saving a few ports from death
ade at FreeBSD.org
Tue Apr 26 05:55:23 UTC 2011
On Apr 25, 2011, at 21:41 , Charlie Kester wrote:
> Maybe freshports could implement a voting system like the one at
"Voting" implies some kind of democracy.
This may come as a shock to folks, but FreeBSD in general is in fact not democratic. It's based around the concept of folks putting in their own time to keep a part of the Project alive.
Whether it's a device driver, some chunk of base userland, ports/<foo>/<bar>, or support for an entire architecture and/or release of FreeBSD -- doesn't matter. Without at least a modicum of active maintainership (hint: MAINTAINER= ports at FreeBSD.org is _not_ active) then it will eventually fall by the wayside and die.
For ports, there's a non-zero cost associated with each and every single one in the tree. Directly, in terms of clusters trying to build packages for the combination of supported releases and architectures, and indirectly, in case of infrastructural changes affecting chunks of the tree, when it comes to determining whether port breakage is a result of said changes, or whether it was broken already.
Generally speaking, such "dead" ports are marked DEPRECATED with a sizable amount of time before being reaped. Honestly, I'd personally prefer the variable to be named PUT_UP_OR_SHUT_UP (but that's just me). The fact remains though, that that is _exactly_ what it is. If, in the period between a port being marked DEPRECATED and it being removed from the tree, and especially in the case of UNMAINTANED=YES (that'd be ports at FreeBSD.org for those in the back), no-one steps up to (a) fix the problem, (b) take maintainership and (c) _continue_ with maintainership.
Well, in that case, the port does not _deserve_ to live. After all, no-one cares about it. If they did, they'd take care of (a) thru (c) above.
More information about the freebsd-ports