cvs commit: ports/devel Makefile
danfe at FreeBSD.org
Mon Dec 26 09:22:41 UTC 2011
On Mon, Dec 26, 2011 at 12:44:22PM +0400, Ruslan Mahmatkhanov wrote:
> Alexey Dokuchaev wrote on 26.12.2011 12:33:
> > It always surprised me how could people come up with dot-containing port
> > names in the first place... Not to mention that this `.core' issue had
> > bitten us already in the past.
> Alexey, this naming was discussed  _before_ i ever starting to add
> all this ports and there was no strong objections (and this is not
> prohibited anywhere in Porters Handbook, Committers Guide etc) about
> this. The point is to name it as close as upstream do - the same policy
> apply to p5 ports, so why there should be exception for py- ports?
Ruslan, I perfectly understand that this naming does not contradict some
letter of PH or CG. I was just making an observation that dot is a poorly
chosen symbol for this purpose. This conflict due to ".core" being magic
extension does not look unexpected at all -- these things are very likely
to happen when a port's compound name uses symbols other than [_-] as
separators. I also respect and appreciate your asking on the ports@, yet I
feel sorry that you've picked up `py-zope.foobar' instead of
While mimicking upstream name is a nice thing to pursue, maintaining
consistency within our own Ports Collection is IMHO more important. Given
that we have to rename .core -> -core for at least two ports already, we've
broken both upstream ill-designed scheme, and consistency with other py-zope
ports. In this light, sticking with dash from the very beginning would be a
win for everyone (including our user).
Sorry I have missed this discussion. Doug had perfectly valid reasoning. I
fully support his assessment that while we already have plenty of portnames
with dot in them, it's not a good reason to add more.
More information about the cvs-all