On 3/15/2010 4:49 PM, Cristopher Ewing wrote:
If it is the case that a buildout should not have an egg-info directory, _and_
it is also the case that paster can't cope with committing a package to svn
without one, then my instinct is to reject this issue as being a problem with
paster and not an problem with zopeskel.
Thoughts from the gallery?
I prefer to see tickets which depend on upstream issues getting fixed
If an upstream ticket is opened (and I don't think from what Ian tells
me that any tickets opened with paster are going to get fixed unless we
fix them), then that ticket could get referenced in the downstream ticket.
The reason I prefer this is because:
a) it encourages the proper cooperation between upstream and downstream,
b) it insures we don't forget about upstream problems,
c) it provides a better picture of the health of the downstream,
d) it keeps us honest.
I know a lot of people in Plone culture just want to see all bugs in
their tracker closed. But I don't believe in closing them until they are
fixed, whatever the reason. I don't have a problem with open issues that
are open issues.
The ticket can simply document what the upstream problem is. By leaving
it open with that documentation, it can provide a pointer to how to
ultimately resolve the issue, should anyone be interested in doing so.
Closing the ticket just makes it more unlikely paster will get fixed.
I'm not saying keeping it open makes it more likely. It's just that
closing it does make it less likely.
Also, this may be more of an abuse issue on our part. Paster was created
to create eggs, not buildouts. It has a right to expect egg-info. This
issue might be telling us something.
office: 332 Chapman Hall phone: (919) 599-3530
mail: Campus Box #3300, UNC-CH, Chapel Hill, NC 27599
ZopeSkel mailing list