svn commit: r422505 - head/archivers/snappy-java

[John Marino]

On 9/21/2016 08:48, Mathieu Arnold wrote:
> Le 21/09/2016 à 15:43, John Marino a écrit :
>>
>> I'm trying to help.
>
> And it is a noble cause. In the end, what it does is give us more work
> because the cache you are using are stuck with one version and will need
> way more work to figure out when there is an update.

Then this is the hangup.  I *never* intended the change to the 
MASTER_SITE to continue past the existing version.  The idea is that 
when somebody updates the version, they change the MASTER_SITE back to 
upstream.  It is a provision and, in my opinion, reliable source for the 
exact distfile being requested.

How exactly am I causing "more work"?  Changing 1 additional line?  If 
so, then I consider the benefits outweight the cost of changing that one 
line.

>
>> It also puts FreeBSD distcache back in the backup role, which is the
>> primary discriminator for marking a port broken.
>
> Not it does not, a cache like our distcache, or those other OS caches
> cannot be the upstream.

first, I don't concede it's a cache in the purest sense and secondly, as 
long as that source is available it serves as a reliable host.

Maybe it's time for portmgr to research this and publish a list of 
forbidden upstreams if they find it really is a cache and find they 
really don't want it.

John

---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus