The vim port needs a refresh

Matthias Andree matthias.andree at gmx.de
Sat May 25 11:36:30 UTC 2013


Am 25.05.2013 13:28, schrieb John Marino:
> On 5/25/2013 13:24, Chris Rees wrote:
>> On 25 May 2013 11:54, Niclas Zeising<zeising+freebsd at daemonic.se>  wrote:
>>> On 05/25/13 10:50, Chris Rees wrote:
>>>>
>>>> Alternatively, perhaps we need an editors/vim-options port????
>>>
>>> Just for the record, editors/vim was (and shells/bash) was converted to
>>> optionsNG not too long ago.
>>
>> Ah, that's at least some good news.  I notice that it was on yet
>> another maintainer timeout, so that criticism stands.
>>
>> It appears that David is no longer interested.
> 
> FWIW, the default on the vim port have taken the dports users by
> surprise.  I've gotten several complaints about the boatload of ports
> that get sucked in (and the amount of bandwidth it requires) by vim.
> They didn't know vim-lite existed.
> 
> I agree the default should be "light" and the kitchen sink version
> should be explicitly requested (if two ports are indeed needed for
> pre-built binary reasons).

I routinely install vim-lite on FreeBSD, and move on.

Yes, it may surprise first-time users, but either you have the GUI stuff
installed anyways and don't care, or you get into the habit of
installing vim-lite on headless and otherwise low-end machines.

Frequent timeouts are a reason to reset the maintainer, though. We have
policies for that, and we have been re-rolling distributions all the
way, from SVN repos, or Github, for example, so why bother.

As to the patch hosting, if some port fails and we modified it in any
way, we're the first contact of the end user anyhow to pre-filter
between what we've botched, and what needs forwarding upstream -- and
there should be plenty of vim mirrors around -- perhaps a case to deploy
phttpget for bulk fetching of a thousand patches from HTTP-based mirrors.

Oh, and look at how snappy portsnap has become on -HEAD. Not sure what
it does differently than 9.1-REL, but it sure feels a lot snappier. If
that's not phttpget we should consider if we can leverage that.



More information about the freebsd-ports mailing list