Maintainership of py-zopetesting and py-zopeevent
Ruslan Mahmatkhanov
cvs-src at yandex.ru
Tue Sep 6 07:19:09 UTC 2011
Doug Barton wrote on 06.09.2011 04:38:
> On 09/05/2011 16:35, wen heping wrote:
>> Would you send a PR of repocopy to rename these ports?
>
> My understanding is that we don't do port names with . in them. Can
> someone who knows more than I confirm one way or the other?
Doug, Mark,
here is my point why there is nothing actually criminal:
a) porters handbook doesn't discourages dot usage in port names (but it
suggest to use PREFIX and POSTFIX for the cases where port name contains
`-` in it).
b) we already have plenty of portnames with dot in them:
[rm at smeshariki3 www]> find /usr/ports -type d -depth 2 -name "*.*"
-print | wc -l
105
c) it's not convenient to maintain ports with some (compeletely
unnecessary) parts of them tweaked. as for now, DISTNAME,
PYDISTUTILS_PKGNAME should be changed to avoid the dots and to make this
ports actually work. Please see this list:
http://pypi.python.org/pypi?%3Aaction=search&term=zope&submit=search
It's a lot of work to change some Makefile parts for all of them if we
ever decide to port them all.
d) while working on that, i erroneously ported already existing deps
just because i wasn't able to find it, and this is something terrible
that i really angry of.
How would one decide which deps is needed to one port or another? He
will consult the port docs, offsite requirements, may be check the code
itself. So, for example, i want to port some code that depends on
"zope.proxy" (just that, as it may be founded in distribution's setup.py
and all the docs), so what should i (the user) do?
`make search name=zope.proxy`. Oh, nothing there, so it seems i should
port it too (OR - so it seems we lacking some dependency so i'm lazy to
go with it and port it too). How could i know that somebody decides to
name it "zopeproxy" just to avoid some dots in the name? I think that
principle of least surprise should be there for such cases, otherwise
user just can't find the port that they need.
I believe that all of this is reasonable enough to pass some new dots to
the tree, isn't it?
--
Regards,
Ruslan
Tinderboxing kills... the drives.
More information about the freebsd-ports
mailing list