| 1 |
Per http://kennethreitz.com/announcing-requests-v100.html, version |
| 2 |
1.0.0 defines the public API (he's saying they're doing SemVer). |
| 3 |
|
| 4 |
So I think we should keep things as is for now, maybe just make sure |
| 5 |
the rdeps either (a) work with 1.x or (b) restrict them from 1.x. |
| 6 |
|
| 7 |
Cheers, |
| 8 |
|
| 9 |
Dirkjan |
| 10 |
|
| 11 |
On Wed, Jan 2, 2013 at 5:28 PM, Mike Gilbert <floppym@g.o> wrote: |
| 12 |
> requests seems to break its api frequently, which sucks for having it |
| 13 |
> installed system-wide. |
| 14 |
> |
| 15 |
> The few packages that depend on it have < or = deps in the tree. This |
| 16 |
> sucks because it could cause dependency conflicts. |
| 17 |
> |
| 18 |
> * These packages depend on requests: |
| 19 |
> dev-python/python-keystoneclient-0.2.1 (dev-python/requests) |
| 20 |
> dev-python/python-keystoneclient-9999 (dev-python/requests) |
| 21 |
> dev-python/python-novaclient-2.10.0 (test ? dev-python/requests) |
| 22 |
> dev-python/python-novaclient-9999 (test ? dev-python/requests) |
| 23 |
> dev-python/sphinxcontrib-issuetracker-0.10.1 (>=dev-python/requests-0.13) |
| 24 |
> net-misc/flexget-1.0_beta3181 (=dev-python/requests-0.14*) |
| 25 |
> net-misc/flexget-9999 (=dev-python/requests-0.14*) |
| 26 |
> net-misc/httpie-0.3.0 (>=dev-python/requests-0.10.1) |
| 27 |
> net-misc/httpie-0.3.0-r1 (>=dev-python/requests-0.10.1) |
| 28 |
> (<dev-python/requests-1.0) |
| 29 |
> net-misc/httpie-0.3.1 (>=dev-python/requests-0.10.1) |
| 30 |
> (<dev-python/requests-1.0) |
| 31 |
> |
| 32 |
> I follow flexget somewhat closely, and their upstream is probably in no |
| 33 |
> hurry to port it to the new api. They suggest using a virtualenv instead |
| 34 |
> of installing it system-wide. |
| 35 |
> |
| 36 |
> The QA team would probably tell us to mask requests-1.0* or the reverse |
| 37 |
> dependencies. Any preference here? |
| 38 |
> |
| 39 |
> Does anyone have any better ideas? |
| 40 |
> |
Archives