a6pm |
| 66 |
-----END PGP SIGNATURE----- |
Archives