[Python-bugs-list] [ python-Bugs-457633 ] _cursesmodules.c does not build on MacOSX

noreply@sourceforge.net noreply@sourceforge.net
Wed, 12 Dec 2001 05:20:38 -0800


Bugs item #457633, was opened at 2001-09-01 16:12
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=105470&aid=457633&group_id=5470

Category: Extension Modules
Group: None
Status: Open
Resolution: Remind
Priority: 7
Submitted By: Jack Jansen (jackjansen)
Assigned to: Jack Jansen (jackjansen)
Summary: _cursesmodules.c does not build on MacOSX

Initial Comment:
_cursesmodule.c suddenly stopped build on MacOSX.
(this is from the CVS repository).

The first error is on "chtype" on line 187, and you don't want the know the amount of errors it causes after that. I guess that some optional feature of
curses was added without an appropriate ifdef or something.

----------------------------------------------------------------------

>Comment By: Jack Jansen (jackjansen)
Date: 2001-12-12 05:20

Message:
Logged In: YES 
user_id=45365

It turns out that the curses supplied with OSX is far too
old to be useable by cursesmodule. Building _curses is now
disabled on OSX.

----------------------------------------------------------------------

Comment By: Guido van Rossum (gvanrossum)
Date: 2001-12-11 08:53

Message:
Logged In: YES 
user_id=6380

I'm reassigning this to Jack because Andrew is shedding
load.

Jack, is this still a problem? If you can't fix it, can you
lower the priority? It's not going to hold up the 2.2
release...

----------------------------------------------------------------------

Comment By: A.M. Kuchling (akuchling)
Date: 2001-09-04 12:38

Message:
Logged In: YES 
user_id=11375

>(are there no other platforms that have this problem?)

Probably not, as most systems are SYSV-based these days.
It wouldn't work on, say, SunOS 4.1, but who uses such an 
old version?  I'm surprised MacOS X has such an old 
version of curses, since current *BSDs don't have a 
problem.



----------------------------------------------------------------------

Comment By: Guido van Rossum (gvanrossum)
Date: 2001-09-04 08:03

Message:
Logged In: YES 
user_id=6380

Reopening, hoping for a better way to fix this.

This should really check for a magic string that occurs
in all recent curses.h files, or one that only occurs
in ancient (unusable) cursees.h files, to distinguish
between a good and a bad curses.

----------------------------------------------------------------------

Comment By: Jack Jansen (jackjansen)
Date: 2001-09-04 02:07

Message:
Logged In: YES 
user_id=45365

It turns out that the curses on MacOSX is a really old
(1994?) Berkeley curses, not
good enough for _cursesmodule.c. I've "manually" disabled it
in setup.py (are there
no other platforms that have this problem?).

I have no idea why the problem only showed up recently,
there appear to have been no changes around this area in
setup.py...

----------------------------------------------------------------------

Comment By: Guido van Rossum (gvanrossum)
Date: 2001-09-01 19:15

Message:
Logged In: YES 
user_id=6380

When did it last build for you without problems?  The last
change was from 19 July, and unless MacOS X defines either
sgi or __sgi, I don't see how that could cause lots of
errors. The 'chtype' typedef isn't affected by the last two
checkins.

Maybe something changed in some header included y Python.h
or pyport.h? Maybe something you changed yourself?

----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=105470&aid=457633&group_id=5470