Things to graph for package-of-the-day work

[Jonathan Lange]

On Mon, Nov 16, 2009 at 12:19 AM, Martin Pool <mbp at canonical.com> wrote:
> 2009/11/16 Jonathan Lange <jml at canonical.com>:
>>>>> 1. Packages with upstream imports
>>>>>  - Number of packages (maybe in an interesting subset)
>>>>
>>>> Packages in Lucid main: 16143
>>>> Packages in Karmic main: 16217
>>>>
>>>>>  - Number of these packages with upstream links
>>>>
>>>> lucid: 16
>>>> karmic: 340
>>>
>>> So 'upstream links' in this context means just a link from a distro
>>> package to a Launchpad product, and from there it may have a link to a
>>> trunk branch?
>>>
>>
>> To a Launchpad product series. From there it has a link to a product,
>> which then in turn has a link to a product series which has a link to
>> a branch.
>>
>> I'm not sure if this answer actually helps.
>
> It does confirm my tentative understanding was correct.
>
> Then your numbers seem to show that the weakest link in the import
> chain may be those upstream links?  And it should be fairly easily
> fixed, considering the data to complete say 90% of them is probably
> already in the system?
>

Is it? It's certainly not in the official linking table.

jml

PS. Expect greater latency on the next reply.