| 1 |
On Tue, 20 Dec 2016 23:57:10 +0100 |
| 2 |
Jonas Stein <jstein@g.o> wrote: |
| 3 |
|
| 4 |
> The shutdown is in two weeks. |
| 5 |
> One can see clearly a drop of broken ebuilds after the first mail on the |
| 6 |
> mailinglist and after the personal mail on 2016-11-24. [1] Thank you all |
| 7 |
> for fixing so many ebuilds already. |
| 8 |
> But still 400 ebuilds use googlecode. |
| 9 |
|
| 10 |
Double checking here, but surely, packages which have newer google code versions |
| 11 |
don't need the old versions killed prematurely. |
| 12 |
|
| 13 |
For instance, dev-perl/Google-Ads-AdWords-Client had[1] version 2.xxx from GoogleCode, |
| 14 |
but has a version 4.xxx now from CPAN/Github. |
| 15 |
|
| 16 |
I don't think I *need* to kill the 2.xxx version, and I'm no hurry to force |
| 17 |
everyone who was using 2.xxx to make a radical major version shift and break their |
| 18 |
code. |
| 19 |
|
| 20 |
So I think as long as: |
| 21 |
|
| 22 |
1. No mirror restrictions |
| 23 |
2. Its mirrored |
| 24 |
3. There is a clear succession away from google code in newer versions. |
| 25 |
|
| 26 |
We're not harming anyone by keeping old versions from googlecode. |
| 27 |
|
| 28 |
So: |
| 29 |
|
| 30 |
> But still 400 ebuilds use googlecode. |
| 31 |
|
| 32 |
There may be significantly less here to worry about. |
| 33 |
|
| 34 |
|
| 35 |
1: When writing this email I realised they might have mirrored a SHA-identical version |
| 36 |
of the old 2.xxx series to CPAN, and lo and behold, they did!, so this email now becomes |
| 37 |
a purely theoretical exercise :) |
Archives