| 1 |
On Mon, Feb 01, 2010 at 11:06:15PM -0700, Denis Dupeyron wrote: |
| 2 |
> Agreed. I would suggest to use this series of GLEPs as examples of |
| 3 |
> what to do for future GLEP writers. |
| 4 |
I've actual considered breaking them up even further. |
| 5 |
|
| 6 |
> replace |
| 7 |
> "We should be prepared to add stronger checksums wherever possible, |
| 8 |
> and to remove those that have been defeated." |
| 9 |
> with: |
| 10 |
> "Stronger checksums shall be added as soon as an implementation is |
| 11 |
> available in Portage. Weak checksums may be removed as long as the |
| 12 |
> depreciation process is followed (see below)." |
| 13 |
+1 on that wording, I'll commit when I am next able to. Do see already |
| 14 |
how I partially adapted for Cardoe's requests. |
| 15 |
|
| 16 |
> And then, in "Checksum depreciation timing" I would prefer that the |
| 17 |
> description of what needs to be done in the present situation was used |
| 18 |
> as an example after a more general rule is stated. Something like: |
| 19 |
I'd propose the following variant: |
| 20 |
===== |
| 21 |
A minimum set of depreciated checksums shall be maintained only to |
| 22 |
support old package manager versions where needed by historically used |
| 23 |
trees: |
| 24 |
- New package manager versions should NOT use depreciated checksums in |
| 25 |
Manifests when newer checksums are available. |
| 26 |
- New trees with that have never used the depreciated checksums may omit |
| 27 |
them for reasons of size, but are still strongly suggested to include |
| 28 |
them. |
| 29 |
- Removal of depreciated checksums shall happen after no less than 18 |
| 30 |
months or one major Portage version cycle, whichever is greater. |
| 31 |
===== |
| 32 |
|
| 33 |
I'm not editing these into the checksum GLEP right now, as I'm sitting |
| 34 |
in the food court at the airport, with my flight boarding shortly. |
| 35 |
Most probably will be committed from my hotel room tomorrow night in |
| 36 |
Brussels. |
| 37 |
|
| 38 |
Signing off now, from an airport with bizarre wireless: |
| 39 |
More than 45% of connections fail to start, but when they do, they're |
| 40 |
good for more than 20Mbit. That's after you deal with the their broken |
| 41 |
DNS in the auth mechanism (I'm probably the only person with working |
| 42 |
wireless right now, so it's all the bandwidth to myself). |
| 43 |
|
| 44 |
-- |
| 45 |
Robin Hugh Johnson |
| 46 |
Gentoo Linux: Developer, Trustee & Infrastructure Lead |
| 47 |
E-Mail : robbat2@g.o |
| 48 |
GnuPG FP : 11AC BA4F 4778 E3F6 E4ED F38E B27B 944E 3488 4E85 |
Archives