| 1 |
On 12/03/10 13:05, Michał Górny wrote: |
| 2 |
> The first question that comes into my mind is -- why do we need |
| 3 |
> to communicate that? I think that USE_PYTHON is a pretty specific |
| 4 |
> variable which should be used only if specially required (i.e. to keep |
| 5 |
> multiple Python versions ready for use). |
| 6 |
> |
| 7 |
> What needs to be fixed IMO is the default value of that variable. |
| 8 |
> If you already disabled switching the active Python version by default, |
| 9 |
> you should also make the Python eclass base its' USE_PYTHON defaults |
| 10 |
> on the active Python version. |
| 11 |
|
| 12 |
USE_PYTHON hits people when emerging a new 2.x slot. |
| 13 |
Say they have 2.6 installed and add 2.7 to that. |
| 14 |
|
| 15 |
I agree, that its default value may need re-consideration. |
| 16 |
After that, patches to the python eclass are needed. |
| 17 |
|
| 18 |
Depending on how fast we can fix that, communicating status quo before |
| 19 |
that may still help reduce user frustration. Especially as we have an |
| 20 |
unmasked 2.7.1 in tree for a few days now. |
| 21 |
|
| 22 |
|
| 23 |
>> - mentioning in 'man make.conf' |
| 24 |
> |
| 25 |
> Not related. We don't describe VIDEO_CARDS, LIRC_DEVICES etc. in that |
| 26 |
> file too. |
| 27 |
|
| 28 |
Yes, but that's USE_EXPAND stuff - USE_PYTHON is different, at least at |
| 29 |
the moment. You may still be right, though. |
| 30 |
|
| 31 |
|
| 32 |
|
| 33 |
Sebastian |
Archives