Re: pkg/45345 (lang/python31 wrongly builds pyexpat)

["OBATA Akio" <obache%netbsd.org@localhost>]

The following reply was made to PR pkg/45345; it has been noted by GNATS.

From: "OBATA Akio" <obache%netbsd.org@localhost>
To: gnats-bugs%netbsd.org@localhost
Cc: 
Subject: Re: pkg/45345 (lang/python31 wrongly builds pyexpat)
Date: Tue, 06 Dec 2011 09:55:58 +0900

 On Tue, 06 Dec 2011 01:45:01 +0900, David Holland 
<dholland-pbugs%netbsd.org@localhost> wrote:
 
 >   >  >   >>>  How do you think about databases/py-sqlite3? it is same 
 > situation.
 >   >  >   >>>  ${PYLIB}/sqlite3/* exists in base python package, but lack of 
 > _sqlite3.so.
 >   >  >   >> I think that's wrong too.
 >   >  >   >
 >   >  >   >  Then, this issue should be send to upstream first.
 >   >  >   >  I think that it is the default behavior of python distribution.
 >   >  > I don't think so. Both py-sqlite3 and py-expat are built out of the
 >   >  >  main python distfile. And look at the first hunk of
 >   >  >  lang/python26/patches/patch-am, which suppresses the standard build 
 > of
 >   >  >  those and other builtin modules.
 >   >
 >   > It's just exactly disabling to built those modules, or builtin
 >   > sqlite3 libraries may be detected (It is not expected behavior).
 > Right. But the reason it does that is that we build the sqlite
 >  material separately in the py-sqlite3 package.
 > That is, that separation is something we did, not part of the default
 >  behavior of Python. There would be no point contacting upstream.
 
 No.
 Your question point is that *.py files depending on expat and sqlite3 modules
 will be installed even if those modules are not installed (disabled or 
depending
 libraries are not found).
 It is the default behavior and you can do change request to upstream.
 
 -- 
 OBATA Akio / obache%NetBSD.org@localhost