Short version: I can confirm that bug under linux, but the patch breaks nis module on solaris. Linux machine is: Linux malhar 2.2.16-3smp #1 SMP Mon Jun 19 17:37:04 EDT 2000 i686 unknown with python version from recent CVS. I see the reported bug and the suggested patch does fix the problem. Sparc box looks like this: SunOS cfa0 5.8 Generic_108528-06 sun4u sparc SUNW,Ultra-Enterprise using python2.0 source tree. The nis module works out of the box, but applying the suggested patch breaks it: 'nis.error: No such key in map'. --Greg Ball
Greg Ball writes:
Short version: I can confirm that bug under linux, but the patch breaks nis module on solaris.
I'm presuming that these were using the same NIS server? I'm wondering if this may be an endianess-related problem. I don't understand enough about the NIS protocols to know what's going on in that module. -Fred -- Fred L. Drake, Jr. <fdrake at acm.org> PythonLabs at Digital Creations
On Fri, Jun 29, 2001 at 10:03:28AM -0400, Fred L. Drake, Jr. wrote:
Greg Ball writes:
Short version: I can confirm that bug under linux, but the patch breaks nis module on solaris.
I'm presuming that these were using the same NIS server? I'm wondering if this may be an endianess-related problem. I don't understand enough about the NIS protocols to know what's going on in that module.
It's my suspicion that it depends how the "aliases" map is built. The patch that "broke" things for the Linux systems includes the comment /* created with 'makedbm -a' */ which makes me suspect that it's dependant on the way the map is constructed. (I couldn't find an online makedbm manpage which documents a -a option) Endian issues should not exist, the protocol below NIS/YP takes care of this. Jeff
participants (3)
-
Fred L. Drake, Jr. -
Greg Ball -
Jeff Epler