[csw-maintainers] /usr/local references, and packages

[Ben Walton]

Excerpts from Philip Brown's message of Thu Jan 27 14:17:23 -0500 2011:

> > Ok, it at least _feels_ new.  I don't recall seeing packages
> > rejected for this until very recently.  More strictly enforced?
> 
> 
> If one were to analyze the packages submitted in the last 6 months
> (SUBMITTED, not only accepted), one might find an odd peak in the
> frequency of things containing /usr/local in the last month or so.
> Dont know why, it just happens sometimes

Well, this indicates a problem in the way the release process is
currently operating then, even when looking at this 6 month window.
The original 5.10 release wasn't flagged for this but you indicate the
tools have not changed.  This means that your process is not being
applied routinely.

> > Again, I don't think the point of this is that what you're saying
> > is wrong, although I think you could argue that anything in
> > share/doc is fair game for an exclusion at the maintainers
> > discretion.  What I'm hearing is that people perceive this as a
> > 'new rule' whether or not you feel that it's new.
> 
> This should not be about "feelings". It is really inappropriate to
> say, "well, some people *feel* like this is new, so even though
> facts say it's not new at all, lets ignore facts and go with peoples
> *feelings*"

Well, the facts aren't solid either, but that aside...  If a large
enough number of people are being hit by this and wondering why
packages are being rejected, then there is a problem.  Lets fix the
problem.

Things to fix:

1. This check needs to be applied to every package, every time,
   consistently.  That's not happening now or the original 5.10
   release would have been snagged too...and the 5.8 before that,
   etc.
2. Maintainers need to understand what the policy is.  If someone like
   Peter who's been around for a long time is being caught off guard
   by this, the policy is not understood widely enough.

Again, I'm not arguing that blocking things with /usr/local is wrong.
I'm just saying that it needs to be fully qualified.  You mentioned
that "it's so completely obvious it hasn't needed definition," so that
should mean it's very simple to lay down the policy, including things
that are valid exceptions.  Please do and then let maintainers comment
on it.  Revise if necessary.  My hunch is that you won't need to, but
let's see.  We can vote on it if there is enough variance in opinion.

At the end of the day, this makes _your_ life easier.

> but it isnt "harmless". its just "very very rarely used".  If
> someone were to actually use that functionality, it would be fatal
> to the compile(I think), if no change was made.  That is not
> "harmless".

No, it is harmless.  Go read the docs in the perl distribution.  It's
not something that is accidentally going to C++ your leg. :)

Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302