[98388] trunk/dports/databases

[Jeremy Lavergne]

It isn't any worse than stealth updates: it would still be out of our hands, a calculated risk.

Blair Zajac <blair at orcaware.com> wrote:

>On 10/05/2012 10:06 AM, Jeremy Lavergne wrote:
>>> Still, this would be only valid for all existing hashes. Imagine a
>new changeset would appear which would increase the default to 8.
>>> If that short hash had been used at MacPorts it would likely cause
>an error, wouldn't it.
>>
>> Given that their "book" says the largest git repositories they've
>seen just barely require 12 digits... I don't think we really need to
>worry. If paranoid we could use 12 or we could use the whole thing: as
>long as it works, great.
>>
>> If there's a collision, it will error out rather than give you bad
>data:
>> http://stackoverflow.com/a/7128558
>
>The problem for us is if we pick a truncated hash and then there's a 
>collision, one needs to figure out which of the two untruncated hashes 
>is the correct one.  Presumably, it would be the older one, but for 
>unpopular ports, it could be a while to figure that out.
>
>Seems like picking something 12 or above then means you don't need to 
>worry about this problem.
>
>Blair